Home Photography ISO is totally FAKE. Seriously.

ISO is totally FAKE. Seriously.

10
42
Advertisement canpers.com

Check out our #1 PHOTOGRAPHY BOOK on Amazon: http://help.tc/s OUR STORE: http://sdp.io/sdpbook coupon ‘YouTube’ ISO is fake for a bunch of reasons: 1) …

source

Facebook Comments
Advertisement canpers.com

42 COMMENTS

  1. High ISO is a post-processing gain from the sensor, but it is different from post-processing after saving the photo. In dark condition, you have to scale up the brightness before saving the photo, because you always have limited bit depth. That is the same story why log scale is better for video recording.

  2. Good thing I don't use digital anymore then, and along with digital "thinking" I also shed the precise or nitpicky mentality. – I mean, I would've been thinking that ISO is indeed a standardized value across all cameras and counted on that. – But now that I switched to film some years ago, which made me adopt a much more "raw" or "rough" mentality in terms of exposure and such, the digital ISO standard (or lack thereof) doesn't apply to me. I just go by rough estimations of ISO-rating of different films now and just make sure I expose properly (still on an estimation) and then also develop carefully. – Whatever the result is will have to do for me and if the image isn't to my liking I'll have to do better. – I grew tired of digital precision, made me feel like a robot anyway, so good riddance…

  3. Tony, check out Dave McKeegan's video on ISO variance. Not so clear cut and he offers some proof. Sorry he says you are only about half correct. Deeper analysis and explanation makes some scientific sense to explain difference. ISO when far over 800 is different than setting an ISO at 100 then boosting in post to match an extreme ISO (100 iso gains much more noise and potential banding when boosting to the extreme due to how signal is sent through sensor and circuitry, noise is amplified).

  4. Dude this is so helpful… knowing I can just let set the ISO to auto eliminates one whole step from setup. Thanks! Btw I was wondering about this before even seeing that this video existed lol guess I had my suspicions.

  5. Are not these sensors (like) photomultipliers whose sensitivity (threshold to relase electrons upong reception of photons) can be boosted by changing the voltage that they are subjected to? This is important conceptually, a higher ISO is not just a digital multiplication, but an actual physical change in the sensor as some have commented. In practice I think the statements in the video are still true for the most part. ISO across brands is not fully comparable and the advertised highest ISO is a joke.

  6. Tony, can you do the same for shots per second? Camera companies don't give you a realistic number under normal circumstances. Sometimes it's just overstated by 15% like on the GH5. Maybe a common practice.

  7. I saw your another video where you explained T-stops. In the part where you compare 3 camera ISO brightness, did you chceck T-stop? Because problem might be in lenses.

  8. Tony did you choose F4 or F2 in Olympus X-T3 camera? You know very well that if you choose F4 on m43 sensors, that's actually F2 in a full frame camera. I'm not sure, i haven't seen most of your videos but one thing that i've noticed is that you are always trying to diminish Olympus. As i said, i haven't seen all your videos, i may have the wrong perception, but from what i've seen so far, you always try to find a way to trash Olympus products when comparing them to other brands. I'm really sorry if i have the wrong perception. Cheers!

  9. I took this to the test and compared two pictures I took indoor in a dark area of the house. One at ISO 800 to get a correct exposure and one at ISO 100 which was way too dark. Put them into Lightroom and bumped up the exposure. After putting them next to eachother, I notice no difference at all.

  10. I don't usually like reviews from Tony but this is fantastic, needed to be said and clearly explained. Totally agree with him and he has debunked so many myths. Well done Tony. I suppose that I should get my Weston Master V out again…..

  11. Best proof of this are the red cameras with the r3d footage format. The r3d raw format allows you to change iso in post and chose what iso value you want for the raw footage. So they actually embraced this fully. Why i do not have this option for my cr2 files in lightroom already? Maybe soon everyone will switch to "dynamic iso in post".

  12. Am I the only one that took out 2 different brands camera and test them at the same setting?

    Anyway, regarding the physical change for ISO. High ISO drains your battery power faster, it's not something you see physically because everything is happening on the sensor, but it did happen.

    Your explanation is more of a "conspiracy theory" or rather personal opinion

  13. When companies boast a high iso, its not just a meaningless number. it means that their engineers have devised a method for deciding what the acceptable signal to noise ratio is for that particular camera, for example they decide that they want a minimum dynamic range of 5ev. Then they amplify the signal until they reach that signal to noise ratio/dynamic range. The better the sensor is at reducing noise, the higher will it be able to be amplified before reaching that set signal to noise ratio/dynamic range of 5.

  14. Does anyone remember the German DIN values like 21°, 24°, etc. for ISO 100, 200, etc.? They are essentially dB values where 3db means half (resp. double) the light sensibility. It starts with DIN 1= ISO 1 and the ISO doubles with every 3 DIN (or dB). So 1 is the common ground. It wouldn't hurt to make use of those figures again and name it something like sdB or dBs (sensitivity in dB) – though nobody will understand it for about 5 years 😁

  15. Tony, this video is not accurate, although i am sure you had fun making it. Ifyou want to standardize on iso, then there are other factors like the brightness of a lens, that can vary by even a stop with different lenses at the same aperature and focal length.

    ISO does give us a workable relative tool, as well as experiential understanding of the dynamic range we can pull from our camera.
    it's kind of like how many steps to walk across a room. We each have different size legs, but we all understand that 100 steps is roughly 10x the distance of ten steps.

    A sensor cell is like a well, that fills up with so many quanta, say 60000 electrons knocked loose by light.
    This is translated to a voltage that is read by an analog to digital converter. The adc, for simplicity of illustration, has a maximum voltage it measures, and a baseline voltage. so for example a 16 bit adc can measure the voltage to ~1 part in 65000.

    Now, if only enough electrons were knocked looses to create a voltage that is 1/1000 of the peak voltage the adc will measure, that means that it can only measure 1 part in 6, about 2 stops. to boost the voltage, it is amplified by the gain, i.e., your iso setting before the adc, so it can benefit from the full benefit of the adc. That is exposing to the right. too much and we are overloading the adc, at least until people can figure out how to make cells that reset themselves when filled up and keep track of how many times it has done this. The more accurate, and greater range the adc has, the more iso invariant the sensor is, along with how clean the amplifier is, leakage, and the various kinds of noise.

    There is also dual (and possibly triple base iso tuned sensors in the future),
    prior to the amplification gain, such as from sony.
    The above is greatly simplistic, however it points out that iso is a physical modification to the image capture, mostly prior to digitization. Just like changing the aperature, you cannot, at least today, reverse it after the pic is taken onto your sd card – that is it is not an overlay.

    The fact that you can pull 5 stops from an underexposed image is a testament to the greatly improved sensors, as well as some smarts built into sensors to adjust things a bit.

  16. Interesting. I always thought high ISO sensors were more sensitive to light, just like high ASA film used to be. But it's still not completely bogus as a standard if it measures the amount of gain that can be applied before a pre-determined level of noise comes into the image. I recall from my old wet processing days in the 70s that grain got worse and worse the higher the film ISO or the more you pushed it in the lab. It's almost the same thing! Push the developing time and temperature to bring more out of film, or increase the electronic gain to get more out of the sensor image, both resulting in more noise. If the camera makers could standardise on the acceptable level of noise they should be able to come up with consistent and comparable ISO ratings for their sensors.

  17. Regarding the fact that "photography uses standard from the dark ages", this is not exclusive to photography. It's true for anything. You don't change a standard and way to represent things without a great reason. Even our DNA is like that: still having many useless parts from another time. And it applies also for ideas and theories: the quantum revolution has been very hard to accept at its time, and even the founders of it (Planck, Einstein) were not believing in it. Planck was very pragmatic and used that new approach as a convenient mathematical tool, thinking it has no reality outside of maths. He also stated that new ideas are not accepted by established people, but they become more and more popular to the new generation and gets accepted as the olders are passing away…

    Many others pointed it out, the ISO invariance is not true with all manufacturers. You used a Nikon to illustrate it, by luck or convenience, but try on a Canon, you'd have a hard time with your explanation. But that doesn't mean your global ISO topic is not true. It is. Which also mean you sometimes have to find the "best" ISO of a new camera (when you don't have that ISO invariance). I remember someone stating that ISO 125 was the actual "base" ISO on a Canon 5D Mark III.

  18. Does ISO adjustment happen in the amplifier, Analog/Digital Converter or Image Signal Processor? If ISO adjustment happens after the amplifier then I think it would essentially be post processing. Unless, in the ISP, higher ISO preserves information from the darker image areas that would otherwise be discarded in compression. Otherwise it seems to me ISO doesn’t do anything that can’t be done in post processing.

  19. Amen, Tony. the "ISO" have now become too large to be useful. Note that each "stop" doubles the "ISO" number. That's why the numbers get too large to make sens — pretty much above about "6400".

    on an exponential scale like this an exponential notation is better– and that is what db is. I think cine cameras are already using this with "ASA 400" as db 0 and then 6 db steps from there — ASA 800 would be +6db, ASA 1600 would be +12 db etc.

    6 is a good step value as 1/3 stop = 2 db and 1/2 stop = 3 db.

    another old term that need to go in the dumpster is "camera roll". I like to use "Events" and with sub "Sets" allowed.

  20. Thanks Tony! That was one part of digital photography that I didn't understand. As an engineer I can totally relate to your feelings about this. And yes, it's ridiculous that we can't set a lower ISO to avoid using ND filters.

  21. Great, informative, and useful. We really appreciate getting true information and explanations!

    The whole world is tired of being fed Fake anything, Please keep giving us the True facts and solid information.

  22. The act or process of 'Assimilating', or of absorbing information, experiences, etc.: Physiology & Ergonomically the Manufactures are advancing an Art Form… Computational Photography is NOW! All that is required is a pin hole camera with a small sensor and a super computer and batteries… I could not help but notice that Olympus M1x …(2) Computers and (3) Batteries???

Leave a Reply