Scott, as usual you do a great job making complicated concepts easier to understand. You are a very good teacher and I really appreciate everything that you do. Even though others have pointed out that they believe there are technical inaccuracies in the video, I don't feel that they are able to articulate their position in a manner that is easy to understand or practical to apply in the field. It is great that we have public forum like this where everyone can chime in, yet I rarely see anyone who can make things as easily understandable as you do. And that is what makes you a great teacher. I look forward to the rest in this series!
The sound example was great to explain how noise increases . The concept of Dynamic range coming down with the increase in ISO also was extremely well explained.
Scott, I appreciate the time, energy and effort you put into educating your audience. That said, you're presentation in these first two videos is filled with misinformation. The foundational error is treating exposure as image lightness. That's not what exposure is. Exposure has been known since the late 1800s to be the light energy per unit area projected by lens upon the glass plate, film, or sensor. As such, exposure is determined by the available light in the scene, lens aperture (f-stop) and exposure time (shutter speed). ISO isn't an exposure setting; never has been. ISO also does not amplify signal. It's not a source of noise. In fact, it's the light used to make the photo that is the primary source of noise in photos. It's called photon noise or shot noise. In simple terms, noise is randomness. If you put small water cups on a table on your deck to collect water during a rainstorm, you'd find the amount of rainwater in each cup will be similar but not exactly the same. That randomness in the amount of water in each pup is similar to noise in photography. If you photograph an evenly lit wall, the pixels on the camera sensor would detect a similar number of photons. They'd be similar but not exactly the same. The randomness in the number of photons each pixel receives is noise. The more total light collected by the sensor, the less visible noise will be. That's because the amount of light - the signal - overwhelms any randomness - noise - in how much light each pixel receives. The less total light used to make a photo, the less signal there is. This reduced signal allows noise to be more easily seen. ISO's role in all this is pretty simple. It's the setting we use to manage the lightness of the photo made by the camera. I say, lightness, to distinguish an ISO setting from the exposure (light per unit area of the sensor) used. Adjustments to exposure alter brightness - an amount of light - while adjustments to ISO alter image lightness or tonality from black to white. When ISO is increased to compensate for a limited amount of light being available, the resulting photo maintains a pleasing lightness *but* that increased lightness reveals shot noise that is present in the image. ISO doesn't create or add noise. It reveals noise that's present in the photo. There's a second kind of noise that can affect digital photos. This is noise that's a product of image and data processing in the camera. It's called read noise. Read noise is someone mistakenly referred to as ISO noise. The thinking goes something like this, "As ISO increases, the in-camera processing introduces more read noise which is what we see at high ISOs." If only that were true. In fact, read noise historically gets lower as exposure decreases and ISO increases. A decade or more ago, the standard practice in digital photography was to use the highest ISO needed for a given exposure. If too low an ISO was used, the photo would need to be lightened in Photoshop. That process would reveal the greater read noise associated with the lower ISO. Today, dual gain sensors have become fairly common. The advantage of this technology is that photos made at lower ISOs enjoy the same low read noise as photos made at higher ISOs. As a result, if you choose ISO 800 in the field to avoid blowing out white plumage in a bird and then lighten the darker plumage in post to bring out detail in the bird, the noise revealed is no worse than the read noise associated with ISO 1600 or whichever higher ISO you'd normally use to bring out detail in dark plumage in that lighting. The bottom line of all this for photographers is that f-stop and shutter speed are the critical settings. They control exposure and two important creative aspects of a photo: depth of field and the rendering of movement. If you choose the largest aperture (smallest f-number) that delivers an acceptable depth of field and the slowest shutter speed that adequately renders movement without blowing out important highlights, you'll optimize exposure within the creative goals you have for the image. After that, select an ISO that results in a photo having a pleasing lightness. In candor, ISO is a rather trivial setting. It's certainly not as important as the fundamentally-flawed exposure triangle would have us believe. It doesn't control exposure or add noise. It's how we manage image lightness.
I read this and appreciate your thoughts. Would love to see an video referenced that explains this easier and more accurately as I have read dozens of articles and watches numerous videos and perhaps my explaination is based on others information. But i have tried to make this concept easy for beginners and not sure who to do this easily for a beginner. The basic point was was to give an idea of what noise was and why we need to boost the signal to create our final output (when light isn't getting to the sensor). Also to show the DR decreases as ISO increases which should be factually accurate. Also appreciate your critique. Thanks.
@@2224jnk ISO communicates information the camera uses to process an image. Increasing ISO results in the camera processing a lighter version of the photo and that will make noise that's already present in the image easier to see. ISO isn't a source of noise.
Nice job Scott! You hit the nail on the head with Digital ISO and used a good analogy of the term GAIN. Comparing this to audio amplifiers is perfect. An amplifier with poor signal to noise ratio sounds terrible when one turns up the volume, hence noise compared to one with a great S/N ratio. This is true for a cameras sensor and the amplifier circuits used to process the signal. Look forward to you next video on dual gain sensors.
thanks! I generally get hammered but a couple electrical engineer types who tell me it's all wrong, but don't seem to understand that I am making video about general concepts and not advanced videos on circuitry and electrical signals lol
As per your dynamic range video loved it and particularly the graphics. Now have a great understanding of ISO and call it what you will. I like sensitivity and who cares. Keep doing these videos.
For me it helps to think as each pixel that records light can have an arbitary value from 1 to 100. In good light, pixels will have any value from 1 to 100 giving it the maximum possible DR for that sensor. In poor light, the pixels might only be recording light in the range 1 to 10. We therefore need to apply gain to the signal in order to fully utilise the recording potential for each pixel. In this case we can apply a gain factor of 10. The signal received in the range of 1 to 10 will then become a signal between 10 to 100. We therefore have a reduced DR in the range of 10-100 rather than the 1-100 in good light. This assumes that 1 is the lowest recordable light intensity. Anything below this is 0. Applying gain to 0 still results in 0 (pure black).
Last week I photographed a White tail Deer after sunset at iso 25,600 and the histogram was ETR on a Z8. I was surprised how good it was, After running it through Lightroom denoise it was decent, the light was flat but considering what I would have ended up with using my D300 and the software we had then I'd say we are lucky to live in the times we do.
Very interesting video, I shoot with a fairly old Pentax K3, I do some auroa and astrophotography. I’ve often wondered if I would see an increase in image quality upgrading. Looking at the charts on Photonstophotos shows there will be some improvements, not sure how much though. Thanks.
I have over 50 years of photography experience but never knew increasing ISO effects the darkest to the lightest rather than the entire spectrum equally. I am not sure how or if this new knowledge will effect my photography since I typically us Manual with Auto-ISO these days ...but I am pretty sure I will be thinking about it next trip out.
One again 'simple' but simply wrong. You don't get away with being wrong by claiming it's not 'high level'. ISO is not really 'sensitivity'. Once again your muddling the concept of exposure, using it both to mean amount of light (the correct meaning) and how light or dark the image looks - you talk about 'creating the same exposure' by using a faster film. No. A faster film produces the same density for a lower exposure. You really do need to work out what 'exposure' means. It is after all teh core concept in photography, and untill you get it straight your videos on this topic will be hopelessly muddled. Your discussion of noise and ISO is, I'm afraid, completely garbled. SO is not at heart a gain. The whole idea of 'boosting' is false and leads to wrong conceptualisation. All that is happening is that the image is being rendered lighter - as in a different LUT. You don't need 'boosting' to render something lighter. Increasing the gain does not cause noise. You're using the wrong photon to photos graph. You need the 'input referred read noise'. The one you are showing is not 'very close to the noise you'll see in the photo'. It's the read noise, which is typically insignificant except in very low light in the shadows. Most of the noise that you see is shot noise, not read noise, and as before - to see the visble effect of read noise you need the 'input referred read noise', not the DN read noise. That graph is not showing what you say it is. Sorry, this is a twenty-minute word salad, You don't understand this subject and you need to do a lot of learning before you try to teach it.
I fully agree. ISO is definitely not the sensitivity of a CMOS sensor. If it is that complicated to make a clear distinction between quantity of incoming light and image brightness, why even try to say anything about ISO and dynamic range? The commentary around the photons to photos graphs didn't improve anything. Wrong charts, wrong interpretations. With the selected graphs, the ISO could very well have been just a number written as metadata in the raw file (not saying it is - i didn't check, but that's one of the 4 ways ISO can be handled). In any case, the read noise chart is indeed irrelevant - is it that difficult to see that the numbers are so low that they don't matter?
It's not 'negativity', it's simply factual correction. Who needs it are people actually trying to learn some photography. They don't know to go 'somewhere else' if nobody lets them know that this channel is teaching them stuff that's wrong.
@@BobN54to suggest that folks can't learn from this channel is complete BS, and you know this. again, if you don't like the content, then go away. you're not some amazing gatekeeper of knowledge. in fact, if you knew anything, you'd try to help others, rather than bringing them down.
@@marshalhedin7538 I'm sure that lots of the content here is good, but the fact of it is that these videos on DR contain very basic mistakes. If people learn from them they will learn stuff that is simply false. They will learn a wrong definition of exposure. They will be unable to conceptualise the difference between exposure and lightness/value. They will not know what ISO does. They will not know where the noise in a photo comes from and how to manage it. We haven't got to DR yet, but what we've had so far displays the conceptualisation difficulties that ensue from these basic mistakes. It's not surprising that Scott has made these mistakes, because the internet is awash with them, and unless someone knows where to get correct information on photography they will continue to promulgate these errors and degrade the knowledge base of photography. That's why I comment when I see mistakes like this being made - it has nothing to do with 'bringing people down'. I get a variety of responses from teh creators, some are interested in finding out where they went wrong, and end up improving their videos and providing a better service. Some just continue putting out the errors. I had put Scott in the second case, but I see there has been a more positive response to Bill Ferris' comment, making the same points. Bill is more diplomatic than I am. If Scott takes time to learn the correct basics his videos will improve and the people he teaches will benefit from it.
I think than the « gain » works in 2 ways… boosting in low light… but reduces light in high light? Where is the neutral , less interference bring by processors to the signal? Minimal ISO (64-100)… or somewhat around 400? 800? (Referring to Nikon Z6-8-9 camera)
Not sure that many people realized that the DR decreases significantly when using lens at F8 or more. Softwares manage the noise quite well… but can not recreate the DR (dynamic range)
Scott, as usual you do a great job making complicated concepts easier to understand. You are a very good teacher and I really appreciate everything that you do. Even though others have pointed out that they believe there are technical inaccuracies in the video, I don't feel that they are able to articulate their position in a manner that is easy to understand or practical to apply in the field. It is great that we have public forum like this where everyone can chime in, yet I rarely see anyone who can make things as easily understandable as you do. And that is what makes you a great teacher. I look forward to the rest in this series!
"Amplification" That's the word I've been looking for!! Perfect explanation, much appreciated.
The sound example was great to explain how noise increases . The concept of Dynamic range coming down with the increase in ISO also was extremely well explained.
Scott, I appreciate the time, energy and effort you put into educating your audience. That said, you're presentation in these first two videos is filled with misinformation.
The foundational error is treating exposure as image lightness. That's not what exposure is. Exposure has been known since the late 1800s to be the light energy per unit area projected by lens upon the glass plate, film, or sensor.
As such, exposure is determined by the available light in the scene, lens aperture (f-stop) and exposure time (shutter speed). ISO isn't an exposure setting; never has been.
ISO also does not amplify signal. It's not a source of noise. In fact, it's the light used to make the photo that is the primary source of noise in photos. It's called photon noise or shot noise.
In simple terms, noise is randomness. If you put small water cups on a table on your deck to collect water during a rainstorm, you'd find the amount of rainwater in each cup will be similar but not exactly the same. That randomness in the amount of water in each pup is similar to noise in photography.
If you photograph an evenly lit wall, the pixels on the camera sensor would detect a similar number of photons. They'd be similar but not exactly the same. The randomness in the number of photons each pixel receives is noise.
The more total light collected by the sensor, the less visible noise will be. That's because the amount of light - the signal - overwhelms any randomness - noise - in how much light each pixel receives.
The less total light used to make a photo, the less signal there is. This reduced signal allows noise to be more easily seen.
ISO's role in all this is pretty simple. It's the setting we use to manage the lightness of the photo made by the camera. I say, lightness, to distinguish an ISO setting from the exposure (light per unit area of the sensor) used. Adjustments to exposure alter brightness - an amount of light - while adjustments to ISO alter image lightness or tonality from black to white.
When ISO is increased to compensate for a limited amount of light being available, the resulting photo maintains a pleasing lightness *but* that increased lightness reveals shot noise that is present in the image. ISO doesn't create or add noise. It reveals noise that's present in the photo.
There's a second kind of noise that can affect digital photos. This is noise that's a product of image and data processing in the camera. It's called read noise. Read noise is someone mistakenly referred to as ISO noise. The thinking goes something like this, "As ISO increases, the in-camera processing introduces more read noise which is what we see at high ISOs."
If only that were true. In fact, read noise historically gets lower as exposure decreases and ISO increases. A decade or more ago, the standard practice in digital photography was to use the highest ISO needed for a given exposure. If too low an ISO was used, the photo would need to be lightened in Photoshop. That process would reveal the greater read noise associated with the lower ISO.
Today, dual gain sensors have become fairly common. The advantage of this technology is that photos made at lower ISOs enjoy the same low read noise as photos made at higher ISOs. As a result, if you choose ISO 800 in the field to avoid blowing out white plumage in a bird and then lighten the darker plumage in post to bring out detail in the bird, the noise revealed is no worse than the read noise associated with ISO 1600 or whichever higher ISO you'd normally use to bring out detail in dark plumage in that lighting.
The bottom line of all this for photographers is that f-stop and shutter speed are the critical settings. They control exposure and two important creative aspects of a photo: depth of field and the rendering of movement.
If you choose the largest aperture (smallest f-number) that delivers an acceptable depth of field and the slowest shutter speed that adequately renders movement without blowing out important highlights, you'll optimize exposure within the creative goals you have for the image.
After that, select an ISO that results in a photo having a pleasing lightness.
In candor, ISO is a rather trivial setting. It's certainly not as important as the fundamentally-flawed exposure triangle would have us believe. It doesn't control exposure or add noise. It's how we manage image lightness.
I read this and appreciate your thoughts. Would love to see an video referenced that explains this easier and more accurately as I have read dozens of articles and watches numerous videos and perhaps my explaination is based on others information. But i have tried to make this concept easy for beginners and not sure who to do this easily for a beginner. The basic point was was to give an idea of what noise was and why we need to boost the signal to create our final output (when light isn't getting to the sensor). Also to show the DR decreases as ISO increases which should be factually accurate. Also appreciate your critique. Thanks.
So, like Scott said, increasing ISO will increase noise visible in the image and reduce dynamic range. Correct?
@@2224jnk ISO communicates information the camera uses to process an image. Increasing ISO results in the camera processing a lighter version of the photo and that will make noise that's already present in the image easier to see.
ISO isn't a source of noise.
I like the analogy with sound. Easy to understand.
Nice job Scott! You hit the nail on the head with Digital ISO and used a good analogy of the term GAIN. Comparing this to audio amplifiers is perfect. An amplifier with poor signal to noise ratio sounds terrible when one turns up the volume, hence noise compared to one with a great S/N ratio. This is true for a cameras sensor and the amplifier circuits used to process the signal. Look forward to you next video on dual gain sensors.
thanks! I generally get hammered but a couple electrical engineer types who tell me it's all wrong, but don't seem to understand that I am making video about general concepts and not advanced videos on circuitry and electrical signals lol
As per your dynamic range video loved it and particularly the graphics. Now have a great understanding of ISO and call it what you will. I like sensitivity and who cares. Keep doing these videos.
Yes the video really did make sense! Thanks for making it easy on my brain!
For me it helps to think as each pixel that records light can have an arbitary value from 1 to 100. In good light, pixels will have any value from 1 to 100 giving it the maximum possible DR for that sensor. In poor light, the pixels might only be recording light in the range 1 to 10. We therefore need to apply gain to the signal in order to fully utilise the recording potential for each pixel. In this case we can apply a gain factor of 10. The signal received in the range of 1 to 10 will then become a signal between 10 to 100. We therefore have a reduced DR in the range of 10-100 rather than the 1-100 in good light.
This assumes that 1 is the lowest recordable light intensity. Anything below this is 0. Applying gain to 0 still results in 0 (pure black).
Live your videos. Always good info. Thanks!
Last week I photographed a White tail Deer after sunset at iso 25,600 and the histogram was ETR on a Z8. I was surprised how good it was, After running it through Lightroom denoise it was decent, the light was flat but considering what I would have ended up with using my D300 and the software we had then I'd say we are lucky to live in the times we do.
Very interesting video, I shoot with a fairly old Pentax K3, I do some auroa and astrophotography. I’ve often wondered if I would see an increase in image quality upgrading. Looking at the charts on Photonstophotos shows there will be some improvements, not sure how much though. Thanks.
I have over 50 years of photography experience but never knew increasing ISO effects the darkest to the lightest rather than the entire spectrum equally. I am not sure how or if this new knowledge will effect my photography since I typically us Manual with Auto-ISO these days ...but I am pretty sure I will be thinking about it next trip out.
Yes, I agree. I think the dynamic range goes down even when I add light to an underexposed photo in post-processing. Or am I mistaken?
With all the variable options of base aperture settings ( F1.4 or F2, F2.8 etc ) , what is considered a "Full Stop" ?
Standard stops would be 2, 2.8. 4. 5.6, 8, 11
Wish you had some camo ducks are assholes hats
One again 'simple' but simply wrong. You don't get away with being wrong by claiming it's not 'high level'. ISO is not really 'sensitivity'. Once again your muddling the concept of exposure, using it both to mean amount of light (the correct meaning) and how light or dark the image looks - you talk about 'creating the same exposure' by using a faster film. No. A faster film produces the same density for a lower exposure. You really do need to work out what 'exposure' means. It is after all teh core concept in photography, and untill you get it straight your videos on this topic will be hopelessly muddled.
Your discussion of noise and ISO is, I'm afraid, completely garbled. SO is not at heart a gain. The whole idea of 'boosting' is false and leads to wrong conceptualisation. All that is happening is that the image is being rendered lighter - as in a different LUT. You don't need 'boosting' to render something lighter. Increasing the gain does not cause noise. You're using the wrong photon to photos graph. You need the 'input referred read noise'. The one you are showing is not 'very close to the noise you'll see in the photo'. It's the read noise, which is typically insignificant except in very low light in the shadows. Most of the noise that you see is shot noise, not read noise, and as before - to see the visble effect of read noise you need the 'input referred read noise', not the DN read noise. That graph is not showing what you say it is.
Sorry, this is a twenty-minute word salad, You don't understand this subject and you need to do a lot of learning before you try to teach it.
I fully agree. ISO is definitely not the sensitivity of a CMOS sensor. If it is that complicated to make a clear distinction between quantity of incoming light and image brightness, why even try to say anything about ISO and dynamic range?
The commentary around the photons to photos graphs didn't improve anything. Wrong charts, wrong interpretations. With the selected graphs, the ISO could very well have been just a number written as metadata in the raw file (not saying it is - i didn't check, but that's one of the 4 ways ISO can be handled). In any case, the read noise chart is indeed irrelevant - is it that difficult to see that the numbers are so low that they don't matter?
if you guys want something else, just go elsewhere. we don't need your negativity ..
It's not 'negativity', it's simply factual correction. Who needs it are people actually trying to learn some photography. They don't know to go 'somewhere else' if nobody lets them know that this channel is teaching them stuff that's wrong.
@@BobN54to suggest that folks can't learn from this channel is complete BS, and you know this. again, if you don't like the content, then go away. you're not some amazing gatekeeper of knowledge. in fact, if you knew anything, you'd try to help others, rather than bringing them down.
@@marshalhedin7538 I'm sure that lots of the content here is good, but the fact of it is that these videos on DR contain very basic mistakes. If people learn from them they will learn stuff that is simply false. They will learn a wrong definition of exposure. They will be unable to conceptualise the difference between exposure and lightness/value. They will not know what ISO does. They will not know where the noise in a photo comes from and how to manage it. We haven't got to DR yet, but what we've had so far displays the conceptualisation difficulties that ensue from these basic mistakes. It's not surprising that Scott has made these mistakes, because the internet is awash with them, and unless someone knows where to get correct information on photography they will continue to promulgate these errors and degrade the knowledge base of photography. That's why I comment when I see mistakes like this being made - it has nothing to do with 'bringing people down'. I get a variety of responses from teh creators, some are interested in finding out where they went wrong, and end up improving their videos and providing a better service. Some just continue putting out the errors. I had put Scott in the second case, but I see there has been a more positive response to Bill Ferris' comment, making the same points. Bill is more diplomatic than I am. If Scott takes time to learn the correct basics his videos will improve and the people he teaches will benefit from it.
I think than the « gain » works in 2 ways… boosting in low light… but reduces light in high light? Where is the neutral , less interference bring by processors to the signal? Minimal ISO (64-100)… or somewhat around 400? 800? (Referring to Nikon Z6-8-9 camera)
Not sure that many people realized that the DR decreases significantly when using lens at F8 or more. Softwares manage the noise quite well… but can not recreate the DR (dynamic range)