S1RII High ISO Raw Noise Comparison

[letyphotographer]

Lumix S1RII vs Sony A7RIII

DPR Studio Test Scene Raw files Downloaded and converted in RawTherapee 5.11 (the Beast)

Same settings for both photos :

51200 ISO, Tungsten white balance 2856 Degrees Kelvin, no exposure change

Custom tone curve Film Alike mode ( visually similar to DPR Test images regard contrast )

Default Capture Sharpening (a kind of up stream/native sharpening in RT) with no visible impact in details for these photos

Lens profiles applied (quite stretchy for the Lumix), maximum jpg quality, crop needed for size limit

SCD+VNG4 demosaic method (good balance details/noise digital artifact)

Absolutely NO NOISE REDUCTION , because RT is a neutral developer and you have full control of all parameters

S1RII







A7RIII







In my personal view probably there is something going on at matrix level in both raw files, but I don't see any difference in microcontrast, eventually the raw baking is at the same amount. If you choice different demosaic method (LMMSE for example) the image could be more grainy.

Sony is slightly less noisy, but the scene maybe is more illuminated.



Sorry for the grammar but I am not an english speaker

Cheers

 

[Interceptor121]

Lumix S1RII vs Sony A7RIII

DPR Studio Test Scene Raw files Downloaded and converted in RawTherapee 5.11 (the Beast)

Same settings for both photos :

51200 ISO, Tungsten white balance 2856 Degrees Kelvin, no exposure change

Custom tone curve Film Alike mode ( visually similar to DPR Test images regard contrast )

Default Capture Sharpening (a kind of up stream/native sharpening in RT) with no visible impact in details for these photos

Lens profiles applied (quite stretchy for the Lumix), maximum jpg quality, crop needed for size limit

SCD+VNG4 demosaic method (good balance details/noise digital artifact)

Absolutely NO NOISE REDUCTION , because RT is a neutral developer and you have full control of all parameters

S1RII



A7RIII



In my personal view probably there is something going on at matrix level in both raw files, but I don't see any difference in microcontrast, eventually the raw baking is at the same amount. If you choice different demosaic method (LMMSE for example) the image could be more grainy.

Sony is slightly less noisy, but the scene maybe is more illuminated.

Sorry for the grammar but I am not an english speaker

Cheers

At ISO 51200 both camera have noise reduction on the raw files

Compare the files between 6400 and 25600 there you will see differences

--
If you like my image I would appreciate if you follow me on social media
instagram http://instagram.com/interceptor121
My flickr sets http://www.flickr.com/photos/interceptor121/
Youtube channel http://www.youtube.com/interceptor121
Underwater Photo and Video Blog http://interceptor121.com
If you want to get in touch don't send me a PM rather contact me directly at my website/social media

 

[NAwlins Contrarian]

Lumix S1RII vs Sony A7RIII

* * *

Same settings for both photos :

51200 ISO, Tungsten white balance 2856 Degrees Kelvin, no exposure change

* * *

Sony is slightly less noisy, but the scene maybe is more illuminated.

I appreciate the effort to provide a reasonable, objective comparison, but there are at least two reasons why this test is not valid:

(1) the effective exposures (i.e., exposure relative to 'gain' settings) are not the same and

(2) both cameras are applying noise reduction to the so-called raw file, i.e., they are not close to raw.

The effective exposures are not the same

Just looking at your two files shows to the naked eye that the Sony one is brighter than the Panasonic one. Maybe the "ISO 51200" setting on the Sony is really ISO 48000 and on the Panasonic it's really ISO 40000, or whatever. DxO tests this and reports the results. I'm not aware of any other scientific organization that does. Unfortunately DxO has not published test results for any camera launched since November 2023. Anyway, I cropped out very similar, near-white sections from each of the files you posted and stacked them right next to each other:



Then I cropped out from each of those samples a 35x35 pixel section 175 pixels from the left edge and abutting the bottom of the crop; using GIMP 3.0.2's cubic resampler, sized them down to 1x1 pixel; and checked the color of each. In GIMP's color picker's LCh panel, the L value of the Panasonic is 49/100, and the L value of the Sony is 53/100. In other words, the Sony image is measurably brighter.

'Raw' noise reduction invalidated the comparison

Here you can look at the Photos to Photos tests for each camera, compared:

https://www.photonstophotos.net/Charts/PDR.htm#Panasonic Lumix DC-S1RM2,Sony ILCE-7RM3

and see that the S1R II starts applying noise reduction (indicated where the circle is replaced by the downward-pointing triangle) to its 'raw' (at that point, not truly raw!) files at its ISO 6400 setting, and the A7R III does so starting at its ISO 32000 setting. So at ISO 51200, you are comparing the noise reduction each camera is applying, not their respective sensor performances. You can't compare their respective sensors' performances past their ISO 5000 settings (and even there, you'd have the issue raised above).

 

[letyphotographer]

If you read my previous post :

" In my personal view probably there is something going on at matrix level in both raw files "

" Sony is slightly less noisy, but the scene maybe is more illuminated. "

I was aware of the fact but I didn't replayed because I was thinking nobody was interested in the thread.

Looking at the luminance istograms and at the scene, there are dfferences in the bright/dark distribution areas but the shooting parameters are strictly equal.

So maybe during DPR relocation headquarter (due to Amazon selling) the lighting set-up was slightly changed , or the lens transmittance is different.

In any case , ISO photo value is a jpg only affaire and you have to manual fine tune the files exposure in post when the shooting settings , and the in camera analog/scaled ISO step, are identical.

By the way, this is the next S1RII Vs A7RIII raw comparison :

ISO 25600, exactly where the S1RII on sensor denoising has a bump in the Bill Claff PDR chart, and the scenario is probably the worst.

Plus 0,2 EV in S1RII file and mid to high value changes in the film-alike tone curve , for brighting shadow and to exacerbate noise structure.

The other settings are equals.

S1RII







A7RIII







The more smeared details in Lumix image than Sony is clearly visibile, in particular if you look at the pointy zone on the backside playcards , and looking at the mid shadow small letters in the banknote , completely disappeared.

But the raw denoise is not anyway too much exaggerated IMHO.

The S1RII sensor is probably noisier than the A7RIII ones, but the Lumix choice is maybe correlated to the tipically compressed S series raw files.

As a matter of fact S1RII file size is over 20 MB less than Sony not compressed ones. A more smooth signal could provides aid to the Lumix algorithm for saving space in the memory card and in the buffer , but this is only my remote ipothesis.

More to come...



Sorry for the grammar, I am not an english speaker

Cheers

 

[letyphotographer]

Hi guys, this is my third S1RII IQ post, and my settimg choice this time is:

12800 ISO S1RII vs A7RIII

A quite interesting parallel , because in Bill Claff PDR chart the dynamic range spread between the two cameras is smaller than 25600 ISO and remains approximately the same down to 6400 ISO, where starts high ISO on sensor noise reduction in Lumix camera.

All other parameters stay equal to the past 25600 ISO S1RII vs A7RIII comparison.

How the Panasonic image processing will affect the fine elements rendering?

Well, let's see it in these new Rawtherapee developed images !

S1RII 12800 ISO



A7RIII 12800 ISO



Looking at points in playcards blue back ( sort of similar to random noise ), at watermark and small letters in banknote shadow, you could check a subtle flattering in details contrast , but to a lesser degree than 25600 situation.The Lumix vs Sony differences are visible only in extreme pixel peeping and you always have 16 MB size benefit in the memory card ( though in other circumstance size benefit could be greater , even 20-30 MB, see DPR ISO Invariance Test at -6 EV )

Also, bearing in mind many camera brands are involved in raw NR at 50 000 ISO and above ( sometimes below ), so the real IQ threat in Lumix camera is only at 25600 ISO.

Not a big concern in my personal view.

More to come..

Cheers