p.3 #1 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
dwweiche wrote:
Perhaps we've dug deep enough to have some evidence that claiming 14-bit files is mostly good marketing copy, but provides little to no actual benefit for the photographer. Just don't send that note to Canon support, Bruce
Does somewhat sound like it. I think I’ll shoot in ES a lot more and see if I can notice any difference in the files 😅
Bruce n Philly wrote:
So quickie question:
Does 1st curtain shutter have less shutter shock than Mechanical?
If 1st does have less, and has full 14-bit (except for H+), then 1st curtain should be the go to mode unless you need 20 fps. In other words, if you value 14-bit.
Or, if you like mechanical, you don't loose anything going to 1st curtain, and gain some sharpness. Or said another way, there is no advantage at all to using Mechanical shutter.
Correct?
UPDATE from Canon: 1st Curtain is just like Mechanical.
I better stop bugging them for a while.....
Peace
Bruce in Philly
At shutter speeds of over 1/1000th sec, EFCS is meant to effect bokeh so many say that using MS shutter is preferred. I used the A7III for 3 years and used the MS less than 5 times, even when shooting over 1/1000th but I personally didn’t notice any difference.
I’ve found less shutter shock in EFCS with no visible negative effects so it’s the standard setting for me. I know one guy who only uses MS as he prefers the sound but that’s it lol.
p.3 #2 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
Bruce n Philly wrote:
Update from Canon.... I think is the last one... the first two responses were from a different support person... this one is a bit terse in tone. So I don't think I will bug them again. This person appears to have more knowledge. To get this response... besides bugging the crap out of them... was to post the links by Jesse Evans and stanj in my last email to them (thanx Jesse and stan for your effort!).
I think my last email with these two links triggered something behind the scenes... anywho... here is the third response from Canon:
"The primary bit depth for shooting RAW images on the EOS R5 camera is a 14-bit. This is what we present to the public in general when it comes to the bit depth. The RAW image is saved as a 13-bit RAW image when the data goes through the analog to digital conversion process (A/D conversion) when shooting with the high-speed continuous plus (+) mode. The RAW image is saved as a 12-bit RAW when it's processed through the A/D conversion process when shooting with an electronic shutter. The difference between a 12-bit and 14-bit image is very minimal and may not even be noticeable in a majority of pictures.
The information you see on the Canon website is correct with the bit depth for RAW images. Unfortunately, we do not provide additional information with how images are processed through the processor on the camera. For any information regarding image files, formats, etc. we would recommend to view the user guide and any published information available on our Canon website."
So.....
For highest 14-bit quality: Mechanical shutter one shot, continuous, continuous H, but NOT H+
For medium 13-bit quality: Mechanical shutter H+
For lowest 12-bit quality: Electronic shutter any mode
p.3 #3 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
Photosbydlee wrote:
Does somewhat sound like it. I think I’ll shoot in ES a lot more and see if I can notice any difference in the files 😅
At shutter speeds of over 1/1000th sec, EFCS is meant to effect bokeh so many say that using MS shutter is preferred. I used the A7III for 3 years and used the MS less than 5 times, even when shooting over 1/1000th but I personally didn’t notice any difference.
I’ve found less shutter shock in EFCS with no visible negative effects so it’s the standard setting for me. I know one guy who only uses MS as he prefers the sound but that’s it lol. ...Show more →
You have to be using fast lenses , about 1.4 or higher, for ECFS to affect bokeh. I will admit that after seeing the difference, I would definitely use full mechanical. The quality of bokeh definitely takes a hit.
p.3 #4 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
Just want to say THANK YOU to the guys who have doggedly hunted down the info here in this thread. Much appreciated!
cputeq wrote:
You have to be using fast lenses , about 1.4 or higher, for ECFS to affect bokeh. I will admit that after seeing the difference, I would definitely use full mechanical. The quality of bokeh definitely takes a hit.
So if I'm shooting say EF 135 f/2, or a faster lens closed down a little, the effect shouldn't be worth worrying about? I'm ashamed to say I haven't put this to the test myself, I've just assumed the worst and switched to full mechanical in such cases.
p.3 #5 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
GammyKnee wrote:
Just want to say THANK YOU to the guys who have doggedly hunted down the info here in this thread. Much appreciated!
So if I'm shooting say EF 135 f/2, or a faster lens closed down a little, the effect shouldn't be worth worrying about? I'm ashamed to say I haven't put this to the test myself, I've just assumed the worst and switched to full mechanical in such cases.
Yeah I don't think it happens at f2 or slower (stopped down or 'native') but I can't remember the specifics. I'm thinking it was f/1.4 and faster is where you'd have to bother with full mech.
Easy enough to just take a nice bokeh picture in both modes and compare, but IIRC, f/2 should be fine.
p.3 #6 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
cputeq wrote:
You have to be using fast lenses , about 1.4 or higher, for ECFS to affect bokeh. I will admit that after seeing the difference, I would definitely use full mechanical. The quality of bokeh definitely takes a hit.
Funnily enough I was actually using f/1.4 lenses on Sony when I tested it and couldn’t see anything myself. I also rarely shoot in light over 1/1000th so that’s why I just always leave mine on EFCS.
p.3 #7 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
For me, it looks like I will shoot only 1st curtain from now on, on H (not H+) unless I want faster frames per second (or the full 20 frames per second).
See page 248 and 249 of the R5 manual.
Key points from these pages relating to bokeh and wide open (doesn't give an f/stop threshold):
From page 248:
- Mechanical ...recommended when shooting with the aperture of a bright lens wide open
From page 249:
- Defocused image areas may be incomplete when shooting near maximum aperture at high shutter speeds, depending on shooting conditions. If you dislike the appearance of defocused image areas, ...shoot with an option other than electronic 1st curtain
p.3 #10 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
Jesse Evans wrote:
.... I’m curious to know how the NR was determined to be applied at all. ... I guess if I want to I’ll need to reach out to Bill Claff to discuss.
Usually I see signal processing ( which is assumed to be Noise Reduction ) in the 2D Fourier Transforms that I create.
p.3 #11 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
bclaff_too wrote:
Usually I see signal processing ( which is assumed to be Noise Reduction ) in the 2D Fourier Transforms that I create.
Hey you're on here too.
After I asked this question I found this thread. I noticed some people had posited some alternative theories to noise reduction. However, when I was going to reply for clarification dpreview informed me the reply limit had been reached.
After reading the thread most of my questions were answered, well, at least that I assumed that would be reasonable to assume you would be able to answer.
But I have one remaining question: do you think that a logical alternative explanation is that what we are seeing there is actually a locality correlation between the magnetic fields that are generated by the IBIS module? It is curious that the correlation is most visible at the dual base ISOs present in the camera. At very the very low charge levels of the test images, couldn't that play an impact here?
Part of the reason I ask is that there doesn't appear to be any correlation of pixels indicating any noise reduction at any higher charge values. The correlation decreasing as ISO increases above the current native could also indicate that the base charge level exceeds that of the interference, making it invisible / destroying the correlation.
I'm just interested in whether this makes sense, or it's nonsense.
Since seeing reports that the R5 uses noise reduction at base ISO, I have tried several tests to find the actual noise reduction being applied, but I cannot find it in any visual test I have tried, but it is difficult to make perfectly controlled tests. After diving in to these raw files using rawdigger, I have grown more suspicious of the claims of actual noise reduction.
p.3 #12 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
Jesse Evans wrote:
Hey you're on here too.
...
After reading the thread most of my questions were answered, well, at least that I assumed that would be reasonable to assume you would be able to answer.
But I have one remaining question: do you think that a logical alternative explanation is that what we are seeing there is actually a locality correlation between the magnetic fields that are generated by the IBIS module?...
The way the 2D Fourier Transform works is that it helps to detect signal processing where a pixel correlates with it's immediate neighbors. I can't see how a magnetic field would do that.
As for the behavior relative to ISO setting, I think the effect becomes smaller relative to the rising noise floor.
p.3 #13 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
bclaff_too wrote:
The way the 2D Fourier Transform works is that it helps to detect signal processing where a pixel correlates with it's immediate neighbors. I can't see how a magnetic field would do that.
As for the behavior relative to ISO setting, I think the effect becomes smaller relative to the rising noise floor.
There are several magnetic fields arranged around the sensor, generated by the magnets for the IBIS module. Wouldn’t a non random interference say, in the shape of a magnetic field also produce something that looks like a correlation between pixels since there is no other data present than noise?
p.3 #14 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
Jesse Evans wrote:
There are several magnetic fields arranged around the sensor, generated by the magnets for the IBIS module. Wouldn’t a non random interference say, in the shape of a magnetic field also produce something that looks like a correlation between pixels since there is no other data present than noise?
I don't think so.
It looks like the R5 uses 5 piezoelectric motors, and 3 magnets for the stabilization mechanism, and their locations don't seem to correlate well with the fourier charts.
p.3 #16 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
molson wrote:
Using the Electronic Shutter mode seems to have a pretty significant effect on dynamic range...
This to me is all the proof you need about the Canon file bit depth with ES no matter what Canon reps claim and no of course they have verified what was known all along!
p.3 #17 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
Jesse Evans wrote:
.... I’m curious to know how the NR was determined to be applied at all. ... I guess if I want to I’ll need to reach out to Bill Claff to discuss.
bclaff_too wrote:
Usually I see signal processing ( which is assumed to be Noise Reduction ) in the 2D Fourier Transforms that I create.
Jesse Evans wrote:
Hey you're on here too.
...
After reading the thread most of my questions were answered, well, at least that I assumed that would be reasonable to assume you would be able to answer.
But I have one remaining question: do you think that a logical alternative explanation is that what we are seeing there is actually a locality correlation between the magnetic fields that are generated by the IBIS module?...
bclaff_too wrote:
The way the 2D Fourier Transform works is that it helps to detect signal processing where a pixel correlates with it's immediate neighbors. I can't see how a magnetic field would do that.
As for the behavior relative to ISO setting, I think the effect becomes smaller relative to the rising noise floor.
Jesse Evans wrote:
There are several magnetic fields arranged around the sensor, generated by the magnets for the IBIS module. Wouldn’t a non random interference say, in the shape of a magnetic field also produce something that looks like a correlation between pixels since there is no other data present than noise?
Exactly how much field strength does it take to disrupt the sensor? Those tiny motors are not an NMR.
p.3 #19 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
ES indeed has lower dynamic range but if I understand the photons to photos dr chart correctly that only applies for iso values below 800. At 800 or above this chart shows no significant difference between mechanic and electronic shutter
p.3 #20 · R5/R6 Bit Depth, MS vs ES - Canon says NO DIFFERENCE
ES indeed has lower dynamic range but if I understand the photons to photos dr chart correctly that only applies for iso values below 800. At 800 or above this chart shows no significant difference between mechanic and electronic shutter
