A7rIV vs A7rV noise levels?

[ilumo]

Hi all. To prevent muddying the other a7rV thread this question-

i have read in various places that the r5 DR is slightly better than the r4s. But then I see in dpreviews image comparison that the r5 high iso actually looks a bit noisier. Am I seeing things or is that the case? And that higher DR is not equal to high iso noise control ?

Just wondering as I had the r4 before and just ordered the r5. The geek side of me is teching out.

 

[MILC man]

those cameras have the same sensor.

 

[No account anymore]

Hi all. To prevent muddying the other a7rV thread this question-

i have read in various places that the r5 DR is slightly better than the r4s. But then I see in dpreviews image comparison that the r5 high iso actually looks a bit noisier.

Such small variations from the same sensor might well be a result from the process of making these test photos (variation in exposure, light level, etc.) or sample variation.

Am I seeing things or is that the case? And that higher DR is not equal to high iso noise control ?

We often see what we want to see. How would such minor differences, igf they are real and more noticeable than sample variations) affect your real world photograpy?

• How often do you expose optimised, and then use data close to the noise floor?
• How much would a 0.3 stop or less noise difference affect images that are NOT directly compared to identical images taken with a different camera?
• And, would a 0.3 stop difference matter? Would it be noticeable?

If you shoot challenging subjects, I would rather go for the best autofocus instead of having those almost sharp frames. This would do much more for your real world results that the DR and/or noise level difference between these cameras.

Just wondering as I had the r4 before and just ordered the r5. The geek side of me is teching out.

If you answer the questions above, my guess is that your inner tech geek won't freak out.

 

[Nokiron]

Hi all. To prevent muddying the other a7rV thread this question-

i have read in various places that the r5 DR is slightly better than the r4s. But then I see in dpreviews image comparison that the r5 high iso actually looks a bit noisier. Am I seeing things or is that the case? And that higher DR is not equal to high iso noise control ?

Just wondering as I had the r4 before and just ordered the r5. The geek side of me is teching out.

The difference is negligible if not within variation. Don't worry.

Read Noise in DNs versus ISO Setting

Photographic Dynamic Range versus ISO Setting

 

[No account anymore]

Hi all. To prevent muddying the other a7rV thread this question-

i have read in various places that the r5 DR is slightly better than the r4s. But then I see in dpreviews image comparison that the r5 high iso actually looks a bit noisier. Am I seeing things or is that the case? And that higher DR is not equal to high iso noise control ?

Just wondering as I had the r4 before and just ordered the r5. The geek side of me is teching out.

The difference is negligible if not within variation. Don't worry.

https://www.photonstophotos.net/Charts/RN_ADU.htm#Sony ILCE-7RM4_14,Sony ILCE-7RM5_14

https://www.photonstophotos.net/Charts/PDR.htm#Sony ILCE-7RM4,Sony ILCE-7RM5

Translated to real world use: About identical.

 

[Nokiron]

Hi all. To prevent muddying the other a7rV thread this question-

i have read in various places that the r5 DR is slightly better than the r4s. But then I see in dpreviews image comparison that the r5 high iso actually looks a bit noisier. Am I seeing things or is that the case? And that higher DR is not equal to high iso noise control ?

Just wondering as I had the r4 before and just ordered the r5. The geek side of me is teching out.

The difference is negligible if not within variation. Don't worry.

https://www.photonstophotos.net/Charts/RN_ADU.htm#Sony ILCE-7RM4_14,Sony ILCE-7RM5_14

https://www.photonstophotos.net/Charts/PDR.htm#Sony ILCE-7RM4,Sony ILCE-7RM5

Translated to real world use: About identical.

Absolutely.

 

[Djehuty]

Practically the same but if you bought the more expensive one you'd magically see the difference.

 

[pollup]

I was posting something ... and then realized that the DPR comparison tool is flawed, because some elements of the scene have been altered!

 

[Nokiron]

I was posting something ... and then realized that the DPR comparison tool is flawed, because some elements of the scene have been altered!

I assume you posted a comparison of the colorchecker?

Those are recommended to be replaced every two years since the color quality degrades over time. So unfortunately they cant keep them forever.

 

[pollup]

I was posting something ... and then realized that the DPR comparison tool is flawed, because some elements of the scene have been altered!

I assume you posted a comparison of the colorchecker?

Those are recommended to be replaced every two years since the color quality degrades over time. So unfortunately they cant keep them forever.

Indeed it was.

I understand the need for replacing things, but ideally DPR would just reshoot the scene with every single camera whenever they replace something.

The hair is also different, also understandably so because it keeps moving. There's also an old image that looks worse and worse over time, there must be some color degradation happening there too, and it hasn't been replaced.

 

[Nokiron]

I was posting something ... and then realized that the DPR comparison tool is flawed, because some elements of the scene have been altered!

I assume you posted a comparison of the colorchecker?

Those are recommended to be replaced every two years since the color quality degrades over time. So unfortunately they cant keep them forever.

Indeed it was.

I understand the need for replacing things, but ideally DPR would just reshoot the scene with every single camera whenever they replace something.

The hair is also different, also understandably so because it keeps moving. There's also an old image that looks worse and worse over time, there must be some color degradation happening there too, and it hasn't been replaced.

Actually, I very much doubt that DPReview have many cameras available at the office at the same time. The industry standard is getting loaners so keeping something like this test scene up-to-date is incredibly hard. Just imagine keeping it as clean as possible and not have dust confused for noise and not disturbing the layout.

To get lab-quality consistency you need to be a company like DXOMark, that provides consultancy, software, services and use of their lab for companies to use. Outside of that I think this test-scene DPReview has is the best we got.

 

[MattPointZero]

Indeed it was.

I understand the need for replacing things, but ideally DPR would just reshoot the scene with every single camera whenever they replace something.

Not likely. Can you imagine the work involved with that??? Also - There's at least 300 cameras listed, they don't keep them all sat there in storage I am pretty sure!

The DPR thing is kind of useful, I go to it a fair bit - but you have to remember how flawed it is as a precise tool especially for sharpness and detail - it very much matters what lens you are shooting with as to outcome, the sony R series are shot with the 85GM, not the sharpest lens, the Nikon Z7 is shot with a non-native mount through an FTZ adapter! These things must have a bearing, so I look at it, but try to remember its limitations

 

[pollup]

Indeed it was.

I understand the need for replacing things, but ideally DPR would just reshoot the scene with every single camera whenever they replace something.

Not likely. Can you imagine the work involved with that??? Also - There's at least 300 cameras listed, they don't keep them all sat there in storage I am pretty sure!

294 cameras in the list.

I said "ideally", if not possible then redoing all the mirrorless is another good option, and again if not possible then redoing all the mirrorless from 2018+ is a good option.

That way, the number of cameras will be greatly reduced.

The DPR thing is kind of useful, I go to it a fair bit - but you have to remember how flawed it is as a precise tool especially for sharpness and detail - it very much matters what lens you are shooting with as to outcome, the sony R series are shot with the 85GM, not the sharpest lens, the Nikon Z7 is shot with a non-native mount through an FTZ adapter! These things must have a bearing, so I look at it, but try to remember its limitations

 

[MattPointZero]

Indeed it was.

I understand the need for replacing things, but ideally DPR would just reshoot the scene with every single camera whenever they replace something.

Not likely. Can you imagine the work involved with that??? Also - There's at least 300 cameras listed, they don't keep them all sat there in storage I am pretty sure!

294 cameras in the list.

I said "ideally", if not possible then redoing all the mirrorless is another good option, and again if not possible then redoing all the mirrorless from 2018+ is a good option.

That way, the number of cameras will be greatly reduced.

The DPR thing is kind of useful, I go to it a fair bit - but you have to remember how flawed it is as a precise tool especially for sharpness and detail - it very much matters what lens you are shooting with as to outcome, the sony R series are shot with the 85GM, not the sharpest lens, the Nikon Z7 is shot with a non-native mount through an FTZ adapter! These things must have a bearing, so I look at it, but try to remember its limitations

Did you count them? That is some commitment to pedantry...

Are you imagining that they buy every single one of these cameras plus some lenses and then stack them on shelves once they shoot the scene??? I am pretty sure that isn't how it works!

 

[ilumo]

those cameras have the same sensor.

Yes. But there is some difference in the processing I’m sure. Otherwise bclaff wouldn’t have measured the increase in DR via his testing. That said I noticed (subjectively) more color noise in the image comparison tool. Regardless, I’m going to be keeping the a7rV for the vastly improved AF/ handling changes over the a7riv.
this was just more for my personal edification (justification?)

 

[ilumo]

Hi all. To prevent muddying the other a7rV thread this question-

i have read in various places that the r5 DR is slightly better than the r4s. But then I see in dpreviews image comparison that the r5 high iso actually looks a bit noisier.

Such small variations from the same sensor might well be a result from the process of making these test photos (variation in exposure, light level, etc.) or sample variation.

Am I seeing things or is that the case? And that higher DR is not equal to high iso noise control ?

We often see what we want to see. How would such minor differences, igf they are real and more noticeable than sample variations) affect your real world photograpy?

• How often do you expose optimised, and then use data close to the noise floor?
• How much would a 0.3 stop or less noise difference affect images that are NOT directly compared to identical images taken with a different camera?
• And, would a 0.3 stop difference matter? Would it be noticeable?

If you shoot challenging subjects, I would rather go for the best autofocus instead of having those almost sharp frames. This would do much more for your real world results that the DR and/or noise level difference between these cameras.

Just wondering as I had the r4 before and just ordered the r5. The geek side of me is teching out.

If you answer the questions above, my guess is that your inner tech geek won't freak out.

I actually shoot a lot of 800-1600 iso shots. A lot of indoor event and portrait photography in non ideal lighting. So high iso performance in general is important but I agree that .2 stops is probably not noticeable. I came from canon so the banding was infinitely worse than the “grain” noise of the Sonys.
definitely not freaking out. Just curious.

 

[MILC man]

those cameras have the same sensor.

Yes. But there is some difference in the processing I’m sure. Otherwise bclaff wouldn’t have measured the increase in DR via his testing.

i suspect that miniscule difference is easily within the range of error; see here for an example of how different pdr measurements for the same sensor in two different cameras can be wrong:

"Bill’s PDR tests don’t take sensitivity into account. My tests show that, at base ISO (64 for X2D, 100 for GFX 100S), the two cameras are within 0.2 stops of having the same sensitivity. Once you correct for that, Bill’s curves for the two cameras are nearly on top of each other.

So the Fuji Rumors claim is at best misleading, and at worst just plain wrong. C’mon folks, these cameras have the same sensors, albeit operated slightly differently. How far apart can the DR be?"

Bill Claff’s PDR ratings for the X2D and GFX 100S - the last word (kasson.com)

That said I noticed (subjectively) more color noise in the image comparison tool.

with what, jpg? or raw that of course uses different profiles? did they use the same lens for both tests? etc.

Regardless, I’m going to be keeping the a7rV for the vastly improved AF/ handling changes over the a7riv.

that's more than enough justification, the a7rv is a hugely improved camera over it's predecessor, just like the a9ii was over the a9, despite having the same sensor.

 

[AlephNull]

Hi all. To prevent muddying the other a7rV thread this question-

i have read in various places that the r5 DR is slightly better than the r4s. But then I see in dpreviews image comparison that the r5 high iso actually looks a bit noisier. Am I seeing things or is that the case? And that higher DR is not equal to high iso noise control ?

Just wondering as I had the r4 before and just ordered the r5. The geek side of me is teching out.

Photons to Photos shows slightly less noise - that's a more scientific measurement than the DP Review. The difference is small, and it's entirely possible that DP Review measured a cleaner-than-normal A7RIV and a noisier-than-normal A7RV.

I'd expect it to be slightly less noisy, because the A7RV is designed to keep the sensor a bit cooler (so it can shoot 8k video) and heat is the source of thermal noise.

 

[ilumo]

those cameras have the same sensor.

Yes. But there is some difference in the processing I’m sure. Otherwise bclaff wouldn’t have measured the increase in DR via his testing.

i suspect that miniscule difference is easily within the range of error; see here for an example of how different pdr measurements for the same sensor in two different cameras can be wrong:

"Bill’s PDR tests don’t take sensitivity into account. My tests show that, at base ISO (64 for X2D, 100 for GFX 100S), the two cameras are within 0.2 stops of having the same sensitivity. Once you correct for that, Bill’s curves for the two cameras are nearly on top of each other.

So the Fuji Rumors claim is at best misleading, and at worst just plain wrong. C’mon folks, these cameras have the same sensors, albeit operated slightly differently. How far apart can the DR be?"

Bill Claff’s PDR ratings for the X2D and GFX 100S - the last word (kasson.com)

That said I noticed (subjectively) more color noise in the image comparison tool.

with what, jpg? or raw that of course uses different profiles? did they use the same lens for both tests? etc.

Regardless, I’m going to be keeping the a7rV for the vastly improved AF/ handling changes over the a7riv.

that's more than enough justification, the a7rv is a hugely improved camera over it's predecessor, just like the a9ii was over the a9, despite having the same sensor.

I used the default. See attached screenshot from my phone. I see noticeably more color noise, but I suppose this could be from a multitude of factors. That said I’m not sure what the comparison tool is actually comparing (jpg vs raw) if it’s raw, it also seems there might be differences between the two files. The r4 raw is much bigger. So is one of them using uncompressed vs compressed raw?

 

[unners]

One thing that caught me out with the A7R5 is I was processing with Lightroom as I hadn't upgraded to the new version of Capture One. I was coming from an A7RIV and then a GFX100s and the new A7RV seemed quite noisy. Turns out the default noise reduction in Capture One made quite a big difference compared to Lightroom and once I opened the A7RV files in CO23, the noise seemed basically the same as my previous A7RIV. Obviously the GFX100s was slightly better than both.