/p/ - Photography » Thread #3526035

[33 / 6 / ?]

1MiB, 1313x909, fake_iso.png

View Same Google ImgOps iqdb SauceNAO

Anonymous Mon 26 Aug 2019 00:16:00 No.3526035 View View Reply Original Report

Quoted By: >>3526036 >>3526046 >>3526436

>ISO is meaningless

I can't believe I never realized this.

Anonymous

Anonymous Mon 26 Aug 2019 00:20:31 No.3526036 Report

Quoted By: >>3526115

>>3526035
It always frustrates me when you try to explain this to people. The only difference has been analogue v digital amplifiers and now there is little difference so yes ISO (on decent digital cameras) is only for JPEG rendering. I found that when I had to explain it to finally resolve the issue you cant use academic papers or electrical engineering just take a few pictures and get people to guess. Then if they are not to obstinate they will finally believe you.

Anonymous

View Same Google ImgOps iqdb SauceNAO dualISO_diagram3.jpg, 201KiB, 755x428

Anonymous Mon 26 Aug 2019 00:43:39 No.3526046 Report

Quoted By: >>3526048

>>3526035
He jumped the shark with that one. He is somewhat correct on some subatomic level he's not even gasping. But he's clearly out of the depth and incorrect on so many levels that he is laughable.

Anonymous

Anonymous Mon 26 Aug 2019 00:50:21 No.3526048 Report

Quoted By: >>3526063

>>3526046
isn't this chart showing the same thing he's saying?

Anonymous

Anonymous Mon 26 Aug 2019 01:34:35 No.3526063 Report

Quoted By: >>3526089

>>3526048
On the contrary, It's showing what iso does. If iso was meaningless shadows/highlights values would be flat. But this one is for filming, analogue iso gains are calibrated separately for filming and for photography. And it's causing all sorts of problem since people are equalling them. Bringing definitions and charts from one to the discussion of the other, and even using techniques independently.

Anonymous

Anonymous Mon 26 Aug 2019 02:43:31 No.3526089 Report

Quoted By:

>>3526063
that chart is saying more about the tonal range that is captured vs ISO, but the video seems to be talking about noise vs iso. I wouldn't say those are the same thing.

Anonymous

Anonymous Mon 26 Aug 2019 03:47:25 No.3526115 Report

Quoted By: >>3526235

>>3526036
Not entirely correct.
It depends on the sensor and its related electronics.
There are ISO variant and ISO invariant sensors.

If the signal gets to the analog/digital converter without amplification, you are indeed correct, the sensor is ISO invariant.

If there is an analog amplifier in use before the analog/digital converter it has a better signal/noise ratio than not using the analog amplifier and multiplying after the analog/digital converter.
This sensor is ISO variant, using higher ISO natively results in a better signal/noise ratio than cranking it up in post.

Anonymous

Anonymous Mon 26 Aug 2019 04:46:25 No.3526132 Report

Quoted By:

these guys are fucking retards and i cannot believe the sheer amount of people who follow them

Anonymous

Anonymous Mon 26 Aug 2019 12:21:28 No.3526235 Report

Quoted By: >>3526260

>>3526115
ISO invariance is old meme from early 2015's. A bunch of pixel peepers jumped on a marketing slogan and found out that you gain nothing in raw at high isos, so they put a name to it. The talk was about signal to noise ceiling of analogue converter. Never that iso does nothing, or that you have actually iso variant and invariant sensor! All sensors are iso invariant to the digital amplification levels. Some more so, and some less. Some documented and some less so. In general you've base iso's at 200 extended up to 50 that do nothing, and real invariance range up to 6400. After that it's just digital gain and irrelevant. At least they were smart enough to know what this means, purely s/n at pixel level, not colour!
Now every month a new blogposter will find it and repeat all the nonsense again. Make his own little test, and add few falsehood into it as well. It's particularly popular since you can see the effects yourself™. And when it finally trickles down to clickbait youtouber shitposters, you get ISO DOESN'T MATTER! At this time it's all bullshit, and shouldn't even be acknowledged.

Anonymous

Anonymous Mon 26 Aug 2019 13:08:14 No.3526260 Report

Quoted By:

>>3526235
The main issue is that amplification at the analog stage results in a better signal/noise ratio, yet manufacturers use this in different ways.
From what I hear some only crank up analog amplification once ("dual ISO") and just multipy by 2 for every stop of higher ISO in digital while others crank it up in increments at lower ISO and resort to digital multiplication later.

Anonymous

Anonymous Mon 26 Aug 2019 13:21:22 No.3526268 Report

Quoted By: >>3526309

He is trying to tell you to use Auto ISO.

It's for the best, and it's the best setting for Jpeg.
You won't get a better result using shooting ISO 100 in Raw and recovering the shadows.

Hence Auto ISO is always just as good, or even better in case of Jpeg.

Anonymous

Anonymous Mon 26 Aug 2019 15:05:46 No.3526309 Report

Quoted By:

>>3526268
or not, if it means clipping highlights. he's indirectly suggesting you shoot at a lower iso to underexpose the image, or at least figure out where the noise level sweet spots are for your camera

Anonymous

Anonymous Mon 26 Aug 2019 15:32:23 No.3526319 Report

Quoted By:

Did I understand correctly that the compromise is dynamic range (underexposure) versus color depth and clean highlights (ETTR)?

Anonymous

Anonymous Mon 26 Aug 2019 16:14:20 No.3526333 Report

Quoted By:

>pro photographer
>literally never taken a good photo in his life.

Anonymous

Anonymous Mon 26 Aug 2019 20:40:13 No.3526436 Report

Quoted By: >>3526854

>>3526035
>ISO is meaningless
yes, no and maybe... if...

This whole thing came about after some engineering nerds and astro geeks did some experiments a few years ago. It has nothing to do with everyday photography.

The original blogs and youtube videos were about iso invariant cameras and how to test for the optimal iso of specific cameras for image stacking.

The misconception is that iso doesn't matter, you can just increase the exposure in post with an iso invariant camera. This does not take into account what is happening to the signal after it has left the sensor.

The whole point was that (certain) cameras have a sweet spot in iso performance, this varies from camera to camera. There are simple tests you can do to determine this for your camera model.

All this only matters if you are photographing very dark/faint objects and want to rape the sliders after stacking dozens of shots.

Anonymous

Anonymous Tue 27 Aug 2019 08:55:08 No.3526602 Report

Quoted By:

aftet watching his followup video, id say it isn't meaningless, but rather arbitrary and abused, and the mechanisms of action vary across cameras

Anonymous

View Same Google ImgOps iqdb SauceNAO otU7pYO.png, 93KiB, 640x452

Anonymous Tue 27 Aug 2019 11:51:12 No.3526645 Report

Quoted By: >>3526652

wait wait wait just to get up to speed on this whole concept, is he imblyin that if you shoot RAW you can just crank the gain in post and get the same result as shooting with higher ISO in the first place? i do nothing but night tography and ISO has been one of the key challenges to try and balance, and messes up a lot of shots once you get home and see it in proper resolution.

so, what do? shoot everything at 100 and just fix in poast?

Anonymous

Anonymous Tue 27 Aug 2019 12:26:55 No.3526652 Report

Quoted By: >>3526659

>>3526645
He intitialy said that ISO doesn't matter at all and you could essentialy just crank it up in post from base to maximum.
That's bullshit indeed as no camera I know of is truly ISO invariant.
I mean sure, you can adjust exposure by 1-2 stops in post just fine, but the build in analog amplifier in the camera can do that with less signal to noise ratio.
That's why ETTR even became a thing.

Anonymous

Anonymous Tue 27 Aug 2019 12:38:32 No.3526659 Report

Quoted By: >>3526670 >>3526671

>>3526652
ok so adjusting everything - including ISO - so the exposure is 1 stop or so higher than the developed product will be is the best way to get the most out of a digital sensor then? so really once I get my aperture and shutter the way I want them for the shot, I just jimmy the ISO to be slightly brighter than I want the final product and I'll have post-production gold? This all makes a lot of sense but honestly I've never really consciously done it. I might just leave the exposure compensation at +1 for my nek outting just to see how it works out.

...course I'll probably end up having to fuck around more with the metering mode as well. sigh I miss Auto sometimes

Anonymous

Anonymous Tue 27 Aug 2019 13:06:31 No.3526670 Report

Quoted By: >>3526674 >>3526676

>>3526659
It only works if you don't clip the highlights and it isn't doing that much either.
But it helps especialy with the darker portions of your image as more precise information can be stored in the "brighter" part of your sensors dynamic range.

Also:
What digital camera has no automatic exposure modes available?

Anonymous

Anonymous Tue 27 Aug 2019 13:07:24 No.3526671 Report

Quoted By: >>3526674

>>3526659
depends on the iso/noise profile of your camera. shoot at the nearest iso sweet spot and then move sliders in post

Anonymous

Anonymous Tue 27 Aug 2019 13:30:42 No.3526674 Report

Quoted By: >>3526678 >>3526864

>>3526670
My camera has Auto and it's actually quite clever (just not appropriate for night or special effect shots of course). And yeah If I manage to not blow out the highlights I might try the ETTR approach just to see how gravy it is in post.

>>3526671
>iso sweet spot
Do you mean the ISO that accomplishes the right exposure, the exposure +1, or just the most efficient ISO without discernible noise? Sorry if I sound like a doofus this is just interesting to me and I've been stuck in my old habits a long time and am looking to try a fresh approach

Anonymous

View Same Google ImgOps iqdb SauceNAO ZA7SDR2-L.png, 25KiB, 702x392

Anonymous Tue 27 Aug 2019 13:38:08 No.3526676 Report

Quoted By: >>3526680

>>3526670
>But it helps especialy with the darker portions of your image as more precise information can be stored in the "brighter" part of your sensors dynamic range.
But you get greater dynamic range at lower ISO

Anonymous

Anonymous Tue 27 Aug 2019 13:47:07 No.3526678 Report

Quoted By: >>3526684 >>3526880

>>3526674
>Do you mean the ISO that accomplishes the right exposure, the exposure +1, or just the most efficient ISO without discernible noise?
dude watch the followup video he made he explains it god damn

certain isos will have less noise on certain cameras and the sensor becomes iso invariant

every camera is different so you have to look at the noise profile for your camera

Anonymous

Anonymous Tue 27 Aug 2019 14:01:21 No.3526680 Report

Quoted By:

>>3526676
It's complicated.
The human eye and film work in a logarythmic way while the sensor of a digital camera works in a linear way.
Each stop of exposure level holds twice as much brightness information as the one before it because of that
If you now move up everything by one stop by using higher ISO, slower shutter speed or larger aperture, you get more detailed brightness information in the shaddows.
In addition to that the analog amplifier used before the analog/digital converter has a higher signal/noise ratio to work with than the the converter itself, meaning you get a cleaner signal.
Especialy with ISO variant sensors that mainly use an analog amplifier to change ISO this can yield good results.

Anonymous

View Same Google ImgOps iqdb SauceNAO img-thing_mb.jpg, 23KiB, 300x300

Anonymous Tue 27 Aug 2019 14:10:10 No.3526684 Report

Quoted By:

>>3526678
okay anon okay

thanks for that

Anonymous

Anonymous Tue 27 Aug 2019 20:32:55 No.3526854 Report

Quoted By: >>3526927

>>3526436
What tests?

Anonymous

Anonymous Tue 27 Aug 2019 21:02:57 No.3526864 Report

Quoted By: >>3526869

>>3526674
Cranking up the iso can be significantly cleaner but also with less dynamic range compared to pushing an underexposed shot to the same exposure.
If your exposure is constant you'll end up with less noise the more you lower the iso as you have to compensate with a longer exposure or wider aperture and thus collect more light and amplify less.

For clean night shots you want to expose as long as possible, with the widest aperture possible and rely on higher iso speed if motion or dof limit result in underexposure. With a tripod the cheapest and most crappy old ass camera is a low light genious.

Anonymous

Anonymous Tue 27 Aug 2019 21:07:26 No.3526869 Report

Quoted By:

>>3526864
>For clean night shots you want to expose as long as possible, with the widest aperture possible and rely on higher iso speed *ONLY* if motion or dof limit result in underexposure.
For better understanding

Anonymous

View Same Google ImgOps iqdb SauceNAO 1563636167725.png, 120KiB, 1000x1000

Anonymous Tue 27 Aug 2019 21:34:27 No.3526880 Report

Quoted By:

>>3526678
>watch the followup video he explains it
yes be a nice goy and watch nice tony nothburger explain every stupidity he made in the first video in the second one!

Anonymous

Anonymous Tue 27 Aug 2019 21:53:54 No.3526894 Report

Quoted By:

Is this yet another stupid discussion like dslr vs mirrorless or theres some reasoning for the range of gear we use, i mean, not only for NASA telescopes and such
Because im interested

Anonymous

Anonymous Tue 27 Aug 2019 23:03:50 No.3526927 Report

Quoted By: >>3529394

>>3526854
>What tests?
Take 7 exposures. For example:
100 iso -3 stops
200 iso -2 stops
400 iso -1 stop
800 iso correctly exposed
1600 iso +1 stop
3200 iso +2 stops
6400 iso +3 stops
Then in post dial the exposure so they all match. If your camera is iso variant the images will look different. If the images look the same your camera is somewhat iso invariant.

If your camera is iso variant it's probably best to just use the iso that has noise you are willing to accept.

If your camera is iso invariant examine the images carefully for noise. The one with the least noise (you'll have to look closely) will be the "best" for processing. The figure could be anything from 200-1600 iso.

In a true iso less camera you can shoot at any iso, they are rare.

If you want to go further, gain can be added in a non linear pattern so check out half and third iso settings.

The website: http://www.photonstophotos.net/ has a lot of data for most modern cameras. The s/n ratio being useful. It gives a guide to where analogue and digital gain are applied.

There was a youtube video i can't find at the moment but here's a couple of links you may find useful:
https://www.lonelyspeck.com/how-to-find-the-best-iso-for-astrophotography-dynamic-range-and-noise/
https://www.skyandtelescope.com/astronomy-blogs/imaging-foundations-richard-wright/choose-iso-astrophotography/

tl;dr It really doesn't matter that much, I wasted days to go "meh".

Anonymous

Anonymous Mon 02 Sep 2019 01:03:14 No.3529394 Report

Quoted By:

>>3526927
You'll be getting wrong colours, and throwing metering modes and auto white balance out of equation. It's only viable for astrophotography since you're already doing false colours because of various factors. There is use for ettr, but again only in certain situations with full knowledge of what it does. Otherwise the smart cam manufacturers are smarter than you, and you should trust them.

Capcode	All Only User Posts Only Verified Posts Only Moderator Posts Only Manager Posts Only Admin Posts Only Developer Posts Only Founder Posts
Show Posts	All Only With Images Only Without Images Only Spoiler Images Only Non-Spoiler Images
Deleted Posts	All Only Deleted Posts Only Non-Deleted Posts
Ghost Posts	All Only Ghost Posts Only Non-Ghost Posts
Post Type	All Only Sticky Threads Only Opening Posts Only Reply Posts
Results	All Grouped By Threads
Order	Latest Posts First Oldest Posts First

On these archives

Your latest searches