Canon R7 bit depth on electronic shutter

[dankenny]

Just thinking about whether to use EFCS or full Electronic shutter for dynamic range and colour critical landscape work.

I've read that the R5's electronic shutter mode yielded less bit depth than other modes. Does anyone know if this is also the case with the R7?

Happy to perform any tests if anyone can suggest what to do.

Thanks in advance.

 

[SigZero]

Just thinking about whether to use EFCS or full Electronic shutter for dynamic range and colour critical landscape work.

I've read that the R5's electronic shutter mode yielded less bit depth than other modes. Does anyone know if this is also the case with the R7?

Happy to perform any tests if anyone can suggest what to do.

Thanks in advance.

If we believe the specs it seems to be 14 bit. In the same place in specs, where R5 states only 12 bit there is nothing about it in R7 specs.

Not easy to check as even with R5 all 14 bits are used in RAW file due to some digital processing.

Probably some DR test should reveal if this is true 14 bits in R7.

Br, Pawel.

 

[bodeswell]

Just thinking about whether to use EFCS or full Electronic shutter for dynamic range and colour critical landscape work.

I've read that the R5's electronic shutter mode yielded less bit depth than other modes. Does anyone know if this is also the case with the R7?

Happy to perform any tests if anyone can suggest what to do.

Thanks in advance.

This is an interesting question because the answer, which one might suppose would be plainly documented, is so hard to find. When Bryan Carnathan published his review of the R5 a couple of years ago he said this in his detailed specs page:

"RAW: RAW, C-RAW 14 bit (14-bit with Mechanical shutter and Electronic 1st Curtain, 13-bit A/D conversion with H+ mode, 12-bit A/D conversion with Electronic shutter)"

Canon EOS R5 Specifications (the-digital-picture.com)

Unfortunately, he did not cite a source and the actual Canon detailed specifications PDF does not say quite the same thing. It does say that for H+ 13 bit A/D conversion applies. See

EOSR5_specifications_FINAL_JULY14.pdf (canon.com)

Googling around reveals that various old timers believe Canon "always" uses 14 bits since time immemorial, but they don't cite sources.

There is an interesting thread on FM:

R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE - FM Forums (fredmiranda.com) See especially the posts from Bruce n Philly, who actually got Canon to reply to him. The Canon reply is at once clear and obscure. It seems to be an implementation detail that they believe the public need not worry about.

The corresponding documents for the R7, both Carnathan's and Canon's, are silent on this point.

I ran exiftool against an R7 .cr3 picked at random and it reported a "bit depth" of 24, which I take to mean eight bits per color.

Carnathan probably had some basis for his numbers, so I would be inclined to go with them in the absence of better documentation. They get some support from the FM thread. I would be inclined to guess that the same numbers apply to the R7.

I have no idea how one would test this at home or if that is even possible.

The obscurity of this issue suggests to me that it doesn't really matter.

 

[Ephemeris]

Our friends at Photons to Photons have some good date for the R5. So we can take solice.

For the R7 I think just a single batch of data. I know he is active in this group so maybe he can chip in. Find his posts really interesting.

 

[bclaff]

...

I've read that the R5's electronic shutter mode yielded less bit depth than other modes. Does anyone know if this is also the case with the R7?

Happy to perform any tests if anyone can suggest what to do.

...

This is easily determined by examining the raw data.
If you don't have the appropriate tools contact me via Private Message (PM) and we'll sort it out.

 

[Ephemeris]

Hey Bill.

Do you have any info on the bit depth question regarding the R7?

I read your website when you published the R7 data but I wondered if this changes with type of shutter?

I'm not a video knowledgeable like many here but would the data change across the many video options these cameras may offer?

 

[dankenny]

Thanks all for thinking with me on this. Messaging you now Bill - legend!

 

[bclaff]

Hey Bill.

Do you have any info on the bit depth question regarding the R7?

I read your website when you published the R7 data but I wondered if this changes with type of shutter?

I'm not a video knowledgeable like many here but would the data change across the many video options these cameras may offer?

Normally, unless it says otherwise, the measurements are with mechanical shutter assuming one is available.

 

[PhotosFlight]

The R3,5,6,7 all use Digic X. As such I'd expect they would behave the same way unless the firmware can change this across bodies. But it looks like others have definitively sorted this out.

 

[SigZero]

The R3,5,6,7 all use Digic X. As such I'd expect they would behave the same way unless the firmware can change this across bodies. But it looks like others have definitively sorted this out.

I think that when Bill will get correct files form R7 ES everything will be clear. Right now I don't think we have clear evidence whether R7 uses 14 or 12 bits in ES.

Br, Pawel

 

[Ephemeris]

Hey Bill.

Do you have any info on the bit depth question regarding the R7?

I read your website when you published the R7 data but I wondered if this changes with type of shutter?

I'm not a video knowledgeable like many here but would the data change across the many video options these cameras may offer?

Normally, unless it says otherwise, the measurements are with mechanical shutter assuming one is available.

 

[bclaff]

I received 3 files which is sufficient to do a quick initial test.
The files were black frames taken with mechanical, EFCS, and electronic shutters.

From the histogram I would say that electronic is 12-bit with some small amount of smoothing (I haven't done spectral analysis yet).
Here's a chart


Best viewed "original size"

And the read noise measurements are not surprising to me, electronic has higher read noise which would mean lower maximum dynamic range at low ISO settings.
Here's a quick chart


Best viewed "original size"

Published is the value that is already published at PhotonsToPhotos.

Looks like similar if not the same amount of signal processing (noise reduction) independent of shutter.

Exaggerated 2D Fourier Transfornms:

[ATTACH alt="Must view "original size" "]3151954[/ATTACH]
Must view "original size"



--
Bill ( Your trusted source for independent sensor data at PhotonsToPhotos )

 

[Ephemeris]

I received 3 files which is sufficient to do a quick initial test.
The files were black frames taken with mechanical, EFCS, and electronic shutters.

From the histogram I would say that electronic is 12-bit with some small amount of smoothing (I haven't done spectral analysis yet).
Here's a chart


Best viewed "original size"

And the read noise measurements are not surprising to me, electronic has higher read noise which would mean lower maximum dynamic range at low ISO settings.
Here's a quick chart


Best viewed "original size"

Published is the value that is already published at PhotonsToPhotos.

Looks like similar if not the same amount of signal processing (noise reduction) independent of shutter.

Exaggerated 2D Fourier Transfornms:


Must view "original size"

--
Bill ( Your trusted source for independent sensor data at PhotonsToPhotos )

Thanks for your efforts bill.

 

[SigZero]

Thank Bill as always.

So we have here exactly the same situation as in R5/R6. The difference is that Canon does not confess this in any part of the specs of R7

So no hope for firmware upgrade to R5 giving 14 bits at least in one shot...

Br, Pawel

 

[PDE94301]

So no hope for firmware upgrade to R5 giving 14 bits at least in one shot...

Br, Pawel

For those of us who are less technically inclined, is there a simple explanation why this is the case? I've been hoping Canon will do something to give the R5 14 bits in ES, at least in single shot mode.

 

[bodeswell]

So no hope for firmware upgrade to R5 giving 14 bits at least in one shot...

Br, Pawel

For those of us who are less technically inclined, is there a simple explanation why this is the case? I've been hoping Canon will do something to give the R5 14 bits in ES, at least in single shot mode.

I believe the tradeoff is that adding more bits would slow down the ES sensor read out. This seems to be pretty common across brands, to the dismay of some of those brand's users too,

But what is the use of ES in one shot mode? Why not just use EFCS? It would give you the 14 bits you want, and in one shot mode there would not be any shutter shock.

 

[PDE94301]

But what is the use of ES in one shot mode? Why not just use EFCS? It would give you the 14 bits you want, and in one shot mode there would not be any shutter shock.

Fair point, but some of us would prefer to go ES for less noise and shutter longevity concerns, but only if there are no compromises on bit depth. Also it would need to be compatible with flash ideally.

 

[bodeswell]

But what is the use of ES in one shot mode? Why not just use EFCS? It would give you the 14 bits you want, and in one shot mode there would not be any shutter shock.

Fair point, but some of us would prefer to go ES for less noise and shutter longevity concerns, but only if there are no compromises on bit depth. Also it would need to be compatible with flash ideally.

I suspect the mechanical shutter will go the way of the reflex mirror … eventually. Compatibility with flash will be tricky. I don’t think this sort of thing can be done in a firmware update. I think it requires new hardware. In the meantime I agree that ES would be quieter. Not sure shutter longevity is a concern in one shot mode. Not sure 12 vs 14 makes that much difference, but if it does then one shot on the R7 EFCS is your best bet, perhaps followed by slow continuous.