My Patreon: www.patreon.com/cuivlazygeek My Merch Store: cuiv.myspreadshop.com/ Sorry about the audio clipping from time to time :( Equipment used in demo: ZWO ASI2600MC Duo (actually Air, but that's not available yet): bit.ly/3M8IkPS (Agena) or bit.ly/3Jti7KQ (HPS) or tinyurl.com/54nkj5n8 (ZWO) UMi 17 Lite Mount: www.proxisky.com/ ZWO EAF Focuser: bit.ly/45agP0Z (Agena) or bit.ly/4avCIcE (HPS) or tinyurl.com/nhfyark8 (ZWO) RedCat51: bit.ly/48hyuVx (Agena) or bit.ly/48pTWXW (HPS) Cable Organizers/Focuser bracket: www.etsy.com/shop/SmartComponents My flat panel: artesky.it/it/15-flat-field-box Cheap alternatives! amzn.to/4cuZBht Other Flat Panels: bit.ly/3ob5xsC or bit.ly/46QnFtQ or tinyurl.com/28vsc4aw or tinyurl.com/99ny3rjm Amazon affiliate: amzn.to/4dpFnXE Agena affiliate: bit.ly/3Om0hNG High Point Scientific affiliate: bit.ly/3lReu8R First Light Optics affiliate: tinyurl.com/yxd2jkr2 All-Star Telescope affiliate: bit.ly/3SCgVbV Astroshop eu Affiliate: tinyurl.com/2vafkax8
Outstanding Cuiv !!! Love the fact that you included DSLRs in your tutorial. There are many of us out here who just don’t have the budget for dedicated Astro cams .
New to this type of thing really. watched many of the videos you have made which are excellent watched more than a few times actually and things are becoming clearer. Acquiring my rig slowly, everything ZWO image capture wise, so going for the Asiair plus awaiting delivery, great video on that by the way I have the 294 MC pro, so very useful info on that, EAF and ASI120mm for guiding. Final thought is right now after watching so many videos of yours is what have a Iet myself in for. Many Thanks Cuiv much appreciated.
Would probably not disagree as your 15 or so years in this hobby is so valuable for us. You’re my technical go to guy for all things astrophotography… CS!!
Very Nice video ! Your points about taking cal frames in a dark environment are great advice. It reminds me of, decades ago, when the early DSLR's were introduced we had to worry about light leaking into the camera via the rear view finder or even the I/O port on the side of the camera. We even put tape over the blinking red LED light. Our solution for insurance, and still something I do today with my ASI6200MC Pro is to slip a soft black velvet bag over the entire camera and filter assembly to ensure no light leaks can occur. I also slip over the front of my scope a black velvet cap to ensure no light leaks. This additional lens cover only because the light cap on my TEC140 is somewhat cheap and not very tight.
Well covered topic! Few additions, though: Very general CCD signal equation is: true_calibrated_signal = (raw_signal - bias - dark) / (flat - bias - dark_flat) - every frame has the bias/zero part, in ideal case it is independent of temperature (it is measured artificially added bias voltage (needed for analog-digital converters, to avoid "negative" (in reality, clipped to 0) values)), in reality it may have temperature dependency. Bias/zero frame describes the read-out electronics of your sensor and that's why it is taken ideally with 0 exposure (reset of sensor, following read-out of "empty" pixels), any longer exposure has dark signal contribution. Another thing there is read-out noise, which is *truelly random* fluctuation of digitized analog signal and it is different *every time* one measures a specific pixel - which means it can't be calibrated out. If one averages or combines bias/zero frames, patterns of instrinsic behaviour of pixels become more defined, because random noise component will be suppressed. I personally take (with CCD cameras!) hundreds of bias frames to create a master bias, my general goal is that square_root(number_of_bias_frames) < read-out-noise in electrons. - every frame has dark frame component, dark frame describes how individual pixels generate signal in the absence of light - due to the temperature only. Also, non-uniformly cooled or warm sensors have pixel coordinate dependency of flat dark signal (that is typically the problem of large(r) sensors). Amp glow is another source of dark signal in the case of CMOS cameras. There is an important thing: every signal which is randomly created has a *random noise* component which is equal to square root of signal (photons/electrons/ADUs). If you average N dark frames, it's signal-to-noise ratio (dark noise) is improving also as square root of N. That's why averaging/combining many dark frames is useful - one will end up with (much) better defined dark signal generation capability of individual pixels. I personally create dark libraries (only for CCD cameras which have overscan region!!!) using hundreds of dark frames, the general number of frames follows the same goal as for bias frames. - because every dark frame includes bias frame, IF exposure of dark frame is equal to exposure of flat frame or night sky frame, bias is not needed at all (in fact, using additional bias frames would lower (by definition) calibrated data quality (by a small bit, but still)). That greatly simplifies the pre-processing, CCD signal equation would become then: true_calibrated_signal = (raw_signal - masterdark) / (masterflat) - bias frames are really needed to *scale* dark frames. That works very well for CCD-s and more-or-less doesn't work properly (or at all?) for CMOS cameras. As I wrote that I use hundreds of raw bias and dark frames, that is exactly for the scaling purposes: master bias created from 100 individual raw bias frames has negligible random scatter in addition to instrinsic bias structure, also master dark frame created from 100 raw dark frames has negligible random scatter (dark *noise*) in addition to instrinsic/true dark *signal*. The result of master dark (when removing bias frames) is the map of sensor which describes how different pixels generate dark signal. Because dark signal increases linearly in time (providing that sensor temperature is constant), it is possible scale the master dark (dark signal image) by the ratio of exposure_starrysky / exposure_masterdark. Unfortunately it can't be done correctly for CMOS cameras, because they have other important random noise sources, which are also in the dark frames itself. When scaling such dark frame, it is easily possible to make calibrated night sky data (a bit) worse than it could be. It's better to use dark frames with the same exposure to the flat and starry sky frames - in the case of CMOS sensors. What I described, is mathematically correct way. In the case of digital sensors, there is no continuous signal level, they have discrete ADU values. So there are plenty of situations, when the process can be simplified - e.g. if exposure is so short that dark signal would be virtually 0 (definitely below 1 ADU). And some intrinsic problems (randomly fluctuating noise(s)) can be solved effectively by dithering and different rejection methods during processing. Still, the math holds. Clear skies!
Sorry, got bored after the first paragraph. Probably quicker to just put a link to your RUclips channel. If you don't have one, suggest you create one 👍
Hi Cuiv and thanks for the video. I have two comments: If i'm not mistaken you didn't specify that if you don't use dark frames you have to subtract the offset from the lights by some other means. This is a requirement and a very important thing to understand imo. In practice, this means you need to subtract the bias from the lights. WBPP will do that automatically if you don't provide dark frames. The other thing is that you don't mention is the benefit of making darks to correct hot pixels. In my experience, even with the newest CMOS sensors you get a lot of hot pixels with longer frames and dark correct them very well. So, if you don't use darks, you have to take care of hot pixels in some other way, either with cosmetic correction or with dithering. And even with dithering, hot pixels can be problematic for registration for instance. This is why I personally would not advise a beginner to skip darks.
Yes, Adam Block recently put out a video explaining why dark frames are the preferred method for correcting hot pixels. Among other things, if you use the PixInsight WBPP Cosmetic Correction method for correcting hot pixels, you may be creating black pixels at the centers of saturated stars. This doesn't happen with dark frames. Also, my understanding is that with DSLRs and uncooled cameras, dark frames that don't match the temperature of the sensor can do more damage than good, so going with just bias frames is best unless you can match the temperature by some means.
Merci Cuiv. Cette vidéo tombe à pic. J’avais d’énormes problèmes récemment avec mes fichiers de calibration. Tu m’as donné plusieurs pistes pour régler le problème.
i am very appreciative! you explain in a way where i can understand. thank you so much for all you do and your attention to detail. I have just subscribed.
Hey Cuiv Good of you to mention the parameter of focus My auto focus will change 3 or 4 time a night depending on temperature changes and I set it to do one after a meridian flip. But the change is usually not more than 100 steps all night. I’ve been doing separate flats with different focus settings and have yet to figure how to stack them. But after watching this I think I’ll just set the focus in the middle of the range used and take flats there. Also was not sure about calibration frames when using my ZWO ASI6200 mm Pro with LRGB and HSO filters and you covered that…excellent! Thanks for all of that…..CS!!
Hi Cuiv. I just bought the asi294mc pro camera. I found out that if I put the gain to maximum and take a 5 minute dark frame, I get amp glow, but if I set the gain to 122, no amp glow even for a 15 minute dark frame. I still have to do more testing to see at what point amp glow happens but under normal gain levels, my asi294 works extremely well. I'm so glad amp glow only happens when your pushing the limits of the gain.
Great video, thank you! I understood most of what you presented from my learnings in Charles Bracken’s “Deep Sky Imaging Primer” but your added insights regarding what is necessary and why for varying sensors was something I had not heard anywhere else. Thank you!
Great video! The one thing is missing - if your flats for some reason overcorrecting - adding darks to the stack will fix that! That's what exactly helped me, I've got a lot of bright dust spots when used only flats+darklats with my cooled IMX571C camera. Adding darks - eliminates any overcorrected dust spots.
I started imaging with a DSLR and took all 4 calibration frames at the end of each imaging session. The Darks were the only hassle requiring up to an hour to collect. Last year the DSLR (Nikon 5600) died, and I replaced it with a cooled astrophotography camera with a 571 chip. I still take Flats during the imaging session using a motorized flat panel. Matching Darks, Bias, and Dark Flats are now taken every 6 months during the day using a NINA script, minimizing the hassle, and used to create master calibration frames in PixInsight. I live in the US in Scottsdale AZ with hot summers so I need to take calibration frames for multiple temperatures to cover how low I can cool the camera. I know there is a lot of discussion on the web for need for all the calibration frames. For me the calibration frames are no longer a hassle, and it is better to have all the calibration frames if the need for them ever arises in the future. For a show Suggestion I would like to see how you handle weather events including inputs from weather services, locale sensors, and the ASCOM drivers necessary to interface with NINA etc...
Calibration frames are so important in this hobby as not using them robs you of signal and introduces artifacts. Excellent discussion here! Thanks for the video!
Thanks for such a clear and comprehensive description of the components for sensor corrections. (I wonder how much editing it might have taken to be sure that you got each term correct each time.) I wish there were a way to communicate exactly which parts of sensor non-ideal performance you are measuring at each step, and how each type of image does its job. I try to do that with fellow amateurs, but most just want to know the 'cookbook', not how the baking works. Great job.
Using calibration frames is a critical step. The most challenging for me are the flats. I've experimented using several types of lighting sources, exposure settings, exposure lengths. I have even processed images using flats from different scopes, just to see what role it would play in the final image. Cuiv the information you have provided is right on and absolutely correct. Thanks for another great informational and helpful session to achieve the very best final image !
Thanks so much for this Cuiv! I've struggled with flats for a loooong time because I have the 294mm - it was frustrating as hell because I was following all the wisdom on youtub/cloudynights and could barely eek out a decent flat. Now the variable brightness on my flatpanel will be put to use! Much appreciated, keep up the great content!! 🙏
I am using the ASI533MC Pro, so I set up a dark library when I first got my camera. It took all day to get all of the combinations of settings that I could think of that I use (gain, binning, exposure time, etc) but I have a decent library now. Bias and flats are usually taken out in the field because they are quick and can be done at the end of a session.
Hi Cuiv, very informative video. I own a 294MC pro camera and never had a problem with calibration. I take skyflats (from the bed, i am lazy too) just before sunrise (typically 200-500 ms). My skyflats (25k-32k ADU) match temperature, bias and gain of light frames. I don’t take bias at all, but flatdarks that match the flats parameters but with cap on (i use libraries generated in a dark room). I am aware of the post on cloudy night, nevertheless when I do all preprocessing in Siril my resulting stack file is perfect with no banding or any artifacts. I love the IMX294 sensor ❤. Giuseppe
@@CuivTheLazyGeek Hi Cuiv, yes I used it successfully with both l-enhance and l-extreme. I don’t know if are the skyflats that make a difference. All the best!
@@giuseppececere9815 that's great! Different samples of the camera have different artifacts, but I have seen some that basically unaffected, glad this was the case for you!
Good video explaining how they should be done. Ive read alot about the 294mm needing 3-4 sec flats.ive been using the auto exposure time in the asi air and it has been calibrating well with less than 1 sec exposures.i need to revisit trying longer exposures.
Thank you, I been wondering about this for a long time. I am thinking of getting a camera and an asi-air for trying some astro photography with my gti-WiFi mount setup.
Hello I am a beginner in astrophotography and I have the ASI 294 MC Pro camera! I thought that 1 sec was for all cameras! It will take me 3 times longer to do the job now but at least I will have a better result. Thank you and keep up your good work it is exciting I would like to learn everything quickly but it is a complicated hobby people do not understand all the work that is behind all these photos. Have a nice day! Ben,from Quebec.
great video Cuiv! in my experience Pixinsight can be a bit finnicky with selecting the correct dark/bias frames, so sometimes selecting these manually is necessary for me in the calibration tab. Also if you calibrate multiple filters at the same time in WBPP (like LRGB or SHO), and two or more flat frames for corresponding filters have the exact same exposure time, PI might select the incorrect corresponding flat dark
Love your videos. I learned something new. I never heard the word Parameters pronounced as Para-meter. As an American, we go with Pa-Ram-A-Tur. Great info on Dark frames. Never heard you didn’t need them. My gear is not that advanced.
Wow! That's a lot of useful information. Thank you very much. I would have appreciated some graphics or PowerPoint type word charts to help with my understanding. I'm working with a ZWO ASI294MC Pro. My light panel doesn't get dim enough to get 3-second flats. So, I plan to add cheap solar film to the panel. Yes, this camera has lots of vignetting.
Thanks for the clarity on calibration frames! Especially appreciate the note regarding darks, for my recently acquired 533MC Pro. Regarding multiple imaging sessions for a single target, or even a mosaic, can you speak to the process of which calibration frames to take for each imaging session, and how you would combine in stacking and pre-processing? Specific context is for cooled astronomy camera, but I also use DSLR; potentially different approaches? Thanks, again!
I got my ASI533 a couple of days ago and am very happy to hear I don't need to shoot any dark frames. I'll have to reprocess the Veil Nebula from last night and check the difference. For my ASI294 I noticed the flat frames becoming quite stable over 1.4 seconds. Will try 3sec. And I'll have to shoot Dark Flats - not a problem. I was actually using 1.4sec bias frames, so I can run them in SIRIL. (I am really surprised that no one ever wrote a script just for the ASI294!) I'll have to use DeepSkyStacker for the 294 from now on. DSS runs so much faster than SIRIL, but how "good" is it!?! Thank you for a very informative video; I've learned something today!
Hello cuiv, thank you so much for this video. I have a question though... You said that I don't have to change the parameters for flats, but in particular for the imx294 sensor I have to shoot at least 3 sec exposure. So this means the only thing adjustable is the brightness of the panel, but what if my panel can't go so dim to have the right histogram? It's always overexposed!!
Great video, Cuiv! I have a ZWO ASI 294MM pro camera and I'm struggling with vignetting. Lately, I discovered that the 240 sec. dark frames have light leaks. I'm doing the dark frames capture at night, in the roll-off observatory with the cap on the telescope. I suspect parasite light from an illuminated USB hub in the observatory enters the telescope from the rear, going around the main mirror assembly (it's a 10 inch Newtonian). I have taken dark frames with the camera taken off the telescope and they are ok, but the vignetting got worse. So I suspect that the flat frames are also affected by the parasitic light.
Should the optical train be at the same temperature that you are imaging at? for example if it is 20 degrees outside, should your flats be taken at the same temperature so focus, filter spacing, etc is equal.
Nah, it's fine. Filter spacing doesn't change significantly enough, and focus doesn't matter near as much as people make it (I actually have a video on the topic!)
Hi, Cuiv, thanks for the video, you answered few things i was confused about. I have an ASI 183GT, i use LRGB filters with it, so i need to be talking flats for any of those filters which i wasn't aware.
Hi. I am new on your channel.. did i missed that you did not explained in detail how to take flats with dslr, what settings to use… please help also with those informations… thanks
Excellent explanations! I’d love to know your thoughts on using bias frames AND flat darks for the 2600. I was always under the impression that flat darks also remove the same noise as bias frames, so you could use one or the other.
Great video as always. I find it still better to use darks with 533 and 2600 type sensors as hot pixels can show up in longer exposures that are not present in the bias. Keep up the great work and thank you.
Great video thank you! I use the ASI 585mc Pro and I take the flats with the auto exposure option, it works fine. The 512GB RAM helps the buffering so it not goes wrong.
Been dealing with a dust mote that shows up with my luminance filter. Even though I have taken flats, darks, and biases a silhouette of the mote still shows up in the masters coming out of WBPP. I have a 533mm Pro. Will take you suggests from the video and recalibrate. Thank you Cuiv.
I have a light panel for tracing and it has 3 brightness settings. It looks like it has some stuck pixels. Like someone drew 3 or 4 tiny lines with a pencil. They are just elongated dots. Will this be an issue? should I use a white cotton sheet to diffuse the light? Should I move the panel between the flat shots to dither?
Excellent video, Cuiv! Wish I had a cooled astro camera, would have made life a lot easier for calibration frames. I have a DSLR and uncooled ASI585MC - had I known the cooled Pro version would be out in a matter of months, would have waited to get that instead of the uncooled one. Also, thanks for the suggestion to take Flats at the same temperature as Lights. Always thought Flats were temperature independent. Now I really want a cooled camera!
Great informative video as always. Can you have a bias frame library for a cooled camera without amp glow if always imaging at the same temperature & gain?
Thanks for clearing up what with which and when for all of these calibration frames! 👍👍 For temp settings... Is that minus 'x' degrees Celsius from ambient, or is it direct reading of the scale? 🤔
So in the camera specs, it's "x degrees below ambient" (e.g. how much it can cool relative to ambient temp), and in the software itself it's the absolute temperature (so if it says -10C in the software, the camera sensor is at -10C). I hope this helps!
Ever since I started calibrating with my ASI2600mc-pro, I only use darks (from a library) flats, and darkflats. No bias. Am I missing some opportunity for better images? I take my flats and darkflats in the morning using the tee-shirt method, and then with the cap on (and the tee-shirt thrown over the cap for good measure). Seems to work well...
I am shooting the same object across multiple nights. Do I need calibration frames for each night of imaging? Can I reuse the same darks, dark flats, and bias when grouping imaging nights together in WBPP?
Bias, dark and flat frames were used to calibrate CCD image sensors. Where uneven sensitivity to pixels is a problem basically for measurenent and scientific use, and read noise is pretty high as well. Longer sub exposures compensates for the relatively high read noise. With CCDs most of time we took fewer but longer exposures to reach a certain total integration time. CMOS cameras have low read noise, basically its no concern, but the shoot noise is more pronounced. As nature of shoot noise is random and can not be calibrated... Lots of shorter CMOS exposures can be stacked as read noise is no practical problem. Just expose until the background is a bit over the noise floor and that is the most efficient exposure. The most dominant source of noise to most CMOS users (unless having access to a dark location) is usually the sky background. Light pollution and moonlight beeing two typical sources for brightening the background sky. Commercial CMOS cameras are made to take splendid images without calibration. No need for bias (unless shooting flats), or flats (when keeping the image sensor clean), or dark frames (software handles hot pixels). My personal experience with CMOS cameras is that plain shooting without calibration works better. Vignetting is handled by software and dust motes are few and far between when taking care before letting loose. So why use calibration routines developed for CCD image sensors when using CMOS image sensors??? My personal opinion based on experience I have is that methods developed for CCD image sensors mainly (not only) deal with problems inherent in one kind of image sensor but not the other one. Discussion around this is never ending... And there is always a people for which is important to feel right by all costs and meens... Just for a feeling... Therefore I see much more sense for my self to use and keep doing what is working for me and not what is recommended by somebody who has different setup, processing workflow, conditions, possibilities objects to capture and so on... Everyone should try what works for him and his situation.
Thank you for your expertise!! I use a 533MC Pro and I have been processing my images with and without calibration frames and have noticed that my images without have more contrast and color with proper integration time. I always felt like data was being lost or corrupted with calibration frames. Every DSO is different so I use different methods for each one. Whatever gives me the best result at the time is what I keep. I have lots of cloudy nights so I am constantly trying unconventional things.
I really disagree with this advice except the conclusion (use whatever works for you)! - why would shot noise be different for CMOS and CCD? It's just SQRT(signal) - Flats calibrate your optics, so are not specific to a sensor type. And I find it infinitely easier to have a clean image in the first place than to have to deal with the vignetting in software (and other weird shadows that can happen with non refractor designs, especially things like RASA) - if you take flats you NEED bia - if your sensor has amp glow (many CMOS do) OR you don't dither you NEED darks for the hot pixels (or to mess with cosmetic correction, but having darks in the first place is imo so much easier) In other words the routines developed for CCD are very much valid for CMOS. If it works for you, great! But advising to not do any calibration is a recipe for frustration at the processing level in my opinion, especially for anyone not using a good refractor. That said, it would make sense to take the calibration frames, and try to stack without them to compare :) then if it looks just as good, no need to bother!
I have the 533 MCpro and have been taking darks but I tested the stacking with and without darks and saw no difference so I will probably will stop using darks. One bit of advice that I'd like to share is that I take my darks and biases inside my refrigerator. It makes the cameras cooling not work very hard and with the cap on my camera I know for sure there are no light leaks.
Bonjour Cuiv. Merci pour ton excellent video. I have a QHY294M-Pro and I had read about the early problems with short exposures but I tried to find out why this sensor would have this kind of a problem. I've searched the web to try to get information on this but all I found is anecdotal references to the problem. Is it really this sensor or is it the implementation by the camera manufacturer that might be the cause? I ran a test where I started with 3 second flats and flat-darks and progressively reduced the exposures all the way down to 0.5 seconds and I have not observed any strange behavior. With exposure time determined by the scope controller could it be erratic timing response from the camera processor or even its usb performance causing milliseconds jitter of the effective exposure time? That being the case, the jitter if significant enough would have a more detrimental effect on short exposures. Perhaps it's already been solved by a firmware update? Keep up the good work. C'est Superb!
Fantastic video. If I'm just using Bias and Flat Frames with my 533MC Pro, what should I tell PixInsight when stacking? I'm thinking I should leave optimize dark unchecked seeing I don't have them?
@@CuivTheLazyGeek Thanks for that. One last question. Is there a need to do Darks or Bias frames for different filters , but same exposure time? One part of me thinks no because it's a dark frame. Another thought is yes because a narrowband filter may come out slightly darker than a broadband filter.
@@yervantparnagian5999 No need, no - for darks and bias (if well done, best to avoid daytime out in the sun in case of light leaks), the sensor is in complete darkness, no photons reach it, filters don't matter!
Salut, j'ai une ASI2600 et c'est bien la première fois que j'entends dire que les BIAS doivent être faits à 1s de pose et pas au minimum. Peux tu expliquer davante ce point STP?Qu'en est il du mode automatique de l'ASIAIR pour les flats? A éviter? Merci bien.
Avec ce capteur, tu peux utiliser le temps de pose minimum. À -10C ou -5C, une seconde ou une milliseconde sont quasiment la même chose (regarde la courbe de dark current pour la température, prend le carré de cette valeur pour une seconde, compare avec le carré du read noise, tu verras que c'est absolument négligeable !)
Awesome video!!!! Love all your stuff!!! Love the fact that you explained the use of different sensors. I’m a beginning astrophotographer, I just bought ASI585mc pro and a new rig. I see some people say they use unity gain of 252 and others say they use a gain of 100. Which gain do you use with the 585?
Hi Cuiv. Yet another great video that cuts through the calibration zealots propaganda. In my first year I used a cooled 533 sensor on a medium focal length (f5.6) refractor, a good dust blower camel hair brush and I kept my light train assembled (filter and all). IMHO, at that skill level, I really didn't need ANY calibration, even when using SIRIL and GIMP and with only rudimentary processing skills. So I went with that. This freed up my attention considerably so I could focus on ... well focus 🧐for one thing, but in general "first things first". That must be the "data capture" challenges. After all, without data what have you? So mastering PA, guide settings, calibration, target and filter selection, exposure, gain, framing, learning the controls (for me asiair) and mount, rig fettling (power, cabling, balance, vibration etc.) are basic skills that need to be forced into muscle memory by repetition. I think that's a sensible approach to recommend because the learning curve in this hobby is close to vertical and simple mistakes can easily ruin your night. I had more than enough on my plate; I didn't really miss the flats for a long time, and I was happy to rely on processing mitigations. Of course that approach needs 'clean' optics (both in design, and dust/dirt) and a modern amp glow free cooled camera at the get go. After a YEAR - once I became competent in capturing data and more able to critically judge my own images (created using free processing tools, and no calibration), THEN I began to look for ways to deal with dust donuts that snuck in, and to optimize those results. By that time I was ready to invest again, to upgrade my PC and choose software (buying PixInsight, blur.. and noise.. and star.. exterminators). At that time I was comfortable with capture sessions and could easily add taking the darks bias flats darkflats into my workflow (using autorun mode to 'remember' my personalized default settings) without the risk of mental overload and forgetting a critical step (eg sharp focus). These days I generally start the night with 20 autoexposure flats, and add 5 more darks to my library (at temp and exposure of the night) along with some darkflats (typically 2-3sec). That way if the battery runs out I will always be able to calibrate however many lights that I manage to capture. I rarely diverge from 120s, -10C, 100 gain lights so I usually have the option to skip darks and save 10 minutes, if I'm worried about weather closing in or delayed by something that needed re-done. Hope this comment helps somebody, and feeds the hungry algorithm! PS On my big APO I keep the flattener/reducer, filter, spacers and 533 permanently assembled in a bag with a dustcap on as my 'DSO' databack, [and I have a similar swapable assembly (with flipmirror) for planets, and can swap in a 2" diagonal and eyepieces for visual too]. My contention is that a refractor's objective (dust and orientation) won't affect flats, and (although Murphy dust will move within the 'databack') most of it is electrostatically fixed, so you'll have a fair chance of rescue by using flats from the previous session (and clonestamp) if it all goes wrong. If you've done the same.
Hi Cuiv, but even my cooled astro cam like the asi533 mc pro at 0°C exhibit lots of hot/cold pixel. So I think it's still necessary to use darks to eliminate it.
Another great vid, many thanks! You solved my UFO problem, as in Unidentified Flat-related Objects 😊 The solution to my problem was that I had a Master Bias at a different gain level. I use NINA, Siril and have a library with masters, B,D,F, but lately was driven mad by UFO's and couldn't identity the cause 😂 . Following your method I took new 1 second Bias frames with different offsets corresponding to the offsets of No-filter Lights and Flats (7) and Dualband Lights and Flats (20) And reprocessed a set of 100 lights of the Bubble Nebula as RGB. And compared with the older results which had UFO's😊 Using Bias with offset7 solved all problems, no dust rings and background gradient is just gone, like really almost zero. So happy😊 This brings up a question: however. I have a manual filterdrawer from svbony, and have the manual filter, set up in NINA, with offsets for the camera set up without the '()' brackets. However, I now think that the Sequencer doesn't handle the offsets and neither does the Flatwizard. Any suggestions anybody? Thanks! Me myself and I
You are not lazy Cuiv! I am lazy. One year into this fascinating hobby I have only taken calibration frames for the notorious 294(only darks). Why? Because I think it is not "worth" the effort/time. At least not for the new cooled sensors. I have even got very nice results with "DSLR's" (I prefer to call them mirrorless *cameras* ) just taking light frames. But I see I have lost some good sessions skipping calibration frames when the image train is not the best (vignetting), and the relevant camera *really* need calibration frames. In those sessions the "stacks" seem to be "impossible" to stretch to a good image regardless of software. Even Pixinsight I just started to use (and like) has no chance to make bad light frames good. 😥 My Canon Ra, R7, OM-1 and the Fujifilm X-H2 seems to be quite "forgiving" without taking calibration frames so far. Edit: And the camera lens is also very important, the cheap TT Artisan 500 F/6.3 impressed me a lot. Much better than expensive Canon lenses (as an example).
Nice! I think you can get away with no calibration frames when using a refractor, but any other designs will demand flat frames (and thus bias frames or flat darks) at minimum :)
I have the L-quad filter, but to reach the target of 20 000, i need to set it to 0.02s for flats with my led panel on the lowest setting. Does that sound right?
Uncooled camera - in Berkeley, Ca, USA, the pacific ocean means nighttime temperature is never over 18C and rarely below 10C. Given that, I'm assuming I don't need dark/bias/flats across temp range, I can just collect one set per exposure-time X gain ( I usually keep at 300 for my SV705c - imx585 sensor) at around 15C and call it good. Make sense?
Thank you never really got me head around the calibration frames… I thought and did my darks with each of my filters and haven’t done the other types yet… at least I know how to do the flats and have I need to use with each filter.
Cuiv, the mention of flat frames using the same parameters (even for cooled astro cameras) worries me. If I'm taking 30 - 60 second lights, surely I can't do the same with flats? In my experience that would make them massively over-exposed.
Hi Cuiv. I have a 571C based camera (Poseidon C from PlayerOne), I use NINA and PixInsight. The problem I have is that NINA's flatwizzard proposes only Flats and DarkFlats, and PI does not propose Darkflats. So I put Darkflats in the bias tab of PI (not the Dark tab). What problem could this generate? Merci mille fois pour ton travail
In ASIAIR I use the Auto Setting for Flats and seem to work well. Cameras ZWO ASI 183GT 2600MM/MC Pro and 6200MM Pro. Have you given the Auto Flats a try?
Are you telling me that with my 294mc pro I can take a set of dark frames, setting for example -10ºC in my ASIAIR plus, and I can reuse it any time I shot a target with the same temperature settings? Even if in one night there are 30ºC and in the next night there are 0ºC outside? Btw i apreciate your videos a lot!!
Hi! I use asi 294mc pro. Calibration frames (dark, flat, bias). After watching the video I realized that instead of bias you need to use dark-flat but the time is not less than 3 seconds. And flat is also not less than 3 seconds. Did I understand everything correctly?
Hi Cuiv, thank you for the excellent video! I have a 294MC Pro and am using Siril for post processing. I specifically apply preset scripts to my narrowband data, OSC_Extract_HaOIII in particular. These scripts ask for bias frames. Any idea how I can get around this problem? Thanks.
When pressing the run button in pxlnsight it says I am missing the master dark and flat files and they will light frames will not be calibrated. I usually run the stacking without the lights to create the master files. You did not mention anything about master darks, lights and bias files. Thank you for everything you do. Master Cuiv. 😢
Thank you! You took together all informations which I heared in the past but not so systematically 🙏 With Asiair in Contrast to you I used up to now the Auto Mode for Flats. Some critics from you?
I noticed that you make your flat frames at gain 100 and yet several people tell us that we need to make the flat frames with a gain of 0. The reason is to give dynamics to the sensor. I have a 294 Mc and my historam is at 2/3 is this correct?. Thank you for your videos, always informative!.J'ai remarqué que vous faites vos flates frame au gain 100 et pourtant plusieurs personnes nous indiquent qu'il faut faire les flates avec un gain 0. La raison est pour donné de la dynamique au capteur. J'ai une 294 Mc et mon historamme est au 2/3 est-ce correct?. Merci pour vos vidéos, toujours instructifs! 👍
Hi Cuiv, Very usefull video, again. What do you think about dark frames helping to get rid of hot pixels (cosmetic correction during lights calibration)? Even the fancy new cooled sensors without amp glow have them! There is a method to remove these hot pixels without a master dark, but this is time consuming during pre-processing, and requires extra disk storage for the corrected images! Also, do you think cosmetic correction adds up to dithering to REALLY remove hot pixels (especially for long narrow band exposures or limited lights count)? Spending rainy days to build a master dark 1 or 2 times per year helps reducing the time needed for cosmetic correction (it's very fast when based on master dark).
I just received my first cooled ZWO camera, a 533. I noticed that when you took a flat frame with your ZWO camera and the ASI air that your flat frame appeared weird looking. I thought that it was just me not knowing what to expect, but I thought that I was making some kind of mistake because my flat frames looked weird just like yours. I do not know why this. When you open up the flat frame in Pixinsight they look normal but on the ASI app they look weird. I don’t recall my flat frames looking weird like that with my DSLR 850. Do you have an explanation as to why this is. Thank you for all your efforts. I’ve learned a lot from you.
Great video, thanks Cuiv! But there is one point that I've never understood... When taking Dark frames, we record both reading noise and temperature noise. So if we substract Dark frames and Bias frames to our Light frames, it would substract the reading noise twice... Wouldn't it?
I have a Nikon Z5 and my Dark frames are statistically the same as my Bias frames. I have stretched them and see no noticeable difference. Am I missing something?
My Patreon: www.patreon.com/cuivlazygeek
My Merch Store: cuiv.myspreadshop.com/
Sorry about the audio clipping from time to time :(
Equipment used in demo:
ZWO ASI2600MC Duo (actually Air, but that's not available yet): bit.ly/3M8IkPS (Agena) or bit.ly/3Jti7KQ (HPS) or tinyurl.com/54nkj5n8 (ZWO)
UMi 17 Lite Mount: www.proxisky.com/
ZWO EAF Focuser: bit.ly/45agP0Z (Agena) or bit.ly/4avCIcE (HPS) or tinyurl.com/nhfyark8 (ZWO)
RedCat51: bit.ly/48hyuVx (Agena) or bit.ly/48pTWXW (HPS)
Cable Organizers/Focuser bracket: www.etsy.com/shop/SmartComponents
My flat panel: artesky.it/it/15-flat-field-box
Cheap alternatives! amzn.to/4cuZBht
Other Flat Panels: bit.ly/3ob5xsC or bit.ly/46QnFtQ or tinyurl.com/28vsc4aw or tinyurl.com/99ny3rjm
Amazon affiliate: amzn.to/4dpFnXE
Agena affiliate: bit.ly/3Om0hNG
High Point Scientific affiliate: bit.ly/3lReu8R
First Light Optics affiliate: tinyurl.com/yxd2jkr2
All-Star Telescope affiliate: bit.ly/3SCgVbV
Astroshop eu Affiliate: tinyurl.com/2vafkax8
Outstanding Cuiv !!! Love the fact that you included DSLRs in your tutorial. There are many of us out here who just don’t have the budget for dedicated Astro cams .
New to this type of thing really. watched many of the videos you have made which are excellent watched more than a few times actually and things are becoming clearer. Acquiring my rig slowly, everything ZWO image capture wise, so going for the Asiair plus awaiting delivery, great video on that by the way I have the 294 MC pro, so very useful info on that, EAF and ASI120mm for guiding. Final thought is right now after watching so many videos of yours is what have a Iet myself in for. Many Thanks Cuiv much appreciated.
Glad this has been helpful! Good luck on this journey!!
Would probably not disagree as your 15 or so years in this hobby is so valuable for us. You’re my technical go to guy for all things astrophotography…
CS!!
Extremely informative and clears up a lot of things that were left unclear by other "informative" videos, thank you so much!
So glad you found it helpful!
Thanks!
Thank you for your support!!
Very Nice video ! Your points about taking cal frames in a dark environment are great advice. It reminds me of, decades ago, when the early DSLR's were introduced we had to worry about light leaking into the camera via the rear view finder or even the I/O port on the side of the camera. We even put tape over the blinking red LED light. Our solution for insurance, and still something I do today with my ASI6200MC Pro is to slip a soft black velvet bag over the entire camera and filter assembly to ensure no light leaks can occur. I also slip over the front of my scope a black velvet cap to ensure no light leaks. This additional lens cover only because the light cap on my TEC140 is somewhat cheap and not very tight.
Hahaha yes! I remember covering the view finder like that!
Great video Cuiv!!! Maybe the best Calibration frames video I’ve seen so far!!! Thank you 😮
This was insanely helpful! Well defined settings, straightforward, yet simple.
So grateful you shared this!
Well covered topic!
Few additions, though:
Very general CCD signal equation is: true_calibrated_signal = (raw_signal - bias - dark) / (flat - bias - dark_flat)
- every frame has the bias/zero part, in ideal case it is independent of temperature (it is measured artificially added bias voltage (needed for analog-digital converters, to avoid "negative" (in reality, clipped to 0) values)), in reality it may have temperature dependency. Bias/zero frame describes the read-out electronics of your sensor and that's why it is taken ideally with 0 exposure (reset of sensor, following read-out of "empty" pixels), any longer exposure has dark signal contribution. Another thing there is read-out noise, which is *truelly random* fluctuation of digitized analog signal and it is different *every time* one measures a specific pixel - which means it can't be calibrated out. If one averages or combines bias/zero frames, patterns of instrinsic behaviour of pixels become more defined, because random noise component will be suppressed. I personally take (with CCD cameras!) hundreds of bias frames to create a master bias, my general goal is that square_root(number_of_bias_frames) < read-out-noise in electrons.
- every frame has dark frame component, dark frame describes how individual pixels generate signal in the absence of light - due to the temperature only. Also, non-uniformly cooled or warm sensors have pixel coordinate dependency of flat dark signal (that is typically the problem of large(r) sensors). Amp glow is another source of dark signal in the case of CMOS cameras. There is an important thing: every signal which is randomly created has a *random noise* component which is equal to square root of signal (photons/electrons/ADUs). If you average N dark frames, it's signal-to-noise ratio (dark noise) is improving also as square root of N. That's why averaging/combining many dark frames is useful - one will end up with (much) better defined dark signal generation capability of individual pixels. I personally create dark libraries (only for CCD cameras which have overscan region!!!) using hundreds of dark frames, the general number of frames follows the same goal as for bias frames.
- because every dark frame includes bias frame, IF exposure of dark frame is equal to exposure of flat frame or night sky frame, bias is not needed at all (in fact, using additional bias frames would lower (by definition) calibrated data quality (by a small bit, but still)). That greatly simplifies the pre-processing,
CCD signal equation would become then: true_calibrated_signal = (raw_signal - masterdark) / (masterflat)
- bias frames are really needed to *scale* dark frames. That works very well for CCD-s and more-or-less doesn't work properly (or at all?) for CMOS cameras. As I wrote that I use hundreds of raw bias and dark frames, that is exactly for the scaling purposes: master bias created from 100 individual raw bias frames has negligible random scatter in addition to instrinsic bias structure, also master dark frame created from 100 raw dark frames has negligible random scatter (dark *noise*) in addition to instrinsic/true dark *signal*. The result of master dark (when removing bias frames) is the map of sensor which describes how different pixels generate dark signal. Because dark signal increases linearly in time (providing that sensor temperature is constant), it is possible scale the master dark (dark signal image) by the ratio of exposure_starrysky / exposure_masterdark. Unfortunately it can't be done correctly for CMOS cameras, because they have other important random noise sources, which are also in the dark frames itself. When scaling such dark frame, it is easily possible to make calibrated night sky data (a bit) worse than it could be. It's better to use dark frames with the same exposure to the flat and starry sky frames - in the case of CMOS sensors.
What I described, is mathematically correct way. In the case of digital sensors, there is no continuous signal level, they have discrete ADU values. So there are plenty of situations, when the process can be simplified - e.g. if exposure is so short that dark signal would be virtually 0 (definitely below 1 ADU). And some intrinsic problems (randomly fluctuating noise(s)) can be solved effectively by dithering and different rejection methods during processing. Still, the math holds.
Clear skies!
Sorry, got bored after the first paragraph. Probably quicker to just put a link to your RUclips channel. If you don't have one, suggest you create one 👍
@@astrokev99 🤣😂😭💀
Hi Cuiv and thanks for the video.
I have two comments:
If i'm not mistaken you didn't specify that if you don't use dark frames you have to subtract the offset from the lights by some other means. This is a requirement and a very important thing to understand imo. In practice, this means you need to subtract the bias from the lights. WBPP will do that automatically if you don't provide dark frames.
The other thing is that you don't mention is the benefit of making darks to correct hot pixels. In my experience, even with the newest CMOS sensors you get a lot of hot pixels with longer frames and dark correct them very well. So, if you don't use darks, you have to take care of hot pixels in some other way, either with cosmetic correction or with dithering. And even with dithering, hot pixels can be problematic for registration for instance.
This is why I personally would not advise a beginner to skip darks.
Yes, Adam Block recently put out a video explaining why dark frames are the preferred method for correcting hot pixels. Among other things, if you use the PixInsight WBPP Cosmetic Correction method for correcting hot pixels, you may be creating black pixels at the centers of saturated stars. This doesn't happen with dark frames. Also, my understanding is that with DSLRs and uncooled cameras, dark frames that don't match the temperature of the sensor can do more damage than good, so going with just bias frames is best unless you can match the temperature by some means.
Great video! You covered all my questions regarding darks, flats, and bias frames 👊🤓
Glad it was helpful!
Thank you Cuiv for your work!
I have a ASI 294 MC pro and i didnt know that i should not take bias frames.
I will use Darkflats now for my images.
Another great tutorial, Cuiv; thanks for the refresher. Michael
Glad it helps!
Headed out for my first image attempt. Thanks for all the great videos.
Merci Cuiv. Cette vidéo tombe à pic. J’avais d’énormes problèmes récemment avec mes fichiers de calibration. Tu m’as donné plusieurs pistes pour régler le problème.
Génial! Bon courage j'espère que ça va fonctionner!
i am very appreciative! you explain in a way where i can understand. thank you so much for all you do and your attention to detail. I have just subscribed.
Welcome to the channel, and glad this is helpful!
Hey Cuiv
Good of you to mention the parameter of focus
My auto focus will change 3 or 4 time a night depending on temperature changes and I set it to do one after a meridian flip. But the change is usually not more than 100 steps all night.
I’ve been doing separate flats with different focus settings and have yet to figure how to stack them.
But after watching this I think I’ll just set the focus in the middle of the range used and take flats there.
Also was not sure about calibration frames when using my ZWO ASI6200 mm Pro with LRGB and HSO filters and you covered that…excellent!
Thanks for all of that…..CS!!
Hi Cuiv.
I just bought the asi294mc pro camera. I found out that if I put the gain to maximum and take a 5 minute dark frame, I get amp glow, but if I set the gain to 122, no amp glow even for a 15 minute dark frame. I still have to do more testing to see at what point amp glow happens but under normal gain levels, my asi294 works extremely well. I'm so glad amp glow only happens when your pushing the limits of the gain.
That is very interesting! Thanks for this tidbit of information!
Great video, thank you!
I understood most of what you presented from my learnings in Charles Bracken’s “Deep Sky Imaging Primer” but your added insights regarding what is necessary and why for varying sensors was something I had not heard anywhere else.
Thank you!
I'm glad this helped! I really wanted to make a "all cases covered" video and hopefully this fits the bill :)
Perfect timing! I have a 533mc coming this week. Upgrading from a mirrorless and I wasn't positive on how to do calibration frames!
Glad it's helpful, enjoy the 533MC!
Great work. Love the wide spectrum approach.
Thank you! I really tried to cover everything:)
Great video! The one thing is missing - if your flats for some reason overcorrecting - adding darks to the stack will fix that! That's what exactly helped me, I've got a lot of bright dust spots when used only flats+darklats with my cooled IMX571C camera. Adding darks - eliminates any overcorrected dust spots.
I started imaging with a DSLR and took all 4 calibration frames at the end of each imaging session. The Darks were the only hassle requiring up to an hour to collect. Last year the DSLR (Nikon 5600) died, and I replaced it with a cooled astrophotography camera with a 571 chip. I still take Flats during the imaging session using a motorized flat panel. Matching Darks, Bias, and Dark Flats are now taken every 6 months during the day using a NINA script, minimizing the hassle, and used to create master calibration frames in PixInsight. I live in the US in Scottsdale AZ with hot summers so I need to take calibration frames for multiple temperatures to cover how low I can cool the camera. I know there is a lot of discussion on the web for need for all the calibration frames. For me the calibration frames are no longer a hassle, and it is better to have all the calibration frames if the need for them ever arises in the future.
For a show Suggestion I would like to see how you handle weather events including inputs from weather services, locale sensors, and the ASCOM drivers necessary to interface with NINA etc...
Cooled cameras make such a difference in terms of laziness, I love them :)
Calibration frames are so important in this hobby as not using them robs you of signal and introduces artifacts. Excellent discussion here! Thanks for the video!
Thank you for explaining about calibration frames, Cuiv! You really explained this so clearly for me!!!
Glad it was helpful!
Thanks for such a clear and comprehensive description of the components for sensor corrections. (I wonder how much editing it might have taken to be sure that you got each term correct each time.) I wish there were a way to communicate exactly which parts of sensor non-ideal performance you are measuring at each step, and how each type of image does its job. I try to do that with fellow amateurs, but most just want to know the 'cookbook', not how the baking works. Great job.
Thank you!
Using calibration frames is a critical step. The most challenging for me are the flats. I've experimented using several types of lighting sources, exposure settings, exposure lengths. I have even processed images using flats from different scopes, just to see what role it would play in the final image. Cuiv the information you have provided is right on and absolutely correct. Thanks for another great informational and helpful session to achieve the very best final image !
I bought a cheap LED tracing panel FROM Amazon. Seems to work well and doesn't require a white tee shirt.
Yep, flats are really difficult to get right - for me getting a flat panel (or a LED tracing panel) was a game changer
Thanks so much for this Cuiv! I've struggled with flats for a loooong time because I have the 294mm - it was frustrating as hell because I was following all the wisdom on youtub/cloudynights and could barely eek out a decent flat. Now the variable brightness on my flatpanel will be put to use! Much appreciated, keep up the great content!! 🙏
That's awesome!! I hope it will work for you!
I am using the ASI533MC Pro, so I set up a dark library when I first got my camera. It took all day to get all of the combinations of settings that I could think of that I use (gain, binning, exposure time, etc) but I have a decent library now. Bias and flats are usually taken out in the field because they are quick and can be done at the end of a session.
Nice! It's good to have the whole flow down like that
Hi Cuiv, very informative video. I own a 294MC pro camera and never had a problem with calibration. I take skyflats (from the bed, i am lazy too) just before sunrise (typically 200-500 ms). My skyflats (25k-32k ADU) match temperature, bias and gain of light frames. I don’t take bias at all, but flatdarks that match the flats parameters but with cap on (i use libraries generated in a dark room). I am aware of the post on cloudy night, nevertheless when I do all preprocessing in Siril my resulting stack file is perfect with no banding or any artifacts. I love the IMX294 sensor ❤. Giuseppe
Nice! Is that even with dual band narrowband filters? This is where typically the 294 can show issues!
@@CuivTheLazyGeek Hi Cuiv, yes I used it successfully with both l-enhance and l-extreme. I don’t know if are the skyflats that make a difference. All the best!
@@giuseppececere9815 that's great! Different samples of the camera have different artifacts, but I have seen some that basically unaffected, glad this was the case for you!
Thank you! This is the best explanation I've seen.
Glad it was helpful!
Thanks Cuiv for another great video. Even after doing this for a few years there's always something new to learn.
Good video explaining how they should be done. Ive read alot about the 294mm needing 3-4 sec flats.ive been using the auto exposure time in the asi air and it has been calibrating well with less than 1 sec exposures.i need to revisit trying longer exposures.
Thank you, I been wondering about this for a long time. I am thinking of getting a camera and an asi-air for trying some astro photography with my gti-WiFi mount setup.
Nice! You may also consider the ASI2600MC Air in that case (camera+guider+asiair in one package)
Hello I am a beginner in astrophotography and I have the ASI 294 MC Pro camera! I thought that 1 sec was for all cameras! It will take me 3 times longer to do the job now but at least I will have a better result.
Thank you and keep up your good work it is exciting I would like to learn everything quickly but it is a complicated hobby people do not understand all the work that is behind all these photos.
Have a nice day!
Ben,from Quebec.
Like always, a very instructive comprehensive video in the subject, thanks for your contribution in our community.
great video Cuiv! in my experience Pixinsight can be a bit finnicky with selecting the correct dark/bias frames, so sometimes selecting these manually is necessary for me in the calibration tab. Also if you calibrate multiple filters at the same time in WBPP (like LRGB or SHO), and two or more flat frames for corresponding filters have the exact same exposure time, PI might select the incorrect corresponding flat dark
That's right on the flat dark! You need to set the dark tolerance to zero in WBPP to avoid that :)
Bonjour. Super Video! Mais où mettre les dark-flat dans l’Asiair quand on a une 294? À la place des Bias?
Love your videos. I learned something new. I never heard the word Parameters pronounced as Para-meter. As an American, we go with Pa-Ram-A-Tur. Great info on Dark frames. Never heard you didn’t need them. My gear is not that advanced.
Hahaha my English pronunciation should NOT be used as a reference :)
Wow! That's a lot of useful information. Thank you very much. I would have appreciated some graphics or PowerPoint type word charts to help with my understanding.
I'm working with a ZWO ASI294MC Pro. My light panel doesn't get dim enough to get 3-second flats. So, I plan to add cheap solar film to the panel. Yes, this camera has lots of vignetting.
Thanks for the clarity on calibration frames! Especially appreciate the note regarding darks, for my recently acquired 533MC Pro. Regarding multiple imaging sessions for a single target, or even a mosaic, can you speak to the process of which calibration frames to take for each imaging session, and how you would combine in stacking and pre-processing? Specific context is for cooled astronomy camera, but I also use DSLR; potentially different approaches? Thanks, again!
I got my ASI533 a couple of days ago and am very happy to hear I don't need to shoot any dark frames. I'll have to reprocess the Veil Nebula from last night and check the difference.
For my ASI294 I noticed the flat frames becoming quite stable over 1.4 seconds. Will try 3sec. And I'll have to shoot Dark Flats - not a problem. I was actually using 1.4sec bias frames, so I can run them in SIRIL. (I am really surprised that no one ever wrote a script just for the ASI294!) I'll have to use DeepSkyStacker for the 294 from now on.
DSS runs so much faster than SIRIL, but how "good" is it!?!
Thank you for a very informative video; I've learned something today!
Nebula Photos has an older video comparing various stacking programs!
Hello cuiv, thank you so much for this video. I have a question though... You said that I don't have to change the parameters for flats, but in particular for the imx294 sensor I have to shoot at least 3 sec exposure. So this means the only thing adjustable is the brightness of the panel, but what if my panel can't go so dim to have the right histogram? It's always overexposed!!
Great video, Cuiv! I have a ZWO ASI 294MM pro camera and I'm struggling with vignetting. Lately, I discovered that the 240 sec. dark frames have light leaks. I'm doing the dark frames capture at night, in the roll-off observatory with the cap on the telescope. I suspect parasite light from an illuminated USB hub in the observatory enters the telescope from the rear, going around the main mirror assembly (it's a 10 inch Newtonian). I have taken dark frames with the camera taken off the telescope and they are ok, but the vignetting got worse. So I suspect that the flat frames are also affected by the parasitic light.
Should the optical train be at the same temperature that you are imaging at? for example if it is 20 degrees outside, should your flats be taken at the same temperature so focus, filter spacing, etc is equal.
Nah, it's fine. Filter spacing doesn't change significantly enough, and focus doesn't matter near as much as people make it (I actually have a video on the topic!)
Hi, Cuiv, thanks for the video, you answered few things i was confused about. I have an ASI 183GT, i use LRGB filters with it, so i need to be talking flats for any of those filters which i wasn't aware.
Hi. I am new on your channel.. did i missed that you did not explained in detail how to take flats with dslr, what settings to use… please help also with those informations… thanks
Excellent explanations! I’d love to know your thoughts on using bias frames AND flat darks for the 2600. I was always under the impression that flat darks also remove the same noise as bias frames, so you could use one or the other.
Yep, any of those work - and flat darks being so short on a cooler camera are effectively equivalent to bias :)
Thanks Cuiv! ASI294mm pro is on the way. Thanks for the info!!!
Awesome!! Tiny pixels for the win! Just calibrate well and you'll have a blast!
Great video as always. I find it still better to use darks with 533 and 2600 type sensors as hot pixels can show up in longer exposures that are not present in the bias. Keep up the great work and thank you.
Yep, good dithering (and rejection stacking) is necessary to completely avoid dark frames but they never hurt for sure!
Thank you for this very good video, very detailed! A question for ASIAIR. For Flats, why don't you use automatic exposure?
I learned a lot about the 294 ! Thank you ever so much. I have a question for you: In Siril do you simply drop the Dark Flats in the Bias folder?
Great video thank you! I use the ASI 585mc Pro and I take the flats with the auto exposure option, it works fine. The 512GB RAM helps the buffering so it not goes wrong.
Once again, and again and again...
thanks Cuiv
My pleasure! 😀
Been dealing with a dust mote that shows up with my luminance filter. Even though I have taken flats, darks, and biases a silhouette of the mote still shows up in the masters coming out of WBPP. I have a 533mm Pro. Will take you suggests from the video and recalibrate. Thank you Cuiv.
Good luck, I hope it helps!
I have a light panel for tracing and it has 3 brightness settings. It looks like it has some stuck pixels. Like someone drew 3 or 4 tiny lines with a pencil. They are just elongated dots. Will this be an issue? should I use a white cotton sheet to diffuse the light? Should I move the panel between the flat shots to dither?
Excellent video, Cuiv! Wish I had a cooled astro camera, would have made life a lot easier for calibration frames. I have a DSLR and uncooled ASI585MC - had I known the cooled Pro version would be out in a matter of months, would have waited to get that instead of the uncooled one. Also, thanks for the suggestion to take Flats at the same temperature as Lights. Always thought Flats were temperature independent. Now I really want a cooled camera!
Great informative video as always. Can you have a bias frame library for a cooled camera without amp glow if always imaging at the same temperature & gain?
Yes!
Thanks for clearing up what with which and when for all of these calibration frames! 👍👍
For temp settings... Is that minus 'x' degrees Celsius from ambient, or is it direct reading of the scale? 🤔
So in the camera specs, it's "x degrees below ambient" (e.g. how much it can cool relative to ambient temp), and in the software itself it's the absolute temperature (so if it says -10C in the software, the camera sensor is at -10C). I hope this helps!
@@CuivTheLazyGeek Yep! Thanks! That does help...
Ever since I started calibrating with my ASI2600mc-pro, I only use darks (from a library) flats, and darkflats. No bias. Am I missing some opportunity for better images? I take my flats and darkflats in the morning using the tee-shirt method, and then with the cap on (and the tee-shirt thrown over the cap for good measure). Seems to work well...
No that's fine as well! Darks contain the bias signal, and dark flats are effectively equivalent to bias :)
I am shooting the same object across multiple nights. Do I need calibration frames for each night of imaging? Can I reuse the same darks, dark flats, and bias when grouping imaging nights together in WBPP?
Bias, dark and flat frames were used to calibrate CCD image sensors. Where uneven sensitivity to pixels is a problem basically for measurenent and scientific use, and read noise is pretty high as well. Longer sub exposures compensates for the relatively high read noise. With CCDs most of time we took fewer but longer exposures to reach a certain total integration time.
CMOS cameras have low read noise, basically its no concern, but the shoot noise is more pronounced. As nature of shoot noise is random and can not be calibrated... Lots of shorter CMOS exposures can be stacked as read noise is no practical problem. Just expose until the background is a bit over the noise floor and that is the most efficient exposure.
The most dominant source of noise to most CMOS users (unless having access to a dark location) is usually the sky background. Light pollution and moonlight beeing two typical sources for brightening the background sky.
Commercial CMOS cameras are made to take splendid images without calibration. No need for bias (unless shooting flats), or flats (when keeping the image sensor clean), or dark frames (software handles hot pixels).
My personal experience with CMOS cameras is that plain shooting without calibration works better. Vignetting is handled by software and dust motes are few and far between when taking care before letting loose.
So why use calibration routines developed for CCD image sensors when using CMOS image sensors???
My personal opinion based on experience I have is that methods developed for CCD image sensors mainly (not only) deal with problems inherent in one kind of image sensor but not the other one.
Discussion around this is never ending... And there is always a people for which is important to feel right by all costs and meens... Just for a feeling...
Therefore I see much more sense for my self to use and keep doing what is working for me and not what is recommended by somebody who has different setup, processing workflow, conditions, possibilities objects to capture and so on...
Everyone should try what works for him and his situation.
Thank you for your expertise!! I use a 533MC Pro and I have been processing my images with and without calibration frames and have noticed that my images without have more contrast and color with proper integration time. I always felt like data was being lost or corrupted with calibration frames. Every DSO is different so I use different methods for each one. Whatever gives me the best result at the time is what I keep. I have lots of cloudy nights so I am constantly trying unconventional things.
I really disagree with this advice except the conclusion (use whatever works for you)!
- why would shot noise be different for CMOS and CCD? It's just SQRT(signal)
- Flats calibrate your optics, so are not specific to a sensor type. And I find it infinitely easier to have a clean image in the first place than to have to deal with the vignetting in software (and other weird shadows that can happen with non refractor designs, especially things like RASA)
- if you take flats you NEED bia
- if your sensor has amp glow (many CMOS do) OR you don't dither you NEED darks for the hot pixels (or to mess with cosmetic correction, but having darks in the first place is imo so much easier)
In other words the routines developed for CCD are very much valid for CMOS.
If it works for you, great! But advising to not do any calibration is a recipe for frustration at the processing level in my opinion, especially for anyone not using a good refractor.
That said, it would make sense to take the calibration frames, and try to stack without them to compare :) then if it looks just as good, no need to bother!
I have the 533 MCpro and have been taking darks but I tested the stacking with and without darks and saw no difference so I will probably will stop using darks. One bit of advice that I'd like to share is that I take my darks and biases inside my refrigerator. It makes the cameras cooling not work very hard and with the cap on my camera I know for sure there are no light leaks.
That refrigerator tip is a great idea!!
Bonjour Cuiv. Merci pour ton excellent video.
I have a QHY294M-Pro and I had read about the early problems with short exposures but I tried to find out why this sensor would have this kind of a problem. I've searched the web to try to get information on this but all I found is anecdotal references to the problem. Is it really this sensor or is it the implementation by the camera manufacturer that might be the cause?
I ran a test where I started with 3 second flats and flat-darks and progressively reduced the exposures all the way down to 0.5 seconds and I have not observed any strange behavior. With exposure time determined by the scope controller could it be erratic timing response from the camera processor or even its usb performance causing milliseconds jitter of the effective exposure time? That being the case, the jitter if significant enough would have a more detrimental effect on short exposures. Perhaps it's already been solved by a firmware update?
Keep up the good work. C'est Superb!
Fantastic video. If I'm just using Bias and Flat Frames with my 533MC Pro, what should I tell PixInsight when stacking? I'm thinking I should leave optimize dark unchecked seeing I don't have them?
It won't matter, since you don't have anything in the Darks tab!
Thank you for the answer
@@CuivTheLazyGeek Thanks for that. One last question. Is there a need to do Darks or Bias frames for different filters , but same exposure time? One part of me thinks no because it's a dark frame. Another thought is yes because a narrowband filter may come out slightly darker than a broadband filter.
@@yervantparnagian5999 No need, no - for darks and bias (if well done, best to avoid daytime out in the sun in case of light leaks), the sensor is in complete darkness, no photons reach it, filters don't matter!
@@CuivTheLazyGeek Thank you for the quick answers. I put my telescope in a dark room when I do Biases and Darks/Flat Darks. Keep up the great work.
Salut, j'ai une ASI2600 et c'est bien la première fois que j'entends dire que les BIAS doivent être faits à 1s de pose et pas au minimum. Peux tu expliquer davante ce point STP?Qu'en est il du mode automatique de l'ASIAIR pour les flats? A éviter? Merci bien.
Avec ce capteur, tu peux utiliser le temps de pose minimum. À -10C ou -5C, une seconde ou une milliseconde sont quasiment la même chose (regarde la courbe de dark current pour la température, prend le carré de cette valeur pour une seconde, compare avec le carré du read noise, tu verras que c'est absolument négligeable !)
Ah, et jamais utilisé le mode automatique de l'ASIAIR pour les flats donc je ne peux pas me prononcer
Awesome video!!!! Love all your stuff!!! Love the fact that you explained the use of different sensors. I’m a beginning astrophotographer, I just bought ASI585mc pro and a new rig.
I see some people say they use unity gain of 252 and others say they use a gain of 100. Which gain do you use with the 585?
Regarding Flat frames. Can the camera be rotated from the position the lights were taken at?
Hi Cuiv. Yet another great video that cuts through the calibration zealots propaganda. In my first year I used a cooled 533 sensor on a medium focal length (f5.6) refractor, a good dust blower camel hair brush and I kept my light train assembled (filter and all). IMHO, at that skill level, I really didn't need ANY calibration, even when using SIRIL and GIMP and with only rudimentary processing skills. So I went with that.
This freed up my attention considerably so I could focus on ... well focus 🧐for one thing, but in general "first things first". That must be the "data capture" challenges. After all, without data what have you? So mastering PA, guide settings, calibration, target and filter selection, exposure, gain, framing, learning the controls (for me asiair) and mount, rig fettling (power, cabling, balance, vibration etc.) are basic skills that need to be forced into muscle memory by repetition. I think that's a sensible approach to recommend because the learning curve in this hobby is close to vertical and simple mistakes can easily ruin your night. I had more than enough on my plate; I didn't really miss the flats for a long time, and I was happy to rely on processing mitigations. Of course that approach needs 'clean' optics (both in design, and dust/dirt) and a modern amp glow free cooled camera at the get go.
After a YEAR - once I became competent in capturing data and more able to critically judge my own images (created using free processing tools, and no calibration), THEN I began to look for ways to deal with dust donuts that snuck in, and to optimize those results.
By that time I was ready to invest again, to upgrade my PC and choose software (buying PixInsight, blur.. and noise.. and star.. exterminators). At that time I was comfortable with capture sessions and could easily add taking the darks bias flats darkflats into my workflow (using autorun mode to 'remember' my personalized default settings) without the risk of mental overload and forgetting a critical step (eg sharp focus). These days I generally start the night with 20 autoexposure flats, and add 5 more darks to my library (at temp and exposure of the night) along with some darkflats (typically 2-3sec). That way if the battery runs out I will always be able to calibrate however many lights that I manage to capture. I rarely diverge from 120s, -10C, 100 gain lights so I usually have the option to skip darks and save 10 minutes, if I'm worried about weather closing in or delayed by something that needed re-done. Hope this comment helps somebody, and feeds the hungry algorithm!
PS On my big APO I keep the flattener/reducer, filter, spacers and 533 permanently assembled in a bag with a dustcap on as my 'DSO' databack, [and I have a similar swapable assembly (with flipmirror) for planets, and can swap in a 2" diagonal and eyepieces for visual too]. My contention is that a refractor's objective (dust and orientation) won't affect flats, and (although Murphy dust will move within the 'databack') most of it is electrostatically fixed, so you'll have a fair chance of rescue by using flats from the previous session (and clonestamp) if it all goes wrong. If you've done the same.
So I have a 294 mcpro. So ur saying take darks at same exposure time as lights? And dark flats at 3 seconds? And that's it no ligh flats
Darks at same exposure and settings as lights.
Flats at same settings as lights, but 3s-5s long
Dark flats at same exposure and settings as flats
Hi Cuiv, but even my cooled astro cam like the asi533 mc pro at 0°C exhibit lots of hot/cold pixel. So I think it's still necessary to use darks to eliminate it.
What ISO to use on a DSLR when using AsiAir to control it? Set it to auto, or to ISO100?
Another great vid, many thanks! You solved my UFO problem, as in Unidentified Flat-related Objects 😊
The solution to my problem was that I had a Master Bias at a different gain level.
I use NINA, Siril and have a library with masters, B,D,F, but lately was driven mad by UFO's and couldn't identity the cause 😂 .
Following your method I took new 1 second Bias frames with different offsets corresponding to the offsets of No-filter Lights and Flats (7) and Dualband Lights and Flats (20) And reprocessed a set of 100 lights of the Bubble Nebula as RGB. And compared with the older results which had UFO's😊
Using Bias with offset7 solved all problems, no dust rings and background gradient is just gone, like really almost zero. So happy😊
This brings up a question:
however. I have a manual filterdrawer from svbony, and have the manual filter, set up in NINA, with offsets for the camera set up without the '()' brackets.
However, I now think that the Sequencer doesn't handle the offsets and neither does the Flatwizard.
Any suggestions anybody?
Thanks!
Me myself and I
You are not lazy Cuiv! I am lazy. One year into this fascinating hobby I have only taken calibration frames for the notorious 294(only darks). Why? Because I think it is not "worth" the effort/time. At least not for the new cooled sensors. I have even got very nice results with "DSLR's" (I prefer to call them mirrorless *cameras* ) just taking light frames. But I see I have lost some good sessions skipping calibration frames when the image train is not the best (vignetting), and the relevant camera *really* need calibration frames. In those sessions the "stacks" seem to be "impossible" to stretch to a good image regardless of software. Even Pixinsight I just started to use (and like) has no chance to make bad light frames good. 😥
My Canon Ra, R7, OM-1 and the Fujifilm X-H2 seems to be quite "forgiving" without taking calibration frames so far. Edit: And the camera lens is also very important, the cheap TT Artisan 500 F/6.3 impressed me a lot. Much better than expensive Canon lenses (as an example).
Nice! I think you can get away with no calibration frames when using a refractor, but any other designs will demand flat frames (and thus bias frames or flat darks) at minimum :)
Curious how USB speed affects the image data. I thought it just adjusted the data flow to the storage medium.
It shouldn't in theory....but? Just in case!
I have the L-quad filter, but to reach the target of 20 000, i need to set it to 0.02s for flats with my led panel on the lowest setting. Does that sound right?
Cuiv, why is the ASI AIR stacking the BIAS Frames and the Flat Frames ? Are they not supposed to be individual shots ? thx in advance, Kenneth
Uncooled camera - in Berkeley, Ca, USA, the pacific ocean means nighttime temperature is never over 18C and rarely below 10C. Given that, I'm assuming I don't need dark/bias/flats across temp range, I can just collect one set per exposure-time X gain ( I usually keep at 300 for my SV705c - imx585 sensor) at around 15C and call it good. Make sense?
It makes sense to me, it sounds like a good compromise!
Thank you never really got me head around the calibration frames… I thought and did my darks with each of my filters and haven’t done the other types yet… at least I know how to do the flats and have I need to use with each filter.
Cuiv, the mention of flat frames using the same parameters (even for cooled astro cameras) worries me. If I'm taking 30 - 60 second lights, surely I can't do the same with flats? In my experience that would make them massively over-exposed.
Question. With 533mc pro, are biases required if I use flats and dark-flats at 2.5 sec? (also using lights and darks obv)
In that case, no need for bias!
Thanks again - i always learn something from you!
👍👍
Always a pleasure!
Cuiv, you are great as usual ! Thanks a lot !
Great video. Maybe you mentioned it in the video and I missed it. How many of each time of calibration frame do you recommend?
I think Cuiv mentioned 20
Sorry for late reply - I don't really have recommendations, but I use 50 bias, 25 flats usually!
Excellent! thank you, loads of useful info
Hi Cuiv. I have a 571C based camera (Poseidon C from PlayerOne), I use NINA and PixInsight. The problem I have is that NINA's flatwizzard proposes only Flats and DarkFlats, and PI does not propose Darkflats. So I put Darkflats in the bias tab of PI (not the Dark tab). What problem could this generate? Merci mille fois pour ton travail
Great video Cuiv. Loved it...
In ASIAIR I use the Auto Setting for Flats and seem to work well. Cameras ZWO ASI 183GT 2600MM/MC Pro and 6200MM Pro. Have you given the Auto Flats a try?
I actually haven't tried it yet! I like doing things manually so I can show the why and how in such a case,but I'll try that next!
Are you telling me that with my 294mc pro I can take a set of dark frames, setting for example -10ºC in my ASIAIR plus, and I can reuse it any time I shot a target with the same temperature settings? Even if in one night there are 30ºC and in the next night there are 0ºC outside?
Btw i apreciate your videos a lot!!
Hi! I use asi 294mc pro. Calibration frames (dark, flat, bias). After watching the video I realized that instead of bias you need to use dark-flat but the time is not less than 3 seconds. And flat is also not less than 3 seconds. Did I understand everything correctly?
Hi Cuiv, thank you for the excellent video! I have a 294MC Pro and am using Siril for post processing. I specifically apply preset scripts to my narrowband data, OSC_Extract_HaOIII in particular. These scripts ask for bias frames. Any idea how I can get around this problem? Thanks.
When pressing the run button in pxlnsight it says I am missing the master dark and flat files and they will light frames will not be calibrated. I usually run the stacking without the lights to create the master files. You did not mention anything about master darks, lights and bias files. Thank you for everything you do. Master Cuiv. 😢
Thank you! You took together all informations which I heared in the past but not so systematically 🙏
With Asiair in Contrast to you I used up to now the Auto Mode for Flats. Some critics from you?
I noticed that you make your flat frames at gain 100 and yet several people tell us that we need to make the flat frames with a gain of 0. The reason is to give dynamics to the sensor. I have a 294 Mc and my historam is at 2/3 is this correct?.
Thank you for your videos, always informative!.J'ai remarqué que vous faites vos flates frame au gain 100 et pourtant plusieurs personnes nous indiquent qu'il faut faire les flates avec un gain 0. La raison est pour donné de la dynamique au capteur. J'ai une 294 Mc et mon historamme est au 2/3 est-ce correct?.
Merci pour vos vidéos, toujours instructifs! 👍
Hi Cuiv,
Very usefull video, again.
What do you think about dark frames helping to get rid of hot pixels (cosmetic correction during lights calibration)? Even the fancy new cooled sensors without amp glow have them!
There is a method to remove these hot pixels without a master dark, but this is time consuming during pre-processing, and requires extra disk storage for the corrected images! Also, do you think cosmetic correction adds up to dithering to REALLY remove hot pixels (especially for long narrow band exposures or limited lights count)?
Spending rainy days to build a master dark 1 or 2 times per year helps reducing the time needed for cosmetic correction (it's very fast when based on master dark).
Pixel rejection during stacking completely removes hot pixels for all my imaging (I don't use cosmetic correction)
Thank you so much for this. Although it made my brain hurt it answered so many of my questions
Glad it helped! Sorry about the headache 😂😆
I just received my first cooled ZWO camera, a 533. I noticed that when you took a flat frame with your ZWO camera and the ASI air that your flat frame appeared weird looking. I thought that it was just me not knowing what to expect, but I thought that I was making some kind of mistake because my flat frames looked weird just like yours. I do not know why this. When you open up the flat frame in Pixinsight they look normal but on the ASI app they look weird. I don’t recall my flat frames looking weird like that with my DSLR 850. Do you have an explanation as to why this is. Thank you for all your efforts. I’ve learned a lot from you.
Great video, thanks Cuiv!
But there is one point that I've never understood... When taking Dark frames, we record both reading noise and temperature noise. So if we substract Dark frames and Bias frames to our Light frames, it would substract the reading noise twice... Wouldn't it?
I eventually got my answer here:
deepskystacker.free.fr/english/theory.htm#CalibrationProcess
I have a Nikon Z5 and my Dark frames are statistically the same as my Bias frames. I have stretched them and see no noticeable difference. Am I missing something?