p.37 #1 · Official: Canon Announces the EOS 6D Mark II
ironmarshal wrote:
Given that you could buy a crop sensor camera using the identical AF and having better dynamic range, the 6D Mark II doesn't scream upgrade potential to me.
Makes buying an UWA for crop a bit more plausible consideration ... albeit, still not FF ... And grab a used 6D for a song.
Still, with all other models showing the lift in DR ... the mind is looking for the 6D2 to follow suit. Count me in the Dave draggin' the last shred of hope for a different reveal territory.
p.37 #2 · Official: Canon Announces the EOS 6D Mark II
OK, please excuse my utter ignorance on this latest issue with the 6DII, but could someone explain to me how an increased DR would have improved this image; https://www.fredmiranda.com/forum/topic/1497323/25#14098785 that Rob Dickinson took with the 6DII?
And to quote Denzel Washington: please explain it to me like I'm a two year old.
p.37 #5 · Official: Canon Announces the EOS 6D Mark II
Ok, here's something else that is weird : I was looking at the press release from Canon and I found a different version on FKP (visible here). The current version on Canon website is slightly different (see here). Here is the interesting paragraph (emphasis mine).
FKP :
“After just one day in Yellowstone National Park with the EOS 6D Mark II DSLR Camera, it was clear that dynamic range has greatly improved over its predecessor,” said acclaimed nature photographer and Canon Explorer of Light Adam Jones
Canon :
“After just one day in Yellowstone National Park with the EOS 6D Mark II DSLR Camera, it was clear that the camera is greatly improved over its predecessor,” said acclaimed nature photographer and Canon Explorer of Light Adam Jones.
Most of the first part of the description is identical. There are other changes in the next sentence. Mostly stylistic. Google cache and Wayback Machine do not have the version before July 5th... Does anybody have an early copy of the PR?
The article is dated FKP article is dated "Jun 29th, 2017" and the Google cache is dated 8 July (the snapshot is a bit too late to make sure the page hasn't been changed recently)...
p.37 #7 · Official: Canon Announces the EOS 6D Mark II
cgarcia wrote:
2) Open the generated PGM file with a image editor and visually determine the area of the masked pixels at the top and the left. For a given sensor it is the same for all the pictures (for the 6D2: 120 pixels at the left and 44 at the top).
I am really a newbie how to calculate the DR.
So, please answer a few questions I have:
1. Doesn't a scenary in motion like the ones in Rob's images disturb the calculations?
2. Why does 6d Mk2 has a mean of +0,7 eV more exposure than 6D across all ISO's. The biggest difference is in ISO100 which is more than +1 eV. (13s vs 30s). Doesn't this disturb the calculations?
3. Won't the vignetting interfere?. The 16-35 at F4 has a vignetting of -2 -> -1,5 eV while EF 35/1.4 at F4 has only about -0,5 eV. Meaning the sensor will gather different amount of photons in the corners.
4. The 16-35 has a transmittans value for 4.5. The 35 has about T4 after stopping down. That will give 6D a slight edge in photon count. But then again, will the calculations cancel this out somehow?
5. The Tree on the hill when boosting the shadows. I'd noticed that the focus could be slighty off. Could this affect the details recovered on the tree? If its slightly off, there are no details, just noise? The trash on the beach on the left is more detailed on the 6D Mk2 while the tree is more detailed on 6D.
6, I also noticed a slight difference in WB even when doing it manually on the same spot on the scene. Could this change interfere? The WB could be due usage of different lenses.
p.37 #8 · Official: Canon Announces the EOS 6D Mark II
M42Nebula wrote:
Thank you for the clarification. I understand the rationale behind the area normalization, however I am not sure about the log2(sqrt(res6D2/res6D)) you are using. Shouldn't this also reflect an area normalization rather than a length normalization?
On the +0.5Ev note : could they have forgotten/disable in firmware the average of the left and right photo-diode images (from the Dual-Pixel sensor) and make us miss another +0.5Ev? (even though that looks really stupid)
My formula is an area normalization (pixels are square). The DR is the ratio between two quantities, which vary differently when pixels are binned; so the DR change is not proportional to the resolution change.
Left/right photodiodes combination is a must to merely avoid image artifacts.
And although I initially thought the same, the 5D4 does not average left/right photodiodes to generate the final image. It likely uses a single shared circuitry for both subpixels: once the signal of one subpixel is integrated, it memorizes the current count and continues integrating the other subpixel. At the end both values are read. The result are thus two images: AB (both subpixels combined) and B (only ther first pixel). And both have the same read noise because both come from the same single read, so their combination doesn't partially cancels the noise. As a curious note, the B subframe is effectively underexposed by 1 EV (anyway, the B subframe is not useful to recover lost highlights; or at least that wouldn't be an easy task for critical image quality, see https://www.fredmiranda.com/forum/topic/1453555/0).
I assume that Canon could have separately read both subpixels, and on this way the final DR after binding them would effectively be +0.5 EV, but manufacturing such a sensor would be more expensive (2x the ADC?). Hope that 5DS2 does the trick (they can still use the current CR2 file by scaling the RAW data). Canon cripples low end cameras, but is generous with the expensive ones (e.g. the 5DS intermediate ISOs are analog gain, not digital). Let's hope.
p.37 #10 · Official: Canon Announces the EOS 6D Mark II
RobDickinson wrote:
I think people do get overly obsessive about dynamic range, you can and usually do have to learn to work round it with any current camera.
I do agree. Saying things like the 80D is a better landscape camera than the 6D II because it has higher engineering DR at base ISO, is turning things totally upside down. The shadow noise goes break even at +3Ev push. At 2E v push and below the 6D II is better. More than 3 Ev push only affects the darkest parts of a small percentage of the images. For all the rest, the 6D II is better in every other way.
That being said, my next camera upgrade will be in search for more dynamic range than the 5DS R. And surprise, lower shadow noise like in the 5D IV is not what I am looking for. What I want is more headroom. Higher saturation or larger sensor (medium format). Lower shadow noise like what we expected to get in the 6D II will not give me that anyway.
As a replacement for the 6D we have, a 6DII is still on the radar. But it will be further out and maybe not at all.
p.37 #11 · Official: Canon Announces the EOS 6D Mark II
I see the Canon Rumors announcement page from June 29th doesn't have the DR in the quote. Not sure if they have changed the page (I don't know how to look at cached pages) since June 29th.
M42Nebula wrote:
Ok, here's something else that is weird : I was looking at the press release from Canon and I found a different version on FKP (visible here). The current version on Canon website is slightly different (see here). Here is the interesting paragraph (emphasis mine).
FKP :
Canon :
Most of the first part of the description is identical. There are other changes in the next sentence. Mostly stylistic. Google cache and Wayback Machine do not have the version before July 5th... Does anybody have an early copy of the PR?
The article is dated FKP article is dated "Jun 29th, 2017" and the Google cache is dated 8 July (the snapshot is a bit too late to make sure the page hasn't been changed recently)......Show more →
p.37 #12 · Official: Canon Announces the EOS 6D Mark II
They are really for noise and dynamic range only, I wasn't that careful with exposure or focus and comparing the 35L mk2 stopped down to f4 vs 16-35 f4 is somewhat unfair.
Astroscapist wrote:
I am really a newbie how to calculate the DR.
So, please answer a few questions I have:
1. Doesn't a scenary in motion like the ones in Rob's images disturb the calculations?
2. Why does 6d Mk2 has a mean of +0,7 eV more exposure than 6D across all ISO's. The biggest difference is in ISO100 which is more than +1 eV. (13s vs 30s). Doesn't this disturb the calculations?
3. Won't the vignetting interfere?. The 16-35 at F4 has a vignetting of -2 -> -1,5 eV while EF 35/1.4 at F4 has only about -0,5 eV. Meaning the sensor will gather different amount of photons in the corners.
4. The 16-35 has a transmittans value for 4.5. The 35 has about T4 after stopping down. That will give 6D a slight edge in photon count. But then again, will the calculations cancel this out somehow?
5. The Tree on the hill when boosting the shadows. I'd noticed that the focus could be slighty off. Could this affect the details recovered on the tree? If its slightly off, there are no details, just noise? The trash on the beach on the left is more detailed on the 6D Mk2 while the tree is more detailed on 6D.
6, I also noticed a slight difference in WB even when doing it manually on the same spot on the scene. Could this change interfere? The WB could be due usage of different lenses.
p.37 #13 · Official: Canon Announces the EOS 6D Mark II
Astroscapist wrote:
I am really a newbie how to calculate the DR.
So, please answer a few questions I have:
1. Doesn't a scenary in motion like the ones in Rob's images disturb the calculations?
2. Why does 6d Mk2 has a mean of +0,7 eV more exposure than 6D across all ISO's. The biggest difference is in ISO100 which is more than +1 eV. (13s vs 30s). Doesn't this disturb the calculations?
3. Won't the vignetting interfere?. The 16-35 at F4 has a vignetting of -2 -> -1,5 eV while EF 35/1.4 at F4 has only about -0,5 eV. Meaning the sensor will gather different amount of photons in the corners.
4. The 16-35 has a transmittans value for 4.5. The 35 has about T4 after stopping down. That will give 6D a slight edge in photon count. But then again, will the calculations cancel this out somehow?
5. The Tree on the hill when boosting the shadows. I'd noticed that the focus could be slighty off. Could this affect the details recovered on the tree? If its slightly off, there are no details, just noise? The trash on the beach on the left is more detailed on the 6D Mk2 while the tree is more detailed on 6D.
6, I also noticed a slight difference in WB even when doing it manually on the same spot on the scene. Could this change interfere? The WB could be due usage of different lenses.
The DR measured has nothing to do with the picture taken... to measure it we look at dark pixels at the sensor side (not exposed to the light) to determine how much noise is in the picture even before receiving light, and then place this in relation with the maximum amount of light each pixel can measure.
In fact, an alternative technique is to take a picture with the lens cap on. On this way, every pixel is black and can be used to measure the noise, not only the ones at the RAW image border. Although some cameras (not Canon) cut the raw values bellow 0, making a bit more difficult the estimation.
The exposure may slightly affect the measured noise due to the contribution of noise sources other than the read noise. And the manipulation by the manufacturer in the raw files (e.g. vignetting correction at really huge apertures, to hide the fact that digital sensors are not yet suited for F1.4 or F1.2) may also affect the DR. But that won't change the big picture a lot...
Talking about vignetting, one of the usefulness of the shadows recovery ability is to apply vignetting correction without increasing a lot the noise in the corners...
p.37 #14 · Official: Canon Announces the EOS 6D Mark II
M42Nebula wrote:
That's not true with the formula you wrote previously, with the square root inside the log.
For the dual pixel operation, I am not sure. Do you please have additional info on that?
The square root does not changes area into lengths... it is merely the accidental result of solving the DR formula (signal/noise) before and after binning and taking into account that while the signal is aggregated, the noises add by the square root of their squared sum (see the quick example on my previous post).
E.g. DXO estimates the 6D ISO 100 DR as 11.43 at screen (physical value measured at 20MP) and 12.11 at print (+0.68 EV) after normalizing to 8 MP. This last is the only comparable measure and the one people always talk about in the forums.
log2(sqrt(20/8)) = 0.66 EV
Not applying the square root the formula would yield 1.32 and would have rated the 6D close to 13 EV of DR.
There is not a lot information about DPRAW or dual pixel... maybe there are some patents, but very difficult to understand :-)
p.37 #15 · Official: Canon Announces the EOS 6D Mark II
cgarcia wrote:
The DR measured has nothing to do with the picture taken... to measure it we look at dark pixels at the sensor side (not exposed to the light) to determine how much noise is in the picture even before receiving light, and then place this in relation with the maximum amount of light each pixel can measure.
In fact, an alternative technique is to take a picture with the lens cap on. On this way, every pixel is black and can be used to measure the noise, not only the ones at the RAW image border. Although some cameras (not Canon) cut the raw values bellow 0, making a bit more difficult the estimation.
The exposure may slightly affect the measured noise due to the contribution of noise sources other than the read noise. And the manipulation by the manufacturer in the raw files (e.g. vignetting correction at really huge apertures, to hide the fact that digital sensors are not yet suited for F1.4 or F1.2) may also affect the DR. But that won't change the big picture a lot...
Talking about vignetting, one of the usefulness of the shadows recovery ability is to apply vignetting correction without increasing a lot the noise in the corners... ...Show more →
FWIW, you describe one way that read noise can be determined.
From read noise one gets pixel ("engineering") dynamic range (EDR).
And, with normalization, measurements like DxOMark's Landscape Score.
However, this is not how something like the PhotonsToPhotos Photographic Dynamic Range (PDR) is measured; that's done with something more like a wedge rather than black frames.
p.37 #16 · Official: Canon Announces the EOS 6D Mark II
If someone offered me to swap the 6D2 sensor (the low DR one, not the magical, class leading, 5D4 slaying one some believe will appear on the shelves in 2 weeks) into my 1DX2 without any loss of other 1DX2 features (like FPS) I would take them up on that. I don't care about ISO 100-400 (really only 100-200) DR and we all know after ISO 400 every DR curve converges together (other than the D5 which is da bomb).
And if someone offered me to swap the D5 sensor into my 1DX2 I'd pay them to do it.
Needles to say if someone offered to swap the 6D2 sensor into my 5D4 I'd tell them to move along.
p.37 #17 · Official: Canon Announces the EOS 6D Mark II
arbitrage wrote:
5D4 slaying one some believe will appear on the shelves in 2 weeks
Not a 5D slayer ... just somewhere between 80D & 5D4.
+1 that critter & BIF shooters have less concern for ISO 100-400 (convergence close enough after that) than others might.
Might be showing my lack of understanding a bit with this question:
So, while there is a certain "expectation" that the new sensor should perform in the same ballpark as the others, and many of us are scratching our heads at how it can be that it (seemingly, per the evidence presented to date) it does NOT perform in that ballpark ... could this be related to decisions about the 6D2 CFA being different from the CFA being used in the others?
Is that something that could explain why the "same sensor" seems to be performing "so differently". By that, the 80D / 5D4 / 1DX II are all tested to be around 64 ISO (for the 100 ISO setting), while their predecessors were tested @ 80 ISO for the 5D3 & 1DX, and 93 ISO for the 70D.
If the 6D2 retains its base ISO of the original 6D (also @ 80 along with the 5D3 & 1DX), could the choice of the CFA be the rationale for why "same sensor" is yielding different results?
I think part of the response has been not just a matter of the technical aspect of it being different (which smaller pixels with same CFA would be understandably "slightly lower" ... but, scratching our heads at how the same sensor could yield different results from the other bodies using the sensor.
Then ... what is the potential for a different CFA showing up (assuming the theory of "original 6D CFA" being used) in the models to be shipped?
p.37 #18 · Official: Canon Announces the EOS 6D Mark II
RustyBug wrote:
Not a 5D slayer ... just somewhere between 80D & 5D4.
+1 that critter & BIF shooters have less concern for ISO 100-400 (convergence close enough after that) than others might.
Yes critters with wings is really all I care about these days.
Still I think the 6D2 will still be a nice update from 6D for many other reasons (especially AF). I think it may be a little over priced without improvements in DR and High ISO but then again I think the 5D4 is way overpriced also and I still ended up buying one (once on sale).
But with any new Canon camera we always get these 40 page threads with tonnes of complaints and doom predictions leading up to actual release into our hands. It is always funny how months and months later most people who talk about the 5D4 love it and recommend it all the time to people in threads. Go back and look at the release thread for the 5D4 and everyone was panning the camera for this and that just like the 6D2. In the end it will still end up being a traditional Canon workhorse that gets results.
If one needs 4K or class leading low ISO DR or whatever else the 6D2 doesn't have then you have to start looking elsewhere.
p.37 #19 · Official: Canon Announces the EOS 6D Mark II
arbitrage wrote:
low ISO DR or whatever else the 6D2 doesn't have then you have to start looking elsewhere.
Agreed ... just expected that if the lowly 80D saw the increase, the FF 6D2 was anticipated to follow suit. If it doesn't, then it's back to the drawing board on some things. I was really looking forward to ISO 64 in FF from Canon (which means 5D4 without flippy screen if not the 6D2).
p.37 #20 · Official: Canon Announces the EOS 6D Mark II
M42Nebula wrote:
Ok, here's something else that is weird : I was looking at the press release from Canon and I found a different version on FKP (visible here). The current version on Canon website is slightly different (see here). Here is the interesting paragraph (emphasis mine).
FKP :
Canon :
Most of the first part of the description is identical. There are other changes in the next sentence. Mostly stylistic. Google cache and Wayback Machine do not have the version before July 5th... Does anybody have an early copy of the PR?
The article is dated FKP article is dated "Jun 29th, 2017" and the Google cache is dated 8 July (the snapshot is a bit too late to make sure the page hasn't been changed recently)......Show more →
UK 6D2 press info as posted on Northlight Images website
UK 6D2 press info
United Kingdom, Republic of Ireland, 29 June 2017 ...........................Push the boundaries with the latest generation technology
With a new 26.2 Megapixel CMOS sensor, the EOS 6D Mark II offers unrivalled image quality and improved performance.................................................Its high dynamic range ensures exceptional exposure latitude so users can push boundaries when shooting in bright light conditions.
