Excellent video. I have been an amateur astronomy for 60 years and I am learning a lot about the new technologies from your videos. I belonged to the Von Braun astronomical society in the 1970's when we had a 21 inch Cassegrain I would enjoy using and making photos with my Nikkormat camera on Tri-X film and then developing and printing - many did not turn so well as they do today with the digital cameras.... I don't know what they have now as I left Huntsville in 1976 for WSMR in New Mexico. The new star trackers are going to allow me to continue to take my camera and lens to the desert as the big telescope is getting too heavy. Thanks for your videos. Please keep up the good work.
Thank you for showing how ISO really works with the camera. It is only brightening the images, not taking in more light. Wow! Presently, I now only use a high ISO to find stars in the city, then lower it afterwards in frame to reduce glare.
I use a Pentax which has built in astro tracker and it surprisingly works fairly well,It uses it's in body image stabilization to physically track the stars as they move. I have done a full minute exposure at iso 200 and f5 for a bit of sharpness and works awesome.
Hi there, I have a question. I get a lot of noise from my 18-55mm kit lens @ ISO 6400 (18mm), 15 sec. exposure and f3.5 and this setting is the one I used to make the Milky Way details appear on my camera (Nikon D3400). I've purchased the Tokina 14-20mm f2 PRO and I am still waiting for it to arrive. Now, my question is will I get less noise from the Tokina lens from its wide aperture and longer exposure (since I'd be able to use 25 sec. exposure on it @ 14mm and at f2 aperture) while still using the ISO 6400? I'm into Milky Way timelapse and the noise on my kit lens is just horrible even after post-processing :(. Thank you and cheers!
Forgive my newbie ignorance. Is this what you're saying? Altering the ISO on a digital camera is simply amplifying the light (the signal from the sensor). It doesn't allow the camera to magically collect more light. So there is no point in doing this as photoshop will do the same thing and possibly with less noise. The only way to increase the light is to increase the exposure by using a tracker and or a faster lens. To me this makes perfect sense. Many experts tell you to crank up the ISO. And, stacking doesn't magically make dim stars brighter either, it only reduces noise, as noise is random and the stars aren't. However, if you do use long exposures etc, aren't you simply increasing the brightness of everything, dim and bright stars together, and running the risk of over exposing everything. Again, apologies if this is in actual fact rubbish. PS. Excellent vids and web site. So glad I stumbled upon you.
Pretty much, yeah! However, there is an important distinction to make. I recommend doing some tests with your own camera to verify what I'm saying. What you should do is go out one night with your camera. Take a 20 second exposure at f/4, ISO 100. Then, increase the ISO to 200 and take another photo. Continue with 400, 800, 1600, 3200, and ISO 6400. Be sure to leave the shutter speed and aperture the same. You'll also want to be shooting in RAW. Once you have these photos, with the ISO increasing 1 Stop each time, you can test your cameras sensor. Load up all of your images in Adobe Camera RAW at the same time. Then, increase Exposure slider by 5 Stops for the ISO 100 image. You can now directly compare it with the ISO 3200 image. If the ISO 100 image brightened by 5 Stops in Camera RAW looks terrible, with odd colors and grain everywhere, your camera sensor is ISO Variant. You can do this for any of your images. For example, take your ISO 400 image, increase the Exposure slider to 4, and compare it directly against the ISO 6400 image. If your camera is ISO Invariant, there should be no discernible difference between them. It's up to everyone to test their own camera sensors to find out whether it's ISO Variant or Invariant. Again, all that means is that with an ISO Invariant sensor you can increase the Exposure in post-processing, with no quality loss. If you try to increase the Exposure slider with an ISO Variant camera, your image will look terrible. If you have an ISO Variant sensor, you'll want to use a higher ISO in-camera, to get the best results. If you have an ISO Invariant sensor, you can pretty much use any ISO you want, and increase the Exposure as needed in Post-processing.
If your camera isn't very new, it likely won't be ISO invariant. However to estimate the best ISO for old digital cameras, you can check out the ISO-Dynamic Range curves. Above a certain ISO threshold, dynamic range decreases linearly with increasing ISO. Try out the 2 ISO values where the change in slope occurs, the one above and below it. They should be the best performer ISO for old digital cameras. Anything above those would probably increase the noise to uncessary amounts and decrease the dynamic range too much, so it's not the best idea to use the maximum iso for old cameras. Keep in mind that lower iso will yield higher read noise to your images, which in return, unless you subtract them with bias (or dark flats not sure about the name) calibration frames, they will end up in your images. Also higher iso images are more likely to show thermal banding if you expose too long or take too many exposures, which is even worse in most cases. Too high ISO will also decrease your dynamic range which in return will yield overall poor image quality. So finding the optimum is important for old cameras.
Super job, thanks for clarification on the fact to the that ISO means little for end results.... the formula is ....... available light, aperature, and exposure time.... period.... post handles the rest...
idrivavipr yep! More or less anyway. Choosing the wrong ISO can have a negative impact, if your camera sensor isn't great. For example, some cameras have badly ISO variant sensors. Therefore, if you try to increase the Exposure by more than 1 Stop in Camera RAW or Lightroom, you will experience severe image degredation. This will mainly manifest as purple noise in the shadows of the image. Then again, if you would just take a longer shutter speed (2+ minutes), that thermal noise would not be an issue because you captured enough signal (light) that the noise is no longer visible.
A slight digression. I have an Olympus Em5-ii which is classed as "semi-invariant". The problem with Photoshop is that is is not as good at processing the image as the in-camera "engine". Compared to the Panasonic GH4 and G7, the E-M5ii image gets more noise at high ISO. The Panasonic sensor is mounted on a heat sink which reduces sensor noise, the Olympus Body Image Stabilisation preludes fitting the sensor to a heat sink. Both Panasonic and Olympus use a Live MOS (LMOS) sensor as opposed to Complimentary MOS (CMOS), so the sensor is biased to sweep stray electrons from the sensor, which means less noise. The GH4 can re-process an image in-camera, and does a far better job of it than Photoshop does. All three cameras have an Auto-ISO even in Manual mode. It can be a case of trust the camera.
Because most monitors are only 8 bit (per colour channel) when comparing photos (recorded in 14 bit RAW) taken at different ISO's, with exposure boosted in post processing you may not be able to see the difference. Depending on the bit-depth of the image sensors analog to digital converter (ADC), the RAW file of the ISO 100 image could contain 32 x less "brightness resolution" than the RAW file at ISO 3200 even with ISO invariant sensor (in other words the 14 bit RAW file of the ISO 100 image would only contains 9 bit of useful information). So when the ISO 100 image is boosted by 5 stops and displayed along side the ISO 3200 image, they will look the same on the 8 bit monitor, even though the ISO 100 image actually has 32 x less "brightness resolution"
Sir, I am very beginner photographer from a small village of india and i am using very basic camera - Nikon Coolpix P530 Which have certain limitations. I want to photograph milky way and some stars. But as i said my camera have certain limitations that my camera have lowest shutter speed of 15" and the lowest aperture i can set is 3.0. But when i set my camera on 15" shutter speed and f. 3.0. The camera locks ISO on 100 and when i try to increase ISO on 1600 the shutter speed is automatically locks on 1". It seems my camera is allowing me to take pictures only on ISO 100 when shutter speed is settled on 15". So my question is, can i shoot milky way and stars? Can i use the Photoshop technique that you showed in your video? By increasing the ISO by stop method in Photoshop?? Please help me sir. Please reply me.
I understand the concept of reducing noise is to collect more light but doesn't long exposure times increase noise as well? Is the long exposure noise reduction option built into my camera sufficient for battling that particular type of noise?
A longer exposure will increase the amount of hot pixels in the image. You can turn on Long Exposure Noise Reduction in-camera to eliminate them completely. Therefore, a long exposure with LENR turned on tends to give the best results. You'll just need to wait longer for the actual image to process.
You should 've been compared the 2 images with 1:1 zoom to really see that they were identical, because from my experience iso isn't just like pumping the exposure up, but making the sensor's pixels more sensitive. I've seen tremendous difference between images myself. Just to clarify it depends on the quality of your camera too of course, crop sensors usually really mess it up from 1600+ iso while some full frame cams can handle way up to the scale like mofos without particular noise.
It's really all about the camera sensor, and if it's ISO Invariant or not. Also, it depends how much light you capture. The less light you capture, the more grain you'll have. If your camera is ISO Variant, you'll have a lot of problems when you increase the Exposure slider in post. However, if you have an ISO Invariant sensor, then you can pretty much use whatever ISO you want, and increase the Exposure slider in post without any quality loss. The main reason you'd intentionally use a lower ISO is to preserve details in the highlights
@@PeterZelinka It definitely depends on the camera sensor. I think the grain of the noise is being produced by the poor quality of the sensor's pixels because they will make grain as a response of thinking there is light coming. I have a d5300 and if i bump up the iso to 400 I will be totaly okay, but if I lack that light shooting at just 100 iso and try to retrieve it from LR I will get much more noise. In opposite, if I shoot with something extreme like 3200 iso it will be almost the same as shooting at 100 and trying to retrieve it as you said. I think there is a "sweet spot" in each camera that you need to know that your iso is totaly fine and after that point it will mess up the image in grain.
@@PeterZelinka Just tested it myself out of curiocity shooting at 400, 800, 1600 iso and the same ones with 100 pumping up the exposure in post and the difference was very notable especially at 1600 iso where the same pic at 100 iso pumped up 4 stops upwards at LR lost significant details. Can send you the samples if you want. I think I read that 5x00 nikon series was iso Invariant somewhere but turns out that this isnt the case. I might try it with extreme iso numbers sometime but i think the outcome will be as expected with the tests.
@@AlexisOGrekos Yeah, what I do is take a photo at 100, 200, 400, 800, 1600, 3200, 6400 ISO. Then open them all up in Camera RAW / Lightroom. Increase the Exposure slider for each photo to match a higher ISO. (In other words, if ISO 3200 is your benchmark, then increase the ISO 1600 by +1, the ISO 800 image by +2, ISO 400 by +3, etc...) Then open them all in Photoshop once they are the same brightness and layer each photo in the same workspace. Now you can find a spot, zoom in, and quickly see the difference between each shot. I know on my d750, when using ISO 6400 as the benchmark, ISO 100 and 200 don't look great. But 400 and up all look identical.
Sensitivity is key in understanding that. ISO 1600? No milky way. ISO 6400? Milky Way! Lol my A6300 is ISO invariant.. (between iso 0-400 and between 800 upwards) That's one crop sensor that rocks imho. Got so lucky on that; bought a compact camera for action and holiday pics. Never knew i'd be taking photos of the milky way but it turned out to be an awesome little camera for that. Too bad the crop factor decreases shutter time so badly. That's why i'd say they're inherently stupid if you're only looking to shoot the milky way. Those 10 seconds matter.
ISO is based on 'gain', not just an in camera process. Gain being more electrical current to increase the sensitivity. It's the same as an Amp, more gain = more electrical distortion (or noise)
Yeah! Exactly! This video is a bit outdated unfortunately. I need to create an updated one to clear that up. Ultimately though, the main reason people see a lot of grain in their night sky images is simply because they didn't capture enough light, not necessarily the high ISO.
I really enjoyed this video. I love the fact that you understand how ISO really works, as a lot of pros don't. Here's a quick way to reduce noise that I came up with (which is very similar to Smart Sharpen on Photoshop): While in the Detail section of Adobe Camera Raw, use the Masking slider to mask away the areas that contain noise and only sharpen the edges you need to. You may further refine that with the remaining sliders. It works wonders! Since I started using that, I've never used noise reduction because it smudges detail. Noise appears more visible when you apply edge contrast (sharpening) to it, so using the Masking slider eliminates that.
So, if I want to take nice photos of the milky way then I'm not looking for a camera that has a high ISO setting, I am looking for one that lets in more light?
@@ikannunaplays woah, whait it doesn't matter how high the iso can go the thing that matters is how well the photo will look at higher iso....and you can't say camera that lets more light. the light depends of your. Shutter speed,aperture, and iso. If you want the camera to let more light in you can but lower f stop number lens......
@@strixxx896 ISO in digital cameras is a "amplification" setting which can be done in post processing on the PC. So, you want more light vs high ISO, yes, bigger aperture. Leaving the shutter open longer is going to introduce blurring so APERATURE is the only thing that you can increase that will not decrease the clarity of the photo.
Hi Peter, thanks for the tip on the 'smart sharpen' feature in PS. Do you think the 'smart sharpen' works better than using the luminance noise correction in combination with radius & masking options in LR? I've always used LR tools and had much better luck so was curious to get your thoughts. I do agree that the best way to reduce noise is to use a star tracker and stack images. My go-to workflow now, when I have the time, is to take four 3 min exposures (so 12 min total) and stack them for best quality.
No, I just wanted to include that filter to be thorough. I normally rely on the Luminance Noise Reduction slider in Camera RAW for all of my noise reduction. Dang! That's definitely gonna create some nicely detailed, and grain-free photos! Personally, I just do a 4 minute exposure, when shooting at wide angles. That's usually long enough to produce a clean photo, without spending too much time on a single composition.
I know this doesn't apply to astrophotography, but I'm wondering if increasing the ISO improves auto focus? My guess is no. I was thinking about this because having a wider aperture and allowing more light to come in improves auto focus.
Increasing the ISO shouldn't have any impact on focusing. The only way to help improve the autofocus would be getting more light in the scene (like a headlamp or light panel). Or, finding something that has a lot of contrast
Hi Peter, excellent videos, i love them. However, i couldn't find the one about reducing amp glow you mentioned at 8:05, can you please tell the title of that video in particular? Thanks.
I ended up removing that one I believe. If you want to reduce amp glow, take a longer exposure. The more light you can capture in a single exposure, the less apparent the amp glow will be. Usually 4 minutes at f/2.8 is enough to cover up the amp glow on most cameras
@@PeterZelinka Thank you for the useful tip! Indeed the amp glow is the worst with the images i had with less exposure, when i had to stretch the image hard to get any detail. Usually with the untracked just on tripod short exposures, and stacking doesn't seem to improve it much either, maybe because i can't remove the falloff well, or my sub exposures are already not long enough, they were far too left in the histogram anyway. Thanks for the tip, you are great! :)
Excellent job Peter.. Very informative and useful. Thank you for the knowledge & inspiration, I had to subscribe to get more info and motivation.. I also subscribed to watch and see your channel grow. Keep up the great work.
You say that the best way to reduce noise is to get more light into the camera, and that means wider aperture and longer shutter speed. Well, my lense has f1.4 and it cannot get any wider than that. So my other option to increase the light would be to have longer shutter speed. Given that I buy a star tracker or I do star-guidance -- so star trail problem is out of the picture -- what then would be the best ideal shutter speed? Why I should do 1 minute longer exposure if I can do 5 hours? I guess my question is: what is the upper limit in the shutter speed.
Great question! As far as I'm concerned, a 4 minute shutter speed is ideal. The longer you go, the more chance you have for star trails. I've gotten up to 10 minutes with sharp stars with my 14-24mm lens and SkyGuider Pro. You will start to notice diminishing returns at some point, mainly depending on your camera sensor and lens' aperture. In my particular case, 4 minutes is the point where I get good results, with minimal grain. If I shoot any longer, the image won't be noticeable cleaner. Therefore, I'd just be wasting time. Also, the sensor will get warmer and warmer and could generate more hot pixels. In some cases, you might want to try for an 8 minute exposure, or maybe a 2 minute exposure. It really just depends on your polar alignment accuracy with the star tracker, camera sensor, and aperture.
so with the iso invariant sensor why would a person choose a higher iso on the camera for this type of picture when the plan is to post process anyways ? other than for test pics at the site to confirm the target is framed and focused as expected by looking at the image after it's been saved on camera? it seems like you get greater flexibility taking the original at low iso like 200 and bumping it up in post processing 1 stop at a time until what you see is what you like without the side affect of initially going with something at a higher iso like 6400 at the camera ??
Dean Johnson exactly! That's the process I follow now. Use a very high ISO to check my composition and light panel intensity, then dial the ISO to around 800 and take a very long exposure to capture enough light and not have any real grain.
thanks! I was hoping that was what you were talking about! took about 100 25 second exposures tonight to test! will look at the data tomorrow night after work! appreciate that tidbit as I hadn't seen anyone else talking about it!
thanks for posting, being a newbie, this helped a lot. I know this was about ISO, but I was wondering. Does white balance make a difference? If so do you have a video on that ? Thank you
Provided you are shooting in RAW, white balance really shouldn't matter. You can always adjust it in Camera RAW / Lightroom later with no quality loss. I personally just leave my White Balance set to Auto or Daylight, and then adjust it when I start editing.
Hello Peter, First off thank you for all of the useful information. I am new to astrophotography, and I just subscribed. I was hoping you might be able to answer one question for me around your shutter speed. I noticed you were shooting at 14mm with a 20 second exposure. Is there a reason you used that shutter speed instead of 30-35 seconds? Is it just for this demonstration or is that your preferred exposure time? Thanks! Matt
Thanks! I use a 20 second shutter speed so that I have sharp stars. The 500 Rule is largely BS and won't give good results. You can always test this on your own, take a series of images - 15, 20, 25, 30 seconds and see at what shutter speed the stars show too much motion. However, it's actually best to take an exposure that is multiple minutes long. This will give you a vastly more detailed and noise-free photo, but you will have to blend images together. I'll be covering this process, and everything else about astrophotography, in an upcoming tutorial which will have over 10 hours of content!
Thank you for getting back to me so quickly. I had heard about the rule of 500 from so many different sources that I just kind of assumed it was right. I trust your judgment though, and I will definitely go out and test a few different exposure lengths. I also just watched your iOptron SkyTracker Pro tutorial and that's a very tempting piece of equipment. I'm looking forward to your upcoming tutorial as well. Thanks again!
The SkyTracker Pro will make a massive difference in image quality, you won't have to worry about grainy images anymore! However, it definitely makes the shooting a bit more complicated. We'll have a full series of tutorials on this star tracker too!
Yes, the voltage is amplified before the signal is changed from analog to digital, if I remember correctly. But you can basically recreate the effect digitally in post processing, if you have an ISO Invariant sensor anywya
Why not stack photos in something like Sequator (free for windows)? Sure, masking foreground out can be a pain sometimes, but with stacking there's basically no need for super fast lenses, tracking mounts etc. And noise is greatly reduced. All it takes is taking 10 frames instead of 1, and then stacking the images.
Great point! Either a star tracker, or photo stacking, is definitely the best way to reduce noise. This video was originally intended to show some ways to reduce noise already present in night sky photos.
Hi just wondering when stacking the photos, say 10shots blend into 1, does each photo have to be the same settings (different ISO, shutter). I actually do not understand the idea of how stacking photo helps to reduce noice. Would love to get some explanation. Thank you
Timur , don’t agree with some of what you say. Ten , one second exposures just blow. You will not receive the data needed to truly show you the color and detail in the sky. I have used a tracker and it will clearly show the data I am speaking of, more color, more resolve in the pic.
Other point to make, the LESS time your sensor (to achieve needed data) has that current running across it, heating it up and creating noise in your pic, is the better for sure.
I found a potentially useful tip, here, on RUclips, relating to the ISO and reducing noise, but since I'm not a seasoned nighttime photographer as you are I cannot attest to its validity. The video in question implies overexposing the ISO by 2-3 stops and reducing the exposure by the same amount in post-processing to decrease the noise. Video link ruclips.net/video/X0WbpelfyaM/видео.html at 4:06 minute mark. I hope this helps.
Would you be willing to receive an email from me with a few of the pictures I’ve taken? They’re awful and I need help. I can explain more in the email. So frustrating.
If you have an ISO Invariant camera, then yes. You can intentionally use a lower ISO, then increase the Exposure slider in Camera RAW or Lightroom to brighten the image as needed. However, if you have an ISO Variant camera, you want to use the "correct" ISO for a good exposure. If you have to increase the Exposure or Shadows in post-processing, you'll notice pretty severe quality loss.
Not exactly: most modern sensors have a dual gain amplifier. For example the Nikon Z7 has a base ISO of 64 and the dual gain kicks in at ISO 400. You can see this e.g. in the dynamic range plot on DxOMark. So you should in this case either stay at ISO 64/100 or go directly to ISO 400 if you want to push the exposure anyway. In short: don't you ISO 320 in this case. Every sensor might have some different transition values for the secondary amplification, for Nikon Z6 it's ISO 800
Excellent video. I have been an amateur astronomy for 60 years and I am learning a lot about the new technologies from your videos. I belonged to the Von Braun astronomical society in the 1970's when we had a 21 inch Cassegrain I would enjoy using and making photos with my Nikkormat camera on Tri-X film and then developing and printing - many did not turn so well as they do today with the digital cameras.... I don't know what they have now as I left Huntsville in 1976 for WSMR in New Mexico. The new star trackers are going to allow me to continue to take my camera and lens to the desert as the big telescope is getting too heavy. Thanks for your videos. Please keep up the good work.
Best iso concept explanation in youtube. Could not understand iso until you explained how the camera handles it.
Thank you for showing how ISO really works with the camera. It is only brightening the images, not taking in more light. Wow!
Presently, I now only use a high ISO to find stars in the city, then lower it afterwards in frame to reduce glare.
I use a Pentax which has built in astro tracker and it surprisingly works fairly well,It uses it's in body image stabilization to physically track the stars as they move.
I have done a full minute exposure at iso 200 and f5 for a bit of sharpness and works awesome.
Hi there, I have a question. I get a lot of noise from my 18-55mm kit lens @ ISO 6400 (18mm), 15 sec. exposure and f3.5 and this setting is the one I used to make the Milky Way details appear on my camera (Nikon D3400). I've purchased the Tokina 14-20mm f2 PRO and I am still waiting for it to arrive. Now, my question is will I get less noise from the Tokina lens from its wide aperture and longer exposure (since I'd be able to use 25 sec. exposure on it @ 14mm and at f2 aperture) while still using the ISO 6400? I'm into Milky Way timelapse and the noise on my kit lens is just horrible even after post-processing :(.
Thank you and cheers!
Nah, with f2 u will let more light into camera, u can easily use max iso1600, f1,8-2 more light f3'5>32 less light
Forgive my newbie ignorance. Is this what you're saying? Altering the ISO on a digital camera is simply amplifying the light (the signal from the sensor). It doesn't allow the camera to magically collect more light. So there is no point in doing this as photoshop will do the same thing and possibly with less noise. The only way to increase the light is to increase the exposure by using a tracker and or a faster lens. To me this makes perfect sense. Many experts tell you to crank up the ISO. And, stacking doesn't magically make dim stars brighter either, it only reduces noise, as noise is random and the stars aren't. However, if you do use long exposures etc, aren't you simply increasing the brightness of everything, dim and bright stars together, and running the risk of over exposing everything. Again, apologies if this is in actual fact rubbish. PS. Excellent vids and web site. So glad I stumbled upon you.
Pretty much, yeah! However, there is an important distinction to make. I recommend doing some tests with your own camera to verify what I'm saying.
What you should do is go out one night with your camera. Take a 20 second exposure at f/4, ISO 100. Then, increase the ISO to 200 and take another photo. Continue with 400, 800, 1600, 3200, and ISO 6400. Be sure to leave the shutter speed and aperture the same. You'll also want to be shooting in RAW.
Once you have these photos, with the ISO increasing 1 Stop each time, you can test your cameras sensor. Load up all of your images in Adobe Camera RAW at the same time. Then, increase Exposure slider by 5 Stops for the ISO 100 image. You can now directly compare it with the ISO 3200 image. If the ISO 100 image brightened by 5 Stops in Camera RAW looks terrible, with odd colors and grain everywhere, your camera sensor is ISO Variant.
You can do this for any of your images. For example, take your ISO 400 image, increase the Exposure slider to 4, and compare it directly against the ISO 6400 image. If your camera is ISO Invariant, there should be no discernible difference between them.
It's up to everyone to test their own camera sensors to find out whether it's ISO Variant or Invariant. Again, all that means is that with an ISO Invariant sensor you can increase the Exposure in post-processing, with no quality loss. If you try to increase the Exposure slider with an ISO Variant camera, your image will look terrible.
If you have an ISO Variant sensor, you'll want to use a higher ISO in-camera, to get the best results. If you have an ISO Invariant sensor, you can pretty much use any ISO you want, and increase the Exposure as needed in Post-processing.
If your camera isn't very new, it likely won't be ISO invariant. However to estimate the best ISO for old digital cameras, you can check out the ISO-Dynamic Range curves. Above a certain ISO threshold, dynamic range decreases linearly with increasing ISO. Try out the 2 ISO values where the change in slope occurs, the one above and below it. They should be the best performer ISO for old digital cameras. Anything above those would probably increase the noise to uncessary amounts and decrease the dynamic range too much, so it's not the best idea to use the maximum iso for old cameras.
Keep in mind that lower iso will yield higher read noise to your images, which in return, unless you subtract them with bias (or dark flats not sure about the name) calibration frames, they will end up in your images. Also higher iso images are more likely to show thermal banding if you expose too long or take too many exposures, which is even worse in most cases.
Too high ISO will also decrease your dynamic range which in return will yield overall poor image quality. So finding the optimum is important for old cameras.
Super job, thanks for clarification on the fact to the that ISO means little for end results.... the formula is .......
available light, aperature, and exposure time.... period.... post handles the rest...
idrivavipr yep! More or less anyway.
Choosing the wrong ISO can have a negative impact, if your camera sensor isn't great. For example, some cameras have badly ISO variant sensors. Therefore, if you try to increase the Exposure by more than 1 Stop in Camera RAW or Lightroom, you will experience severe image degredation. This will mainly manifest as purple noise in the shadows of the image.
Then again, if you would just take a longer shutter speed (2+ minutes), that thermal noise would not be an issue because you captured enough signal (light) that the noise is no longer visible.
A slight digression. I have an Olympus Em5-ii which is classed as "semi-invariant". The problem with Photoshop is that is is not as good at processing the image as the in-camera "engine". Compared to the Panasonic GH4 and G7, the E-M5ii image gets more noise at high ISO. The Panasonic sensor is mounted on a heat sink which reduces sensor noise, the Olympus Body Image Stabilisation preludes fitting the sensor to a heat sink. Both Panasonic and Olympus use a Live MOS (LMOS) sensor as opposed to Complimentary MOS (CMOS), so the sensor is biased to sweep stray electrons from the sensor, which means less noise. The GH4 can re-process an image in-camera, and does a far better job of it than Photoshop does. All three cameras have an Auto-ISO even in Manual mode. It can be a case of trust the camera.
Because most monitors are only 8 bit (per colour channel) when comparing photos (recorded in 14 bit RAW) taken at different ISO's, with exposure boosted in post processing you may not be able to see the difference.
Depending on the bit-depth of the image sensors analog to digital converter (ADC), the RAW file of the ISO 100 image could contain 32 x less "brightness resolution" than the RAW file at ISO 3200 even with ISO invariant sensor (in other words the 14 bit RAW file of the ISO 100 image would only contains 9 bit of useful information).
So when the ISO 100 image is boosted by 5 stops and displayed along side the ISO 3200 image, they will look the same on the 8 bit monitor, even though the ISO 100 image actually has 32 x less "brightness resolution"
Thanks a ton for sharing the concept of 'ISO Invariance'. This is a gem.
Sir, I am very beginner photographer from a small village of india and i am using very basic camera - Nikon Coolpix P530 Which have certain limitations. I want to photograph milky way and some stars. But as i said my camera have certain limitations that my camera have lowest shutter speed of 15" and the lowest aperture i can set is 3.0. But when i set my camera on 15" shutter speed and f. 3.0. The camera locks ISO on 100 and when i try to increase ISO on 1600 the shutter speed is automatically locks on 1". It seems my camera is allowing me to take pictures only on ISO 100 when shutter speed is settled on 15". So my question is, can i shoot milky way and stars? Can i use the Photoshop technique that you showed in your video? By increasing the ISO by stop method in Photoshop?? Please help me sir. Please reply me.
I understand the concept of reducing noise is to collect more light but doesn't long exposure times increase noise as well? Is the long exposure noise reduction option built into my camera sufficient for battling that particular type of noise?
A longer exposure will increase the amount of hot pixels in the image. You can turn on Long Exposure Noise Reduction in-camera to eliminate them completely.
Therefore, a long exposure with LENR turned on tends to give the best results. You'll just need to wait longer for the actual image to process.
You should 've been compared the 2 images with 1:1 zoom to really see that they were identical, because from my experience iso isn't just like pumping the exposure up, but making the sensor's pixels more sensitive. I've seen tremendous difference between images myself. Just to clarify it depends on the quality of your camera too of course, crop sensors usually really mess it up from 1600+ iso while some full frame cams can handle way up to the scale like mofos without particular noise.
It's really all about the camera sensor, and if it's ISO Invariant or not. Also, it depends how much light you capture.
The less light you capture, the more grain you'll have. If your camera is ISO Variant, you'll have a lot of problems when you increase the Exposure slider in post.
However, if you have an ISO Invariant sensor, then you can pretty much use whatever ISO you want, and increase the Exposure slider in post without any quality loss. The main reason you'd intentionally use a lower ISO is to preserve details in the highlights
@@PeterZelinka It definitely depends on the camera sensor. I think the grain of the noise is being produced by the poor quality of the sensor's pixels because they will make grain as a response of thinking there is light coming. I have a d5300 and if i bump up the iso to 400 I will be totaly okay, but if I lack that light shooting at just 100 iso and try to retrieve it from LR I will get much more noise. In opposite, if I shoot with something extreme like 3200 iso it will be almost the same as shooting at 100 and trying to retrieve it as you said. I think there is a "sweet spot" in each camera that you need to know that your iso is totaly fine and after that point it will mess up the image in grain.
@@PeterZelinka Just tested it myself out of curiocity shooting at 400, 800, 1600 iso and the same ones with 100 pumping up the exposure in post and the difference was very notable especially at 1600 iso where the same pic at 100 iso pumped up 4 stops upwards at LR lost significant details. Can send you the samples if you want. I think I read that 5x00 nikon series was iso Invariant somewhere but turns out that this isnt the case. I might try it with extreme iso numbers sometime but i think the outcome will be as expected with the tests.
@@AlexisOGrekos Yeah, what I do is take a photo at 100, 200, 400, 800, 1600, 3200, 6400 ISO. Then open them all up in Camera RAW / Lightroom. Increase the Exposure slider for each photo to match a higher ISO. (In other words, if ISO 3200 is your benchmark, then increase the ISO 1600 by +1, the ISO 800 image by +2, ISO 400 by +3, etc...)
Then open them all in Photoshop once they are the same brightness and layer each photo in the same workspace. Now you can find a spot, zoom in, and quickly see the difference between each shot.
I know on my d750, when using ISO 6400 as the benchmark, ISO 100 and 200 don't look great. But 400 and up all look identical.
Sensitivity is key in understanding that. ISO 1600? No milky way. ISO 6400? Milky Way!
Lol my A6300 is ISO invariant.. (between iso 0-400 and between 800 upwards) That's one crop sensor that rocks imho. Got so lucky on that; bought a compact camera for action and holiday pics. Never knew i'd be taking photos of the milky way but it turned out to be an awesome little camera for that. Too bad the crop factor decreases shutter time so badly. That's why i'd say they're inherently stupid if you're only looking to shoot the milky way. Those 10 seconds matter.
ISO is based on 'gain', not just an in camera process. Gain being more electrical current to increase the sensitivity. It's the same as an Amp, more gain = more electrical distortion (or noise)
Yeah! Exactly! This video is a bit outdated unfortunately. I need to create an updated one to clear that up. Ultimately though, the main reason people see a lot of grain in their night sky images is simply because they didn't capture enough light, not necessarily the high ISO.
I really enjoyed this video. I love the fact that you understand how ISO really works, as a lot of pros don't. Here's a quick way to reduce noise that I came up with (which is very similar to Smart Sharpen on Photoshop): While in the Detail section of Adobe Camera Raw, use the Masking slider to mask away the areas that contain noise and only sharpen the edges you need to. You may further refine that with the remaining sliders. It works wonders! Since I started using that, I've never used noise reduction because it smudges detail. Noise appears more visible when you apply edge contrast (sharpening) to it, so using the Masking slider eliminates that.
So, if I want to take nice photos of the milky way then I'm not looking for a camera that has a high ISO setting, I am looking for one that lets in more light?
@@ikannunaplays woah, whait it doesn't matter how high the iso can go the thing that matters is how well the photo will look at higher iso....and you can't say camera that lets more light. the light depends of your. Shutter speed,aperture, and iso. If you want the camera to let more light in you can but lower f stop number lens......
@@strixxx896 ISO in digital cameras is a "amplification" setting which can be done in post processing on the PC. So, you want more light vs high ISO, yes, bigger aperture. Leaving the shutter open longer is going to introduce blurring so APERATURE is the only thing that you can increase that will not decrease the clarity of the photo.
Hi Peter, thanks for the tip on the 'smart sharpen' feature in PS. Do you think the 'smart sharpen' works better than using the luminance noise correction in combination with radius & masking options in LR? I've always used LR tools and had much better luck so was curious to get your thoughts. I do agree that the best way to reduce noise is to use a star tracker and stack images. My go-to workflow now, when I have the time, is to take four 3 min exposures (so 12 min total) and stack them for best quality.
No, I just wanted to include that filter to be thorough. I normally rely on the Luminance Noise Reduction slider in Camera RAW for all of my noise reduction.
Dang! That's definitely gonna create some nicely detailed, and grain-free photos! Personally, I just do a 4 minute exposure, when shooting at wide angles. That's usually long enough to produce a clean photo, without spending too much time on a single composition.
I know this doesn't apply to astrophotography, but I'm wondering if increasing the ISO improves auto focus? My guess is no. I was thinking about this because having a wider aperture and allowing more light to come in improves auto focus.
Increasing the ISO shouldn't have any impact on focusing. The only way to help improve the autofocus would be getting more light in the scene (like a headlamp or light panel). Or, finding something that has a lot of contrast
Hi Peter, excellent videos, i love them. However, i couldn't find the one about reducing amp glow you mentioned at 8:05, can you please tell the title of that video in particular? Thanks.
I ended up removing that one I believe.
If you want to reduce amp glow, take a longer exposure. The more light you can capture in a single exposure, the less apparent the amp glow will be. Usually 4 minutes at f/2.8 is enough to cover up the amp glow on most cameras
@@PeterZelinka Thank you for the useful tip! Indeed the amp glow is the worst with the images i had with less exposure, when i had to stretch the image hard to get any detail. Usually with the untracked just on tripod short exposures, and stacking doesn't seem to improve it much either, maybe because i can't remove the falloff well, or my sub exposures are already not long enough, they were far too left in the histogram anyway. Thanks for the tip, you are great! :)
Excellent job Peter.. Very informative and useful. Thank you for the knowledge & inspiration,
I had to subscribe to get more info and motivation.. I also subscribed to watch and see your channel grow. Keep up the great work.
You say that the best way to reduce noise is to get more light into the camera, and that means wider aperture and longer shutter speed. Well, my lense has f1.4 and it cannot get any wider than that. So my other option to increase the light would be to have longer shutter speed. Given that I buy a star tracker or I do star-guidance -- so star trail problem is out of the picture -- what then would be the best ideal shutter speed? Why I should do 1 minute longer exposure if I can do 5 hours? I guess my question is: what is the upper limit in the shutter speed.
Great question! As far as I'm concerned, a 4 minute shutter speed is ideal. The longer you go, the more chance you have for star trails. I've gotten up to 10 minutes with sharp stars with my 14-24mm lens and SkyGuider Pro.
You will start to notice diminishing returns at some point, mainly depending on your camera sensor and lens' aperture. In my particular case, 4 minutes is the point where I get good results, with minimal grain. If I shoot any longer, the image won't be noticeable cleaner. Therefore, I'd just be wasting time. Also, the sensor will get warmer and warmer and could generate more hot pixels.
In some cases, you might want to try for an 8 minute exposure, or maybe a 2 minute exposure. It really just depends on your polar alignment accuracy with the star tracker, camera sensor, and aperture.
Whoa. You just blew my mind about ISO!
so with the iso invariant sensor why would a person choose a higher iso on the camera for this type of picture when the plan is to post process anyways ? other than for test pics at the site to confirm the target is framed and focused as expected by looking at the image after it's been saved on camera? it seems like you get greater flexibility taking the original at low iso like 200 and bumping it up in post processing 1 stop at a time until what you see is what you like without the side affect of initially going with something at a higher iso like 6400 at the camera ??
Dean Johnson exactly!
That's the process I follow now.
Use a very high ISO to check my composition and light panel intensity, then dial the ISO to around 800 and take a very long exposure to capture enough light and not have any real grain.
thanks! I was hoping that was what you were talking about! took about 100 25 second exposures tonight to test! will look at the data tomorrow night after work! appreciate that tidbit as I hadn't seen anyone else talking about it!
thanks for posting, being a newbie, this helped a lot. I know this was about ISO, but I was wondering. Does white balance make a difference? If so do you have a video on that ? Thank you
Provided you are shooting in RAW, white balance really shouldn't matter. You can always adjust it in Camera RAW / Lightroom later with no quality loss. I personally just leave my White Balance set to Auto or Daylight, and then adjust it when I start editing.
Hello Peter,
First off thank you for all of the useful information. I am new to astrophotography, and I just subscribed. I was hoping you might be able to answer one question for me around your shutter speed. I noticed you were shooting at 14mm with a 20 second exposure. Is there a reason you used that shutter speed instead of 30-35 seconds? Is it just for this demonstration or is that your preferred exposure time?
Thanks!
Matt
Thanks! I use a 20 second shutter speed so that I have sharp stars. The 500 Rule is largely BS and won't give good results. You can always test this on your own, take a series of images - 15, 20, 25, 30 seconds and see at what shutter speed the stars show too much motion.
However, it's actually best to take an exposure that is multiple minutes long. This will give you a vastly more detailed and noise-free photo, but you will have to blend images together. I'll be covering this process, and everything else about astrophotography, in an upcoming tutorial which will have over 10 hours of content!
Thank you for getting back to me so quickly. I had heard about the rule of 500 from so many different sources that I just kind of assumed it was right. I trust your judgment though, and I will definitely go out and test a few different exposure lengths.
I also just watched your iOptron SkyTracker Pro tutorial and that's a very tempting piece of equipment. I'm looking forward to your upcoming tutorial as well.
Thanks again!
The SkyTracker Pro will make a massive difference in image quality, you won't have to worry about grainy images anymore! However, it definitely makes the shooting a bit more complicated. We'll have a full series of tutorials on this star tracker too!
But isn’t the electrical amplification on the sensor actually what’s happening when the ISO is cranked-up in value?
Yes, the voltage is amplified before the signal is changed from analog to digital, if I remember correctly. But you can basically recreate the effect digitally in post processing, if you have an ISO Invariant sensor anywya
You can get longer Shutter speed by decreasing focal length with the same aperture setting. That what 500rule says
Does it work for video?
Why not stack photos in something like Sequator (free for windows)? Sure, masking foreground out can be a pain sometimes, but with stacking there's basically no need for super fast lenses, tracking mounts etc. And noise is greatly reduced. All it takes is taking 10 frames instead of 1, and then stacking the images.
Great point! Either a star tracker, or photo stacking, is definitely the best way to reduce noise. This video was originally intended to show some ways to reduce noise already present in night sky photos.
Hi just wondering when stacking the photos, say 10shots blend into 1, does each photo have to be the same settings (different ISO, shutter). I actually do not understand the idea of how stacking photo helps to reduce noice. Would love to get some explanation. Thank you
Timur , don’t agree with some of what you say. Ten , one second exposures just blow. You will not receive the data needed to truly show you the color and detail in the sky. I have used a tracker and it will clearly show the data I am speaking of, more color, more resolve in the pic.
Other point to make, the LESS time your sensor (to achieve needed data) has that current running across it, heating it up and creating noise in your pic, is the better for sure.
Wow, amazing learned a lot. Thank you.
I found a potentially useful tip, here, on RUclips, relating to the ISO and reducing noise, but since I'm not a seasoned nighttime photographer as you are I cannot attest to its validity. The video in question implies overexposing the ISO by 2-3 stops and reducing the exposure by the same amount in post-processing to decrease the noise. Video link ruclips.net/video/X0WbpelfyaM/видео.html at 4:06 minute mark. I hope this helps.
I Think I Will prefer the stack method
This video is old, but your point works only because you have an iso-invariant sensor. Use a Canon, and then you realize this is not that simple.
Would you be willing to receive an email from me with a few of the pictures I’ve taken? They’re awful and I need help. I can explain more in the email. So frustrating.
Sure thing! You can find my email over on my website:
www.peterzelinka.com
Jak se mas?
ISO is cameras sensor sensitivity to light. The higher, the more sensitive to light.
nikons iso is not good
So actually ISO is unnessecary
If you have an ISO Invariant camera, then yes. You can intentionally use a lower ISO, then increase the Exposure slider in Camera RAW or Lightroom to brighten the image as needed.
However, if you have an ISO Variant camera, you want to use the "correct" ISO for a good exposure. If you have to increase the Exposure or Shadows in post-processing, you'll notice pretty severe quality loss.
Not exactly: most modern sensors have a dual gain amplifier. For example the Nikon Z7 has a base ISO of 64 and the dual gain kicks in at ISO 400. You can see this e.g. in the dynamic range plot on DxOMark.
So you should in this case either stay at ISO 64/100 or go directly to ISO 400 if you want to push the exposure anyway. In short: don't you ISO 320 in this case.
Every sensor might have some different transition values for the secondary amplification, for Nikon Z6 it's ISO 800
Video quality is so bad
It is a RUclips setting. The highest is 720p on RUclips. It defaulted to 320p. No one that liked your comment knew this?