Are Fuji JPGs deterministic?

Take multiple frames of the same scene with the same lens, same settings.  Are the resulting .jpg images the same?  Or do they vary in color/contrast/whatever?

While doing some testing shortly after getting my X-T3 I seem to recall that I got variation between jpg frames but the raw files when processed with the same setting in CaptureOne were consistent.

I have not tried to repeat the experiment but may if I get a chance some day.  Just wondering what you have seen.

kdarwin-photo wrote:

Take multiple frames of the same scene with the same lens, same settings. Are the resulting .jpg images the same? Or do they vary in color/contrast/whatever?

While doing some testing shortly after getting my X-T3 I seem to recall that I got variation between jpg frames but the raw files when processed with the same setting in CaptureOne were consistent.

I have not tried to repeat the experiment but may if I get a chance some day. Just wondering what you have seen.

Unless you were shooting with some sort of bracketing mode or Auto-DR mode, if the RAWs look consistent, the jpegs should be consistent too (but not identical).

orlovsn wrote:

this is normal for every camera, to keep files identical you need external spot metering and manual white balance

Even spot metering is not going to give you the same look in all the files.

The only way to absolutely guarantee that the photos are identical is to be shooting a static lab test scene where the camera is on a tripod (locked in position), the lights are controlled, and the camera is in completely manual mode for all settings.

Of course, no one is shooting that.

What do you mean by the same?  If you mean if you loaded two such jpegs into a computer and subtracted one from the other the results would be zero - no. First there are too many variables.  While the lighting might be close it will not be the same.  The exposure and WB - unless set manually are recalculated on each shot.  Then there is noise that is independent from shot to shot. Not even two raw images are exactly pixel to pixel the same.   Jpeg is a compressed format and slight differences will result in slight differences in the compression.

If you mean that the output images will appear for all practical purposes to a viewer at a normal viewing distance they should look the same.

When you load the raws and process in Capture One, LR, etc. using common settings for all shots, WB, exposure adjustments, etc., a few variables are removed and jpegs exported might be closer but I don't expect anyone would notice it.

https://www.merriam-webster.com/dictionary/deterministic

'jus sayin'

Chris Dubea wrote:

https://www.merriam-webster.com/dictionary/deterministic

'jus sayin'

wrong link. Should point here:

Chris Dubea wrote:

https://www.merriam-webster.com/dictionary/deterministic

'jus sayin'

That’s not exactly the definition used in this context - this might be better - we’re not discussing philosophy

https://en.wikipedia.org/wiki/Deterministic_algorithm

Chris Dubea wrote:

https://www.merriam-webster.com/dictionary/deterministic

'jus sayin'

I suppose the grammar lessons might be marginally useful to some folks here, but I really think the OP was pretty clear on what information he was looking for.

Might I gently suggest that we keep the discussion focused on the question posed in the first part of his post rather than turning this thread into a Grammar 101 lesson, please?

Thanks in advance.

Greybeard2017 wrote:

Chris Dubea wrote:

https://www.merriam-webster.com/dictionary/deterministic

'jus sayin'

That’s not exactly the definition used in this context - this might be better - we’re not discussing philosophy

https://en.wikipedia.org/wiki/Deterministic_algorithm

Based on that definition I'd say that the output from Fuji (and every other camera manufacturer) is deterministic.  Given the exact same input you will get the same output.

The problem?  You will never get the exact same input so you really can't prove it one way or the other.

michaeladawson wrote:

Greybeard2017 wrote:

Chris Dubea wrote:

https://www.merriam-webster.com/dictionary/deterministic

'jus sayin'

That’s not exactly the definition used in this context - this might be better - we’re not discussing philosophy

https://en.wikipedia.org/wiki/Deterministic_algorithm

Based on that definition I'd say that the output from Fuji (and every other camera manufacturer) is deterministic. Given the exact same input you will get the same output.

The problem? You will never get the exact same input so you really can't prove it one way or the other.

In reality a camera sensor does not capture the same values at for multiple shots of the same scene because there is randomness involved in the data - photon shot noise and electrical noise. Clearly the exact same numbers on the post ADC side will be processed the same. One the other hand the same light energy profile illuminating the detectors will not result in identical numbers on the post ADC side.

Chris Dubea wrote:

https://www.merriam-webster.com/dictionary/deterministic

'jus sayin'

While that is well and good and it is nice that some can use a dictionary, the concept of determinism in physics was a mistake. It started to become well established starting with Einstein’s explanation of the photoelectric effect, that determinism was not a property of the real world.

https://link.springer.com/article/10.1140/epjh/s13129-021-00012-x

Truman Prevatt wrote:

michaeladawson wrote:

Greybeard2017 wrote:

Chris Dubea wrote:

https://www.merriam-webster.com/dictionary/deterministic

'jus sayin'

That’s not exactly the definition used in this context - this might be better - we’re not discussing philosophy

https://en.wikipedia.org/wiki/Deterministic_algorithm

Based on that definition I'd say that the output from Fuji (and every other camera manufacturer) is deterministic. Given the exact same input you will get the same output.

The problem? You will never get the exact same input so you really can't prove it one way or the other.

In reality a camera sensor does not capture the same values at for multiple shots of the same scene because there is randomness involved in the data - photon shot noise and electrical noise. Clearly the exact same numbers on the post ADC side will be processed the same. One the other hand the same light energy profile illuminating the detectors will not result in identical numbers on the post ADC side.

That's all well and good, but I don't think the OP was probably looking for that level of nuance. His RAWs were consistent enough with the same processing, but his SOOC jpegs were not. Auto-WB, Auto DR, or a bracketing mode could affect the SOOC jpegs significantly, but not necessarily the RAWs with identical development settings.

Truman Prevatt wrote:

Chris Dubea wrote:

https://www.merriam-webster.com/dictionary/deterministic

'jus sayin'

While that is well and good and it is nice that some can use a dictionary, the concept of determinism in physics was a mistake. It started to become well established starting with Einstein’s explanation of the photoelectric effect, that determinism was not a property of the real world.

https://link.springer.com/article/10.1140/epjh/s13129-021-00012-x

How can light be both a wave and a particle in what seems to be both at the same time. Still boggles my mind every time I think about it.

The universe is full of mysterious paradoxes waiting to be discovered, revealed, solved, questioned, only to reveal more questions…

Erik Baumgartner wrote:

Truman Prevatt wrote:

michaeladawson wrote:

Greybeard2017 wrote:

Chris Dubea wrote:

https://www.merriam-webster.com/dictionary/deterministic

'jus sayin'

That’s not exactly the definition used in this context - this might be better - we’re not discussing philosophy

https://en.wikipedia.org/wiki/Deterministic_algorithm

Based on that definition I'd say that the output from Fuji (and every other camera manufacturer) is deterministic. Given the exact same input you will get the same output.

The problem? You will never get the exact same input so you really can't prove it one way or the other.

In reality a camera sensor does not capture the same values at for multiple shots of the same scene because there is randomness involved in the data - photon shot noise and electrical noise. Clearly the exact same numbers on the post ADC side will be processed the same. One the other hand the same light energy profile illuminating the detectors will not result in identical numbers on the post ADC side.

That's all well and good, but I don't think the OP was probably looking for that level of nuance. His RAWs were consistent enough with the same processing, but his SOOC jpegs were not. Auto-WB, Auto DR, or a bracketing mode could affect the SOOC jpegs significantly, but not necessarily the RAWs with identical development settings.

I’m not really sure we know what he is asking since the question was - same scene, same lens, same settings (which means auto WB, DR, sharping, etc) are the jpegs identical. If everything is identical- the only difference is result to the fact the lighting was not exactly the same or randomness.

Clearly is one image in camera was generated with Auto DR and the next with auto DR off then the out of camera would likely be different but the raw could be processed to produce comparable jpeg exports. However, that is not the same settings.

kdarwin-photo wrote:

OK, OP back here - I noticed it while testing some lenses. I'm pretty sure I was set to AWB and shooting aperture priority mode. What I remember seeing is that the raw files on several shots all looked the same but there was noticeable variability between the jpgs. Enough that it could make me think the lens had color shift at different apertures. Yet those raws when given the same processing recipe all looked the same.

My 1/2 baked conclusion is that AWB may vary a bit and if I want to really compare one lens to another I had better be looking only at the raw files.

If you use the "as shot" WB values, you should see any AWB differences with the RAWs as well.

I have not tired to duplicate the results.

Pocket Lint wrote:

Truman Prevatt wrote:

Chris Dubea wrote:

https://www.merriam-webster.com/dictionary/deterministic

'jus sayin'

While that is well and good and it is nice that some can use a dictionary, the concept of determinism in physics was a mistake. It started to become well established starting with Einstein’s explanation of the photoelectric effect, that determinism was not a property of the real world.

https://link.springer.com/article/10.1140/epjh/s13129-021-00012-x

How can light be both a wave and a particle in what seems to be both at the same time. Still boggles my mind every time I think about it.

The universe is full of mysterious paradoxes waiting to be discovered, revealed, solved, questioned, only to reveal more questions…

The discovery around black body radiation - bid goodbye to the concept of "pure waves." The discovery explanation verified many times which won Einstein his Nobel - was the beginning of quantum mechanics. The world is not comprised of two dichotomies - one being particles ( a big set of marbles in a bag ) and waves. For example the electron which has mass - also acts like a wave.

Even the discovery of electromagnetic radiation started the process of ripping the bandage of the of a deterministic physical universe. How can something be a wave - when there is nothing to "wave in."

The problem is really more having to do with people trying to related concepts to things that are common to them. On the other hand Richard Feynman was I think correct. "Their "paradox" is only a conflict between reality and your feeling of what reality "ought to be." Or maybe Neils Bohr is the right reference, "Everything we call real is made up of things that cannot be regarded as real."

What quantum mechanics tells us is "nothing exist until it is measured!"

https://theconversation.com/explainer-what-is-wave-particle-duality-7414

kdarwin-photo wrote:

OK, OP back here - I noticed it while testing some lenses. I'm pretty sure I was set to AWB and shooting aperture priority mode. What I remember seeing is that the raw files on several shots all looked the same but there was noticeable variability between the jpgs. Enough that it could make me think the lens had color shift at different apertures. Yet those raws when given the same processing recipe all looked the same.

My 1/2 baked conclusion is that AWB may vary a bit and if I want to really compare one lens to another I had better be looking only at the raw files.

I have not tired to duplicate the results.

Thanks for the clarification.  Minor changes in light can change how AWB modifies the scene and could in fact change the in camera jpeg a little.  For example if the sun goes behind a cloud the scene is not lite by more skylight (blue) than white light. AWB would compensate for that.

On the other hand if you shoot AWB Capture one will open the Raw and apply the settings you had in the camera.  You can change them and the Raw is not modified in camera because of this but when they are opened, the camera settings will be used as default including film simulation.

Your conclusions are correct, if you want to compare lenses, look only at the Raws but also applying the same settings in Capture One.