Well done Alyn! I'm glad you did this video so I didn't have to 😜 You did a very nice job of explaining the major concepts in plain language and how they relate to astrophotography. I think your lead-up explanation of exposure stops and then the visual difference between ISO100 pushed 5 stops vs. ISO3200 at 04:51 is especially effective, and will hopefully get more people thinking about this in the correct way, and not just insisting to me that it's as simple as: higher ISO = higher noise. Also, that's an epic filming location for your talking head stuff!
I do remeber Tony Northup had a video about this a while back, but he didn't give the astro take on it which is what most of us want. Nico your last video on the ISO was very helpful also.
Hey man, it was one of your videos where I first learned about this. I think it was a video about andromeda galaxy. You put up a link to a website which had a drop down menu list for pretty much every camera model and it showed on a graph all the jumps of the iso and where it levelled off. It was a great resource for anyone who wants to find out their best iso for low light without doing all the work. Great video all round too
You will continue to produce information for all of us even in your absence. I hope that those stars swallowed you up and we can all continue to hear you speak through them.
17:40 - Sony a7iii has a function called "Bright Monitoring" that enables you to see in the dark before the shooting without changing the ISO. Just assign a custon button. Thank you for the video!
Yes bright monitoring is great, I've mentioned it in a few vids! But at that moment I was talking more about the impracticality of being able to preview your RAW files. Yes you can compose with Bright Monitoring and you can take your images at 640 but what happens when you get home, boost the exposure and then there's all sorts of rubbish and things in the composition you didn't notice
Good video, I just have two comments on it: 1. You mention that higher ISO means lower dynamic range. While true, it's important to realize that simulating higher ISO by underexposing and correcting in postproces does exactly the same. When you underexpose by e.g. 3 EV, it means that you only use 1/8 of the available range, i.e. lose 3 bits (or 3 EV) of dynamic range. And when you check the graphs of dynamic range by ISO, in the intervals of ISO invariance you can see exactly 1 bit (or 1 EV) of loss for each ISO step. Which is not really a surprise, considering that in these intervals you essentially just choose if it's camera processing chips or your software in PC who does the multiplication or division, the values produced by the sensor are the same. In other words, dynamic range is not really an argument for underexposing and correcting in postprocess, the only reason to do that is to prevent clipping. 2. People often tend to think that an ISO invariant camera automatically means it's better. This is not always true, all it means is that the sensor has only one actual gain value and every other ISO setting is simulated by multiplying or dividing the values from the sensor by a constant - i.e. exactly what the exposure correction in postprocess does. While more convenient, it does not always mean a multi-gain sensor could not achieve better results (in the sense of less noisy image) for a particular combination of exposure settings. Often the reason for using a multi-gain sensor is to get better results in some parts of the range, it would be easier and cheaper to use just one gain and simulate everything else by recalculating in firmware.
Finally!.. This is the first video/article I've found which explains it to my satisfaction, as a photographer (and as a Sony shooter). Seems like most people seem to get bogged down in more of the technical aspects of it all, causing me to involuntarily tune out. You spoke to us as a photographer, rather than an engineer, and I followed you from beginning to end. Thank you.
Good explanation, in fact the best I've seen. As I use a Fuji XT-2 I've come to set the ISO at 1600 for everything and adjust in post with the exposure module (Darktable). Works great. I'm pretty sure that Sony developed this technology and Fuji uses the Sony sensors (although they don't like to admit it). One additional point - while the noise stays the same the dynamic range does not. As you increase exposure (in post) the dynamic range decreases. (You covered this later in the video!).
I couldn't agree more. It is exactly how I choose my iso. I just add that I reduce the iso if I am doing a single shot or stack images. If it is a panorama I reduce too but keep an iso where I still can see the composition in each shot.
Total noob here! Thanks, I've never come across this topic regarding ISO before. All I have interpreted prior was the higher the ISO, the more noise introduced. Always keep it as low as possible. Thanks to you, not only do I realize that is not the case, I also know how to figure out what is likely to suit my camera best. Super job of conveying a complex perspective so I could understand. Worthy of a 3rd "Thanks!" for this video.
Good job explaining this complicated topic! As a long time user of the Fuji X-T1 I can confirm that the camera is pretty much ISO invariant, so much so that I was taking milky way images at iso200 at one point to not blow out the highlights in a campfire. As it turned out, I unfortunately overestimated the performance of the sensor and the darkest regions of the image resolved nearly no detail compared to a ISO1600 exposure. So even with the X-T1 there seems to be a limit of how strong you can underexpose. But in general its still pretty amazing how much freedom you have with the X-T1s raw files.
I've been trying to get my head around ISO invariance for months but couldn't grasp it. That was until I just saw this video which explains it perfectly. Thank you. i also own an A7iii (bought after watching many of your videos). It's nice to know I have the tools, if not yet the skills, to attempt astro photos like yours.
Well shit, I have some homework to do this weekend. I knew about this issue but I tend to go to iso 100 every time and now it seems that may not be ideal. Thanks for this Alyn!
Wow! This is incredibly useful information. I've gone all this time assuming ISO100 was always best when possible. I am eager to do this test with my camera and learn at what point it truly performs best.
Thank you so much Alyn! Your repetition is perfect for teaching new concepts. I learned so much from this video - even what a full stop is pertaining to ISO.
Many thanks for this Alyn - I've learnt something new today! And the best news for me - my camera is an XT-1 😁. So I never need to worry about ISO ever again - Auto ISO for me from now on 🥳🎉
Great explanation. As a software developer who worked with computational photography, iso invariance is a very hard problem to solve. Ideally a short timelapse helps solve the iso issues. Sadly you do get star trailing if done incorrectly.
Well, it's night time and a near full moon, I can't wait to use this info to test out my Nikon Z6! I have been defaulting to ISO4000 for my Milky Way shots for the last year.
Nora by Space Light! that explains a lot and even simple enough for Noggin here to understand! Thanks so much Alyn - certainly wasn't aware of this and very much appreciated!
Great Video as Always! Very informative! In your video you mentioned you could raise your iso to it's highest setting to compose then lower it back down to the minimum invariant iso.. what I do on all my Sony cameras is turn off the live view setting effect.. to compose.. then turn it back on for the shot.. I have setting effect on/off as a toggle on one of the programmable buttons. If you know your manual settings are correct and you are going to shoot all low light.. you can leave live view setting effect off the whole time. Works for me.. your mileage may vary.. lol.
Your awesome alyn I have learned more from you than i have from 3 years of my own experience I'm happy there are people like you that would like to share their information instead of hiding it Thank you alyn Thank you
Alyn, This is an excellent video and introduced me to something I was completely unaware of. I'm just starting out with Milky Way photography (I love it) and was under the impression for shooting foregrounds/dark parts of my image to shoot a longer exposure at lower ISO to try to keep the noise down, but I wasn't having success at all and my low ISO images were very noisy. This completely explains what is going on and that I should not be afraid to shoot much higher ISO to capture information in the shadows. I'm shooting with a Canon R5 and ran this test myself and found that between ISO 800 and 6400 my camera is ISO invariant which is a huge bit of knowledge for me and should greatly improve my images.
They never ever did. Higher ISO always lead to dropping read-noise. After the point, were higher ISO stop the decreasing of read-noise is called "ISO-invariant". This was (not only) described in 2010 within the canon-cosmos.
For those of you using the Canon Rebel T7/2000D/1500D/Kiss X90 (same camera, different names). I tested the noise performance at every ISO level. I heard it is the best-selling camera on the market right now so I thought it'd be useful to share the results. All the shots were taken in a relatively dark room at night. The images were shot at 50mm F/1.8 and 1/30s shutter. The images were matched to the ISO 6400 shot in RawTherapee 5.8 with the neutral processing profile. So here are the results: ISO 6400: Reference image, and highest ISO the camera will go with stock software ISO 3200: Identical noise performance to 6400 ISO 1600: Performace is very, very, very slightly worse. As an analogy, it would be comparable to measuring the height of your room floor to ceiling and being off by a millimeter or two. It's only noticeable if you're specifically looking for it and then you really have to be looking for it. This seems to be the ideal ISO as far as I can tell. ISO 800: Still okay. If you can afford to take some more exposures for stacking, you'll get away with it, but I still recommend going 1600 ISO 400: Only use if you really have no other options and you're fine with a very sub-optimal end result. ISO 200: Just stay away from this one. ISO 100: Don't even think about it. TL;DR: Use ISO 1600 for very slightly better dynamic range, use ISO 3200 for very slightly better noise performance, ISO 800 is mweh, but okay. Hopefully, that'll save you some effort :)
Oh man! I just bought a Fuji X-T1. I've heard of ISO invariance in the past but never thought I'd get to play with one! That's awesome! Great video! Cheers from Costa Rica
Thank you for explaining that concept very clearly ! Will use the 640 ISO on my A7III for astrophotography, low light scenes and day to night time-lapse from now. Nice to know there's always a way to improve our skills and image quality 😉
Excellent vlog! 👌👍💪 Goes well with the one by Tony Northrup where he demonstrated that noise is due to lack of light (wrong aperture/shutter speed) not a high ISO setting.
This is very good. I knew very well before watching the video what ISO invariance is, but I still believe this might just be the best explanation of the topic.🙏🏼 Thank you.
my 6dii was not quite what I wanted and when I switched to xt3 I was way over pleased even after coming down to APSC, For the life in me I couldn't figure out why. Now I know! Perhaps I needed more finesse in my skill but only if I had known, so Thank you!
Excellent video Alex. Had the Fuji XT1 which I used for a few Milky Way shots. Was always impressed with its noise performance, especially when boosting exposure in post.
Thanks, Alyn, spot on! First times out with my Nikon D7500 I was afraid to shoot over 800 because that is what people were saying, and still are. I have since started shooting by looking at the histogram to move the black up so they are not clipping and the whites so in a log-scale are not blown out on the right side, using whatever ISO gets me there. That would make for another great explanation from You or Nico. Daytime photographers talk about it all the time, but not hardly a word from astrophotographers as to what the histogram would typically look like for great photos.
Thanks Alyn. I've been tempted to switch back to Canon from my Sony A7 and this has given me second thought. Maybe I'll have both for a while and do some experiments :D
Complicated...but i got it. I'll test out my camera... in the past, I just used what looked good in my viewfinder...now I know to take test shots so i can check focus and my framing and then pick the best iso for my camera for post processing and final image quality!
It’s easy to blow highlights at each base ISO start (100 and 640 for my Sony A7R3) because all the DR is in the shadows. To not blow out clouds when they are backlit, it’s better to shoot the highest ISO you can get away with up to and including 500. Check the exposure by spot metering on the clouds at +2.
Thanks Alyn for this detailed Video, this gives me a better understanding of my A7 III while out in the field for NightSky photography. Great content as usual !!
Agree with you on the noise and on better dynamic range, but over exposing a RAW file in post over +2 EV tends to show discoloration and/or artifacts depending on what image editing software you're using. I like to play it safe by shooting a well exposed image the scene dictates, but will give it a try in a high dynamic range situation and see. Great video. Clear skies!
this was really interesting, and I'm surprised to see my Nikon D5600 seems to have the same noise after ISO 400. Surprising for such an entry-level camera. This helped me understand ISO much better. Until now I thought ISO was pretty pointless and was basically the same as the exposure slider in lightroom. After watching this video I guess I was right, but now I know why! and I also know that won't be the case for all cameras. Also didn't know ISO effected dynamic range, that'll be super handy in future!
Excellent video as always Alyn! As a Canon R5 owner I have to point out that the R5 has a dual gain sensor, ISO invariant from ISO 400 up. Folks can read about this and compare to other cameras (yes, the a7R IV performs better in this regard!) in the - excellent as always - review of the R5 at DPReview.
brilliant , well done sir . learned some more , as 15 y photographer . And now go to sleep you look a bit exhausted , althought sounded very clear . grtz Geerts Johny
That was a fun little experiment! I was not expecting the drastic color cast at the lower ISOs though. Proved very helpful in understanding the capabilities of my camera better. Thanks.
Very informative & interesting Alyn. Another consideration for me is megapixels. I shoot a lot of timelapse’s & have been using the Sony A7R2 & A7r4. I recently got the A7s3 & have found for 4k timelapses it is so much cleaner doing a day to night timelapse (using aperture priority & iso minimum shutter speed). The quality of the finished sequence at 12mp is, at 4k much better/sharper/cleaner than the 42mp & 61mp sequences. Also by limiting the shutter speed & letting the iso ramp up, I don’t get that annoying speed up effect that I have to correct in post. 🥰👍🙏😀Thankyou for producing great content & I very much appreciate the very informative videos that you produce for us. Cheers!
thanks for that mate , now I can save the moon in my all night long time lapses with a moonrise . I always struggle with that inherently Good On Ya!!!!!!
With aurora photos I have found it more practical to keep ISO in 1600-5000 range to be sure that the exposure (judged from the histogram of the captured image) is exactly where I want is to be so that the adjustments in post are minimized. This because intensity of the auroras and the ideal exposure is constantly changing and I can't evaluate it from the live view (especially when you are in the dark staring at a led display which is showing an underexposed image...). The only way to be sure about the exposure is the histogram of a captured image (at least with Sony a7iii where in case of a long exposure the live view image and its histogram are not anymore wysiwyg).
A good tool I've seen Nico Carver from Nebula Photos use was the website called "Photons To Photos". They have an input-referred read noise chart for a bunch of cameras which shows the levels of noise depending on the ISO settings.
Awesome mate. Explained it very well. Would love a video of your post production methods when shooting at iso640 and how you boost the exposure back to get the correct exposure for the shot. Assuming you don’t do a global exposure slider adjustment if you have objects like the moon and street lights? Is your post done in photoshop?
I quickly googled to see which cameras are iso invariant only to discover that my nikon d5500 is among them. This was a surprise and makes me wanna test what you just told us. I am not delusional of course, it cannot be compared to a full frame camera but it is certainly a good feature to have.
@@caviestcaveman8691 my limit is 1600 for acceptable photos for this camera. I rarely will go more than that. However I have yet to try the new denoise programs. I would like to try using them on older photos with higher iso settings buy haven’t managed to do so yet.
Your videos are always interesting, even if I can't use your explanations for making astrophotography, it is interesting and fun to watch them because it makes me understand the process and the nature of a made photo as opposed to seeing it in real yourself. I have some experience with other photos though. ❤👍🏻
So, Im mostly in dark scenes ... using iso 3200/6400 with flash on body to get correct exposure and balanced ambient and flash light. I was so worried about ISO, this makes me relieved about using even iso 9000/12800 on my D750 if I need to, because basicly....software increases brightnes above ISO200 the same way as I would in LR. 😎 This is amazing, thanks 😁
So do I understand correctly? With a iso variant camera, increasing iso is truly raises the signal compared to the noise level making the SNR better up until a point where it stops and at least not getting worse even with crazy high values. And with an invariant camera you can choose whatever iso you want it won't change the Signal compared to your noise level. So actually raising your iso doesn't make your image more noisy in either case it gets noisy because your exposure time and aperture is not enough to increase your signal out of the noise level. Thus the only thing is lost with higher iso is dynamic range and iso variant cameras are somewhat forced to lose some of it to have less noise.
You first need to understand that noise is introduced into the image before ISO amplification is applied and afterwards. ISO invariant cameras introduce very little noise after ISO amplifcation. So ISO setting makes little difference to the signal to noise ratio. ISO-variant cameras introduce more noise after ISO amplification. So if you boost the signal before that noise is introduced, the signal to noise ratio improves (but you're not collecting more signal). Check out the technical explanation video linked in the comments
There is a fault in your real-life-nighttime-test: You will have to exposure the 1600 well. But in your test you underexposured the 1600. It is importend that the base-image for that test is right lit, or in other words, the histogramm in the embeded jpg for yout base-image (the ISO1600) should be in the middle, not to the left, like in your example.
Follow the link and use code "alyn" for your free month and surprise gift with a 2-year subscription to NordVPN - nordvpn.com/alyn
)W
BTW, the A7iii has a bright monitoring option so you can compose images in low light without changing your settings.
Well done Alyn! I'm glad you did this video so I didn't have to 😜 You did a very nice job of explaining the major concepts in plain language and how they relate to astrophotography. I think your lead-up explanation of exposure stops and then the visual difference between ISO100 pushed 5 stops vs. ISO3200 at 04:51 is especially effective, and will hopefully get more people thinking about this in the correct way, and not just insisting to me that it's as simple as: higher ISO = higher noise. Also, that's an epic filming location for your talking head stuff!
hey Nico great seeying you here ^_^ !
grtz Johny Geerts
Thanks Nico! Glad you approve. Yup, understanding exposure stops is not only important but very useful!
I do remeber Tony Northup had a video about this a while back, but he didn't give the astro take on it which is what most of us want. Nico your last video on the ISO was very helpful also.
Hey man, it was one of your videos where I first learned about this. I think it was a video about andromeda galaxy. You put up a link to a website which had a drop down menu list for pretty much every camera model and it showed on a graph all the jumps of the iso and where it levelled off. It was a great resource for anyone who wants to find out their best iso for low light without doing all the work. Great video all round too
@@JustinMcNeil Exactly I was going to mention the same. I moved to the Comments and Nico was already there lol
Rest in Peace man, you've taught me so much.....
You will continue to produce information for all of us even in your absence. I hope that those stars swallowed you up and we can all continue to hear you speak through them.
I realized that my whole photography life was a lie after watching that video...yeah that makes sense Alyn. Great presentation! :)
17:40 - Sony a7iii has a function called "Bright Monitoring" that enables you to see in the dark before the shooting without changing the ISO. Just assign a custon button. Thank you for the video!
Yes bright monitoring is great, I've mentioned it in a few vids! But at that moment I was talking more about the impracticality of being able to preview your RAW files. Yes you can compose with Bright Monitoring and you can take your images at 640 but what happens when you get home, boost the exposure and then there's all sorts of rubbish and things in the composition you didn't notice
Good video, I just have two comments on it:
1. You mention that higher ISO means lower dynamic range. While true, it's important to realize that simulating higher ISO by underexposing and correcting in postproces does exactly the same. When you underexpose by e.g. 3 EV, it means that you only use 1/8 of the available range, i.e. lose 3 bits (or 3 EV) of dynamic range. And when you check the graphs of dynamic range by ISO, in the intervals of ISO invariance you can see exactly 1 bit (or 1 EV) of loss for each ISO step. Which is not really a surprise, considering that in these intervals you essentially just choose if it's camera processing chips or your software in PC who does the multiplication or division, the values produced by the sensor are the same. In other words, dynamic range is not really an argument for underexposing and correcting in postprocess, the only reason to do that is to prevent clipping.
2. People often tend to think that an ISO invariant camera automatically means it's better. This is not always true, all it means is that the sensor has only one actual gain value and every other ISO setting is simulated by multiplying or dividing the values from the sensor by a constant - i.e. exactly what the exposure correction in postprocess does. While more convenient, it does not always mean a multi-gain sensor could not achieve better results (in the sense of less noisy image) for a particular combination of exposure settings. Often the reason for using a multi-gain sensor is to get better results in some parts of the range, it would be easier and cheaper to use just one gain and simulate everything else by recalculating in firmware.
Recalculating the firmware?
@@rebeldown771 recalculating *in* firmware. The firmware simulates higher/lower ISO by mulitplying/dividing measured values by a factor.
Finally!.. This is the first video/article I've found which explains it to my satisfaction, as a photographer (and as a Sony shooter). Seems like most people seem to get bogged down in more of the technical aspects of it all, causing me to involuntarily tune out. You spoke to us as a photographer, rather than an engineer, and I followed you from beginning to end. Thank you.
This was so useful Alyn. I've struggled with articles about ISO invariance but you make it so very clear and understandable. Thank you so much.
Por fin un fotógrafo que nos enseña técnicas, valores y todo tipo de tips con seriedad y estudios en el campo y en la edición.
Good explanation, in fact the best I've seen. As I use a Fuji XT-2 I've come to set the ISO at 1600 for everything and adjust in post with the exposure module (Darktable). Works great. I'm pretty sure that Sony developed this technology and Fuji uses the Sony sensors (although they don't like to admit it). One additional point - while the noise stays the same the dynamic range does not. As you increase exposure (in post) the dynamic range decreases. (You covered this later in the video!).
Fantastic video as always Alyn and thank you so much for recommending my video :)
I couldn't agree more. It is exactly how I choose my iso. I just add that I reduce the iso if I am doing a single shot or stack images. If it is a panorama I reduce too but keep an iso where I still can see the composition in each shot.
Finally, a clear and complete explanation about ISO invariance. I’ve got a Sony a7III and I’m gonna do it from now on.
Thank you very much, Alyn!😊😉
makes total sense!
Alyn... You blew my mind
Explained sooo beautifully...
RIP sir
Thanks for another really informative & useful video. “Protect the highlights & increase the dynamic range will be my new mantra”!
Total noob here! Thanks, I've never come across this topic regarding ISO before. All I have interpreted prior was the higher the ISO, the more noise introduced. Always keep it as low as possible. Thanks to you, not only do I realize that is not the case, I also know how to figure out what is likely to suit my camera best. Super job of conveying a complex perspective so I could understand. Worthy of a 3rd "Thanks!" for this video.
Thank you very much. I think this was one of the most important videos I watched the last years. A7iii.
Thanks a lot, I just spend the night on Hearth Nebulea, at 3200, but I just figured out that my best ISO is 800 !
Need to take it again !
This is turning my head upside down on my go-to astro camera settings, incredibly useful info! Looking forward to trying this out.
This video is PURE GOLD!!! Thanks Alyn, much appreciated
Good job explaining this complicated topic! As a long time user of the Fuji X-T1 I can confirm that the camera is pretty much ISO invariant, so much so that I was taking milky way images at iso200 at one point to not blow out the highlights in a campfire. As it turned out, I unfortunately overestimated the performance of the sensor and the darkest regions of the image resolved nearly no detail compared to a ISO1600 exposure. So even with the X-T1 there seems to be a limit of how strong you can underexpose.
But in general its still pretty amazing how much freedom you have with the X-T1s raw files.
sorry for copying your profile picture idea by the way ;)
I've been trying to get my head around ISO invariance for months but couldn't grasp it. That was until I just saw this video which explains it perfectly. Thank you. i also own an A7iii (bought after watching many of your videos). It's nice to know I have the tools, if not yet the skills, to attempt astro photos like yours.
Well shit, I have some homework to do this weekend. I knew about this issue but I tend to go to iso 100 every time and now it seems that may not be ideal. Thanks for this Alyn!
Wow! This is incredibly useful information. I've gone all this time assuming ISO100 was always best when possible. I am eager to do this test with my camera and learn at what point it truly performs best.
Thank you so much Alyn! Your repetition is perfect for teaching new concepts. I learned so much from this video - even what a full stop is pertaining to ISO.
Many thanks for this Alyn - I've learnt something new today! And the best news for me - my camera is an XT-1 😁. So I never need to worry about ISO ever again - Auto ISO for me from now on 🥳🎉
Great explanation. As a software developer who worked with computational photography, iso invariance is a very hard problem to solve. Ideally a short timelapse helps solve the iso issues. Sadly you do get star trailing if done incorrectly.
Well, it's night time and a near full moon, I can't wait to use this info to test out my Nikon Z6! I have been defaulting to ISO4000 for my Milky Way shots for the last year.
For the Z6, any ISO above 800 is going to be the best option!
Nora by Space Light! that explains a lot and even simple enough for Noggin here to understand! Thanks so much Alyn - certainly wasn't aware of this and very much appreciated!
Great Video as Always! Very informative! In your video you mentioned you could raise your iso to it's highest setting to compose then lower it back down to the minimum invariant iso.. what I do on all my Sony cameras is turn off the live view setting effect.. to compose.. then turn it back on for the shot.. I have setting effect on/off as a toggle on one of the programmable buttons. If you know your manual settings are correct and you are going to shoot all low light.. you can leave live view setting effect off the whole time. Works for me.. your mileage may vary.. lol.
I feel like i should pay you for this lesson haha. Top stuff mate!
Nordvpn does
Video starts @3:16
Your awesome alyn
I have learned more from you than i have from 3 years of my own experience
I'm happy there are people like you that would like to share their information instead of hiding it
Thank you alyn
Thank you
Alyn, This is an excellent video and introduced me to something I was completely unaware of. I'm just starting out with Milky Way photography (I love it) and was under the impression for shooting foregrounds/dark parts of my image to shoot a longer exposure at lower ISO to try to keep the noise down, but I wasn't having success at all and my low ISO images were very noisy. This completely explains what is going on and that I should not be afraid to shoot much higher ISO to capture information in the shadows. I'm shooting with a Canon R5 and ran this test myself and found that between ISO 800 and 6400 my camera is ISO invariant which is a huge bit of knowledge for me and should greatly improve my images.
I'm gonna have to watch this again with a strong coffee. Amazed that high ISO does not equal higher noise!
It's not the high ISO that causes noise, but the lack of light.
They never ever did. Higher ISO always lead to dropping read-noise. After the point, were higher ISO stop the decreasing of read-noise is called "ISO-invariant". This was (not only) described in 2010 within the canon-cosmos.
i know i watched this once but wow tonight i understand about
5% more. there is hope for me. i enjoy every minute 🤩
For those of you using the Canon Rebel T7/2000D/1500D/Kiss X90 (same camera, different names).
I tested the noise performance at every ISO level. I heard it is the best-selling camera on the market right now so I thought it'd be useful to share the results. All the shots were taken in a relatively dark room at night. The images were shot at 50mm F/1.8 and 1/30s shutter. The images were matched to the ISO 6400 shot in RawTherapee 5.8 with the neutral processing profile. So here are the results:
ISO 6400: Reference image, and highest ISO the camera will go with stock software
ISO 3200: Identical noise performance to 6400
ISO 1600: Performace is very, very, very slightly worse. As an analogy, it would be comparable to measuring the height of your room floor to ceiling and being off by a millimeter or two. It's only noticeable if you're specifically looking for it and then you really have to be looking for it. This seems to be the ideal ISO as far as I can tell.
ISO 800: Still okay. If you can afford to take some more exposures for stacking, you'll get away with it, but I still recommend going 1600
ISO 400: Only use if you really have no other options and you're fine with a very sub-optimal end result.
ISO 200: Just stay away from this one.
ISO 100: Don't even think about it.
TL;DR: Use ISO 1600 for very slightly better dynamic range, use ISO 3200 for very slightly better noise performance, ISO 800 is mweh, but okay. Hopefully, that'll save you some effort :)
If that hadn't been so well explained, it could have been a very difficult topic to grasp. Amazing job dude.
That's something no one talks about.. thanks man 💫👍
Oh man! I just bought a Fuji X-T1. I've heard of ISO invariance in the past but never thought I'd get to play with one! That's awesome! Great video! Cheers from Costa Rica
You made perfect sense but my mind is blown.
Testing my cameras tomorrow fro Invariance.
Thanks Alyn
Brilliantly explained Alyn ! Looks like my old D750 is coming out of retirement for my astro shoots !! Diolch !
Thank you for explaining that concept very clearly ! Will use the 640 ISO on my A7III for astrophotography, low light scenes and day to night time-lapse from now. Nice to know there's always a way to improve our skills and image quality 😉
It DID make sense! Very conclusiv explanation - even for the German listener. Thank you very much, Alyn!
Excellent vlog! 👌👍💪 Goes well with the one by Tony Northrup where he demonstrated that noise is due to lack of light (wrong aperture/shutter speed) not a high ISO setting.
This is very good. I knew very well before watching the video what ISO invariance is, but I still believe this might just be the best explanation of the topic.🙏🏼 Thank you.
Makes me even more happy with my decision to purchase a R6 in late 2020.
Another excellent discussion. Something no one else discuses. But in the end you still have to make compromises,
my 6dii was not quite what I wanted and when I switched to xt3 I was way over pleased even after coming down to APSC, For the life in me I couldn't figure out why. Now I know! Perhaps I needed more finesse in my skill but only if I had known, so Thank you!
very informative Alyn, never looked at my camera performance in this way so will definitely try it out
Excellent video Alex. Had the Fuji XT1 which I used for a few Milky Way shots. Was always impressed with its noise performance, especially when boosting exposure in post.
thank you so much, this is perfect timing as I just changed camera
Thanks, Alyn, spot on! First times out with my Nikon D7500 I was afraid to shoot over 800 because that is what people were saying, and still are. I have since started shooting by looking at the histogram to move the black up so they are not clipping and the whites so in a log-scale are not blown out on the right side, using whatever ISO gets me there. That would make for another great explanation from You or Nico. Daytime photographers talk about it all the time, but not hardly a word from astrophotographers as to what the histogram would typically look like for great photos.
Something I talk about in all my workshops, maybe I should make a video yes! Thanks
Looks like I'm going to get busy and shoot some test shots...interesting video, thanks.
A great explanation Alyn now I’ll just have to test my Fujifilm. Thanks for a very educational vlog dude. 👍🙏
Thanks Alyn. I've been tempted to switch back to Canon from my Sony A7 and this has given me second thought. Maybe I'll have both for a while and do some experiments :D
Great video Alyn, this will give me something to work on in the next few evenings..
That was insanely useful. Thank you for the detailed explonation
First time I've heard of this. You explained it very well and easy to digest. Thanks for sharing.
Complicated...but i got it. I'll test out my camera... in the past, I just used what looked good in my viewfinder...now I know to take test shots so i can check focus and my framing and then pick the best iso for my camera for post processing and final image quality!
It’s easy to blow highlights at each base ISO start (100 and 640 for my Sony A7R3) because all the DR is in the shadows. To not blow out clouds when they are backlit, it’s better to shoot the highest ISO you can get away with up to and including 500. Check the exposure by spot metering on the clouds at +2.
Indeed!..thankyou so much for subjecting yourself to so many non variants of iso invariance and even variations of iso variance.. ;)
Wow, that's super useful. I'm buying my first dslr in march and these tip videos will most likely be very useful for me. Keep up the good work Alyn
Cheers Alyn, as an A7iii user that was great for me, very well explained.
Thanks Alyn for this detailed Video, this gives me a better understanding of my A7 III while out in the field for NightSky photography. Great content as usual !!
Close to 80K subscribers Alyn :) An interesting vide once again, thank you.
Agree with you on the noise and on better dynamic range, but over exposing a RAW file in post over +2 EV tends to show discoloration and/or artifacts depending on what image editing software you're using. I like to play it safe by shooting a well exposed image the scene dictates, but will give it a try in a high dynamic range situation and see. Great video. Clear skies!
Brilliant.... Learnt something new and so clear, going to test my Canon 600D now to get the best ISO for my astrophotos. Great tutorial 👍
this was really interesting, and I'm surprised to see my Nikon D5600 seems to have the same noise after ISO 400. Surprising for such an entry-level camera. This helped me understand ISO much better. Until now I thought ISO was pretty pointless and was basically the same as the exposure slider in lightroom. After watching this video I guess I was right, but now I know why! and I also know that won't be the case for all cameras. Also didn't know ISO effected dynamic range, that'll be super handy in future!
Thanks Alyn - wish I'd watched this a couple of days ago!
Excellent video as always Alyn! As a Canon R5 owner I have to point out that the R5 has a dual gain sensor, ISO invariant from ISO 400 up. Folks can read about this and compare to other cameras (yes, the a7R IV performs better in this regard!) in the - excellent as always - review of the R5 at DPReview.
First, a direct greeting from Brazil. Your video is excellent, very well explained and didactic. Congratulations on the beautiful work.
brilliant , well done sir .
learned some more , as 15 y photographer .
And now go to sleep you look a bit exhausted , althought sounded very clear .
grtz Geerts Johny
That was a fun little experiment! I was not expecting the drastic color cast at the lower ISOs though. Proved very helpful in understanding the capabilities of my camera better. Thanks.
Very informative & interesting Alyn. Another consideration for me is megapixels. I shoot a lot of timelapse’s & have been using the Sony A7R2 & A7r4. I recently got the A7s3 & have found for 4k timelapses it is so much cleaner doing a day to night timelapse (using aperture priority & iso minimum shutter speed). The quality of the finished sequence at 12mp is, at 4k much better/sharper/cleaner than the 42mp & 61mp sequences. Also by limiting the shutter speed & letting the iso ramp up, I don’t get that annoying speed up effect that I have to correct in post. 🥰👍🙏😀Thankyou for producing great content & I very much appreciate the very informative videos that you produce for us. Cheers!
Yup, those 12 megapixels are light sucking beasts. Megapixel count has quite an impact on low light performance. Love my A7Siii :)
thanks for that mate , now I can save the moon in my all night long time lapses with a moonrise . I always struggle with that inherently Good On Ya!!!!!!
Great video thanks. I was aware of the ISO invariance of my D850 but this is a great practical example of what that actually means in real life!
Alyn, thanks for the time explaining, I watched many others, you only one for me that make this clear, I thought was a myth, great job!
Thanks for all the work you put into making the videos
I might have to watch it a few times to get it in my head
My second rime around with this video. Great explanation! I get it now!
Excellent video Alyn, and the best one I've seen in the topic.
With aurora photos I have found it more practical to keep ISO in 1600-5000 range to be sure that the exposure (judged from the histogram of the captured image) is exactly where I want is to be so that the adjustments in post are minimized. This because intensity of the auroras and the ideal exposure is constantly changing and I can't evaluate it from the live view (especially when you are in the dark staring at a led display which is showing an underexposed image...). The only way to be sure about the exposure is the histogram of a captured image (at least with Sony a7iii where in case of a long exposure the live view image and its histogram are not anymore wysiwyg).
Wow this was so eye opening and informative thank you so much Alyn !
A good tool I've seen Nico Carver from Nebula Photos use was the website called "Photons To Photos". They have an input-referred read noise chart for a bunch of cameras which shows the levels of noise depending on the ISO settings.
Great video Al!
Thanks dude!
Excellent work and thorough description on the ISO-invariance topic. Excellent info.
Alyn this video is pure gold 🤯 💪
Awesome mate. Explained it very well. Would love a video of your post production methods when shooting at iso640 and how you boost the exposure back to get the correct exposure for the shot. Assuming you don’t do a global exposure slider adjustment if you have objects like the moon and street lights? Is your post done in photoshop?
Good idea thanks. Sometimes I edit the RAW file twice and merge it into a HDR image :)
I quickly googled to see which cameras are iso invariant only to discover that my nikon d5500 is among them. This was a surprise and makes me wanna test what you just told us. I am not delusional of course, it cannot be compared to a full frame camera but it is certainly a good feature to have.
I have nikon d5500 as well what have you found your iso setting you like
@@caviestcaveman8691 my limit is 1600 for acceptable photos for this camera. I rarely will go more than that. However I have yet to try the new denoise programs. I would like to try using them on older photos with higher iso settings buy haven’t managed to do so yet.
Your videos are always interesting, even if I can't use your explanations for making astrophotography, it is interesting and fun to watch them because it makes me understand the process and the nature of a made photo as opposed to seeing it in real yourself. I have some experience with other photos though. ❤👍🏻
So, Im mostly in dark scenes ... using iso 3200/6400 with flash on body to get correct exposure and balanced ambient and flash light.
I was so worried about ISO, this makes me relieved about using even iso 9000/12800 on my D750 if I need to, because basicly....software increases brightnes above ISO200 the same way as I would in LR. 😎 This is amazing, thanks 😁
So do I understand correctly?
With a iso variant camera, increasing iso is truly raises the signal compared to the noise level making the SNR better up until a point where it stops and at least not getting worse even with crazy high values.
And with an invariant camera you can choose whatever iso you want it won't change the Signal compared to your noise level.
So actually raising your iso doesn't make your image more noisy in either case it gets noisy because your exposure time and aperture is not enough to increase your signal out of the noise level.
Thus the only thing is lost with higher iso is dynamic range and iso variant cameras are somewhat forced to lose some of it to have less noise.
You first need to understand that noise is introduced into the image before ISO amplification is applied and afterwards.
ISO invariant cameras introduce very little noise after ISO amplifcation. So ISO setting makes little difference to the signal to noise ratio.
ISO-variant cameras introduce more noise after ISO amplification. So if you boost the signal before that noise is introduced, the signal to noise ratio improves (but you're not collecting more signal).
Check out the technical explanation video linked in the comments
There is a fault in your real-life-nighttime-test: You will have to exposure the 1600 well. But in your test you underexposured the 1600. It is importend that the base-image for that test is right lit, or in other words, the histogramm in the embeded jpg for yout base-image (the ISO1600) should be in the middle, not to the left, like in your example.
Yeah that makes sense! You explained so well. Thank you, thank you, thank you for this lesson of gold!!
SO so so helpful! Explains some issues I was having. Made notes! Thanks! 😁
Top notch as ever bud 👍 no on ever talks about the d7500 and being a newbee really helps witn settings. Appreciated lots thanks.
Very well done on a complex subject. Loving my D750!