Thanks for the comment. It is true that ISO involves more than simple amplification of image data, but the statement that ISO is amplification is correct either way. I am going to explore this in a later video, but in my article I explain the difference between ISO-variant and ISO-invariant cameras, terms you are probably familar with. The challenge with explaining these concepts is that to a beginner, the simple foundation that ISO is amplification is just the foot in the door to the underlying physics that governs and controls how cameras operate. It is logical to explain the basic foundation before delving deeper into the niche field of scientific photography. So, while you're right that amplification is not the whole story, really, if you wanted to explain the science of photography fully you would need to record an hour-long video. Communicating science is a fine balance of getting the "jist" of the concepts across without confusing the audience and without misconstruing the very technical science that underlies everything.

@@Apalapse There are no ISO-invariant cameras or, viewed an other way, every camera gets ISO-invriant when raising the ISO to a certain value. (So canon-cameras did 10 years ago, when and were the term 'ISO-invariant' popped up first time) Yes, you are right. Beginners need some straight answers. :-) But: I have talked to so many beginners in low-light (astro-landscape)photography and all of them fear, that a high-ISO would make the images noisy. In fact high-ISO is a need to have low-light images with nearly no noise (when having a fixed exposure-time). So the test, you explained will give not the result you told, when talking on low-light situations (and fixed exposure time) I understand your intention to keep things simple. But keeping it simple should not lead to a result, which can be mistaken. So over all: Yeah, a good video for total beginners. :-) Apologys if my comment seemed to harsh. :-( (I dont know why youtube send me here.)

Your comment wasn't harsh at all and I really do appreciate your feedback and our discussion since this is how others can learn. Discourse is essential in science! The result I posted was actually a low-light shot, and based on my empirical data I have found that while ISO- invariant, my camera produces better images when I boost them in LR.

@@Apalapse I guess you are both right 👍

Yeah, I would bet ISO applies amplification of the signals at a hardware/voltage level before pixel readout and software processing, that is why you cannot change ISO after the fact.

Yes!

No such thing when it comes to ISO and digital cameras :)

The efficiency of the sensor is just better in those cameras. Lower read noise, larger photosites, a beter microlens design, back-side illumination, and better signal processing all can make a photo look "cleaner" at the same ISO as another camera with similar specs. I have a Sony a7s and, yes, you can literally see in the dark with that camera, but that is largely because it is a 12-megapixel camera using a full-frame sensor and the sensor tech and signal processing is very good. Low-light performance of cameras does not change how the principle of ISO amplification.

The video is absolutly correct. The sensor is an anlog "tool" and sensitivity can not be changed. Furthermore, the amount of photons reaching a sensor in a fixed time can not be changed (only by lower f-stops, a 'brighter" aperture). So a given amount of photons will "fill" the sensor and "produce electrons". The analog-digital-converter will convert these fixed abount of analog photons to digital values. And at this point ISO "kicks in" an amplifies the signal. The sensor never ever can get more sensitive. Physics just dont allow this to happen. Beside the amplifing there are some software-developments on the (now digital) data, which depend on the set ISO. But the sensor still is as sensitive or not as always. No matter how you change ISO. So the video is absolutely right in that point.

Haha well now you know what it really means ;)

Thank you very much, glad I could help.

Thanks. The ISO myth is really perpetuated in the photography community and I thought it was time I addressed it.

Yes this is correct. Camera ISO behaves nearly the same way.

Yes it will! ISO is a concept that is hotly debated among photographers. I have taken two photography classes from two different universities and when I tell them ISO is fake, they think I'm crazy. Even after proving it to them RIGHT IN FRONT OF THEM, they still don't believe me. Some people are too set in their ways to realize the institution has deceived them, but proper education is the first step in setting people straight.

I made sure to pronounce it eye-so this time around, since that is originally how it was meant to be said.

Yep, the "ISO is sensitivity," "ISO is exposure" BS comes from the fact that in the days of film, all of that WAS TRUE! People just haven't caught up to the fact that digital does not work remotely the same as film did.

Glad I could help!

This theory is a one that a lot of others have brought up, and it is true that not all ISOs will show equal noise compared to under-brightening in-camera and raising the brightness in post. As a result of this, shooting at higher ISOs can sometimes produce less noise than shooting at low ISOs and brightening later. This is because the noise in your photos is comprised of shot, upstream and downstream noise. Shot noise comes from the randomness of light and follows the rules of SNR; more light = more quality, etc. Upstream noise is all the noise that is inputted by the camera's electronics into the light collected as an analog signal. This usually comes from the sensor's electronics. Then, the light passes through the analog-to-digital converter (ADC) where it is converted from light to digital signal. In the ADC the signal is also amplified, which we now know as the ISO. After the amplification there is still some noise that is contributed by other electronics in the camera as the digital signal is written to a file. This noise is called downstream noise. Most cameras typically have very little upstream noise, and the amount is almost always more than the downstream noise. These values are mostly fixed as well, meaning that they don't change with ISO; they are just an inherent part of the camera and how it was designed. So with the final image all the other noise sources besides from the light itself (shot noise) is upstream and downstream. In a low-ISO shot, we know that there is very little amplification, but because the upstream and downstream noise amouints don't change, since downstream noise is a much higher value, it becomes the majority of the noise we see. In other words, if the total amount of amplification is LOW coming OUT of the camera, the downstream noise will make up a larger total fraction of the noise than the upstream and shot noise will. Downstream noise usually presents as banding or strange color artifacts and it looks objectively worse than normal random noise. Raising the brightness in post means that we are also amplifying the downstream noise which like I said represents a larger component of the total noise in the file. At high ISOs we don't have to worry as much about downstream noise since the amplification of the upstream noise and shot noise makes the fraction of downstream noise very small. If we choose to under-brighten in camera by choosing a low ISO but because the scene is dark we amplify in post, then we hvae a situtation where the downstream noise represents a larger fraction of the noise and brightening in post may cause fixed-pattern noise which is not good! So, with all this information, what is the application? Like I said, different cameras will have different amounts of downstream noise, meaning that amplifying in-camera may make more sense than in post since the downstream noise is ultimately reduced that way. This leads to the discussion of ISO-invariant vs. ISO-variant cameras. ISO-invariant cameras have very very little downstream noise so there is usually no downside to amplifying signal in post. ISO 100 +5 stops looks the exact same as ISO 3200 in an ISO-invariant camera, and since dynamic range DECREASES as ISO increases, boosting in post is better. ISO-variant cameras have a lot of downstream noise so in those cameras it is better to shoot at a higher ISO than boost in post since there will always be a high noise floor of downstream noise. Most cameras are a combination of ISO-variant and ISO-invariant, including the Sony cameras. Look at this: www.photonstophotos.net/Charts/RN_e.htm The Sony A7III uses a dual-gain ADC where the downstream noise (input-referred read noise in the case of the graph I linked) is significantly lowered around ISO 640. This is why you may hear Sony users say "shoot at ISO 640 for low light then boost in post," because between ISO 640 and the maximum ISO, while the downstream noise does decrease a slight bit the hit to dynamic range you experience (see www.photonstophotos.net/Charts/PDR.htm). TL;DR: ISO always amplifies and how exactly noise changes with ISO is not a function of what ISO is doing but rather has more to do with the camera's electronics and how light signals are processed inside cameras. It's fascinating to think about, but only for really really niche types of photography is this level of detail useful. Hope this explanation helps.

@@Apalapse Amazing explanation man, it helped a lot :)

@@Bluesky35102 Glad to hear. I plan on making a video explaining this at some point but it will take a lot more research and like all science, once you think you understand it you find out what you learned is just a generalization of some more complicated theory and you have to start all over again. ISO is one of those concepts where on the surface, it's just amplification, but if you keep going deeper you realize it implies much more.

@@signofzodiac High ISOs amplify image data, so whenever this is appropriate is a good time to use high ISOs, which would be good when you don't have enough light to adequately view the photo on the back of the camera (make the image an adequate brightness).

Thanks for mentioning the animation. I've gone to a simpler but more elegant style recently and I really like how it turned out, glad you liked it too! Take care.

Awesome, thank you!

I am not sure, because, objectively, if you kept the brightness the same in each photo but integrated longer at a lower ISO, your picture quality should be greater. A longer shutter speed would give you more light and thus you should see fainter stars than on the high-iso but shorter shutter speed shot.

Some cameras are ISO variant where the boosted low ISO shot will look worse than the high-ISO shot. This is because the analog-to-digital (ADC) converter just applies too much noise to the image and the ratio of this fixed-pattern noise compared to random noise is higher in low-ISO shots than high-ISO shots.

jpg has compression while RAW doesn't

What don't you understand? Let me know and I'll see if I can make things clearer.

@@ApalapseI watched the video again and I think I'm the only one who didn't understand..so if as you explained and the radio example first question , why the manufacturer didn't put fixed iso ? And the second question what iso for ? Thank you in advance , appreciate your time and effort

@@bazrezk7218 Well there's an obvious downside to shooting at a fixed ISO all the time: a lot of your low light images will be dark! Using ISO means you can easily see the image on the back of your camera, otherwise since the images would be dark you wouldn't have a clear idea of what exactly you were capturing (you can't chimp if the pictures are super dark). ISO is helpful when you need to see what you are photographing, and it's a tool used in live view on DSLR cameras and in the viewfinders of mirrorless cameras for you to see your exposure in low-light conditions.

@@Apalapse got it , thank you very much for your explanation and your time..highly appreciate

Raising the ISO in low light situations only means you are amplifying the weaker signal entering the camera. ISO is equivalent to amplification which means that whatever light the camera receives is whatever light makes it into the final photo; ISO does not change that. If you have more light available to you, which you can achieve either by increasing the size of the lens' aperture, lengthening the shutter speed, or adding more light to your scene (adding a flash, strobe, or moving to a brighter environment), it will be easier to use a lower ISO.

Raising the ISO in LowLight will give a a correct exposured image, which will have less noise, than a underexposured image with low-ISO, which was brightened in postproduction, like Lightroom. Readnoise, darkcurentnoise decrease with increasing ISO. Furthermore the ISO setting changes the way the in-canera pre-raw-development works, becouse ISO is more than only amplification of the signal.

​@@sternenhimmelfotografierende First off, dark current noise is not something controllable by the camera. Also, although this is semantics, raising ISO does absolutely nothing to image exposure. Exposure, as I have stated, is determined solely by the integration time, lens t-stop, and scene luminance. In ISO-variant cameras the read noise is high and thus at low ISOs, boosting a dark photo will reveal the fixed-patterned noise that is a characteristic of this type of noise. Canon cameras are notorius for this. ISO-invariant cameras have a lower noise floor and thus there is no disadvantage to boosting brightness in post as opposed to doing so in the camera using ISO. Remember that increasing ISO also decreases the dynamic range of your image since a smaller range of light is quantized by the ADC. You are right that ISO is much more than amplification, but arguing that it does one thing over another is a moot point when you fail to acknowledge the fundamental reasoning behind it and the fact that it varies wildly from camera to camera.

@@Apalapse There is no iso-invariant-camera from ISO100 up to whatever! It is a myth! Get out at night to a really dark location, take your a7s, a 50 mm lens, open it to f1.4 and take images for 5 seconds (Make sure you will have some foreground and some sky within the frame). Change ISO from ISO100 to ISO12800. Bring them to lightroom and equal brightnes. Your a7s ISO100 image will be much more noisy than the ISO3200. Compaire ISO 1600, 3200 and other high-ISOs. Yes, they are ISO-less, exacltly like canon, nikon and other brands. But for low ISO-settting no camaera is ISO-invariant. BTW: You will like this: clarkvision.com/imagedetail/digital.sensor.performance.summary/

sternenhimmel-fotografieren.de I have read that article. Roger has great information. Unfortunately I have done your test before and the image raised in Lightroom as opposed to in camera always wins. This only works at ISO 640 because the camera I use (Sony a7r2) has a dual gain setting and downstream electronic noise is masked at ISO 640. Dynamic range is also a function of ISO, and I prefer to have more of it if the result using in-camera ISO looks the same as boosting exposure in LR. I can find examples and post them here later.

I don't think so. I think simply saying that ISO is "amplification" is enough explanation to bust misconceptions without confusing the viewer. The exmaple I showed illustrates the fact that amplification in Lightroom is the same process by which ISO operates. The loss in quality you are referring to is only a consequence of the camera model. A lot of cameas have a very very low noise floor, so any downstream noise will not be pereceptible in the final image, even at low ISOs boosted with software. These are ISO-invariant cameras. Some cameras, especially Canon cameras have a lot of downstream noise, so higher ISOs will produce less noise, but at what cost? Increasing ISO DECREASES dynamic range, and in my opinion this is a worse effect than a slight difference in noise. In the end when you get into the itty bitty details of sensor architecture and signal processing there is just too much variation between camera models to make me comfortable including that in this simple video. A longer, more in-depth guide is in the works, though.

​@@Apalapse "Increasing ISO DECREASES dynamic range" I've also heard that very low ISO settings can decrease dynamic range, but so far I haven't found much reliable data on that. I think I understand what audience you're trying to reach, and from that perspective, you're making a valid point. What bugs me is that what a lot of people will take away from videos like this is that every camera is iso invariant, which is definitely not the case.

That's the goal. I had to take a break to focus on my studies and career but now I am back in it.

The reason is that over the summer and during fall semester I was extremely busy with school and research. I found a little more free time this semester so I am hoping I can upload more consistently moving forward.

Of course.

That's definitely an interesting read, and I actually plan on doing a video similar to that where I FULLY explain how cameras work and how a digital image is created. The underlying science is quite complex to a beginner but my job as an educator is to try and distill that down to a level everyone can comprehend. Thankfully, that book you linked does a really good job of that.

Thanks, and sorry for the late response. I use Adobe Illustrator to animate the vectors then import everything into After Effects to animate the layers.

No problem!

This is true and in the other comments I addressed that in this video my implication was just to show a case scenario where ISO was the same as EV adjustment. I never suggested that someone should use that method in their workflow. In order for me to make that recommendation I would have to explain more about upstream noise, shot noise, downstream read noise, and the concept (and a little bit of a misconception) of ISO-invariance and ISO-variance with camera models.

Apalapse thank you for answering and the clarification.

Perhaps.

@@ApalapsePerhaps

Indeed.

@@brendenm1284 Quite.

The voltage of the signal is increased OFF of the sensor but the sensor itself cannot magically increase its voltage when gain is applied, that just doesn't make sense. If you think about it, the sensor is not powered at all. The voltage you describe comes from light and the amount of voltage produced per unit amount of light is determined by the sensor's quantum efficiency. Also, if there is a way to magically increase the sensitivity of the silicon, why wouldn't all sensors behave like this? What would be the point of making an ISO-invariant sensor at all? There is no such thing as an ISO-invariant "sensor," and even the concept of ISO-invariance is not really true since downstream read noise decreases as a function of ISO. The reason some cameras are labeled as "invariant" is because the level of noise improvements with using high ISOs is simply not worth the loss in dynamic range.

Yes, and the result of if it's ISO invariance? Both shots look the exact same. The result of ISO variance? The lower ISO shot has more noise, which is still contrary to the misconception. I understand that's not really fair to the photographer since they may not get the same level of noise in each shot which is why I am making a follow-up video where I explain ISO invariance vs. ISO variance. In reality, though, there really is no truly ISO-invariant camera. How a camera captures, digitizes, and analyzes signal is such a deep subject to cover thoroughly, and this video was just a stepping stone for photographers to start asking these types of scientific questions and care more about their photos.

@@Apalapse Viewers who don't know whether their camera is ISO invariant won't know whether this is good information or entirely misleading. I would expect a competent photographer working within the limits of the camera to get a good exposure pretty much every time, and not pray that it's recoverable in post. I have seen reviewers demonstrating recovery of under exposed images and making a big deal of it, good or bad. If you can't get a good exposure with your camera, brushing up on your skills is far cheaper than searching for a camera that can cover over your incompetence.

I agree that having the foundational skills to know how to craft a properly-exposed photo is more important than knowing how to recover an under-exposed photo. It's also important photographers realize what ISO is doing to their images so that they don't make uninformed decisions regarding camera settings. One of the biggest pitfalls of obeying a "traditional" misinformed view of ISO is photographers think that increasing ISO while decreasing exposure through either shutter speed or aperture to maintain the same image brightness produces images with equivalent exposures. This is false since the ISO is merely amplifying the signal to make the image appear properly "exposed." Furthermore, the manipulation of ISO has nothing to do with camera exposure. In an ideal scenario, you would know your shutter speed and aperture needed to produce the desired amount of motion blur and depth of field you need. Setting the camera in manual mode with auto-ISO ensures the maximum signal is collected with amplification occuring in-camera. Unfortunately, like many realize, ISO also decreases dynamic range, so in some situations it is better to under-brighten the image in camera and raise the brightness using post-processing software like Lightroom. There are also instances where the image will be underexposed no matter what you do, i.e. astrophotography and low-light action. Here, a photographer's only choice to return the image to a properly-brightened setting is to raise ISO or increase brightness later. Which choice the photographer makes depends on how their camera behaves; whether it is ISO-variant or ISO-invariant. This video did not mention ISO-invariance nor ISO-variance. I merely showed a case where ISO and post-processing EV produce identical results.

There's no response needed.

This comment is confusing. You are using "digital" and "film speed" in the same sentence. Digital ISO =/= film ISO. One is an amplification setting that is applied to an analog signal (voltage) before it is quantized into 1s and 0s to be written into a RAW file while the other is physical silver-halide crystals that are sensitive to light. Also, SNR has nothing to do with ISO. It's SIGNAL (light) to NOISE (found in light). The ratio of noise in light does not change as you increase ISO. If your SNR is 200:1 at ISO 1600, at ISO 3200 it is 2(200:1) = 400:2 = 200:1 The ratio is the same. If we want to really go into it, the SNR would actually increase at higher ISOs since the amplification drowns downstream read noise produced by the camera's internal electronics, but for the sake of this response I am not getting into that. Hope this helps, and, sorry, but the video is not BS :)

Good evening!

@@Apalapse Good Evening to you too :)