5DS scores at DXO **now posted**

ritholtz said:
jrista said:
ritholtz said:
jrista said:
I think that pretty much sums up a large part of my interest in Exmor in a nutshell right there! :P There isn't ever anything to fix in the first place with an Exmor. The signal is so clean SO deep, you just don't ever have to think about having to fix an issue with the data. It's clean, pristine data, stops and stops deeper than any Canon could ever go.

I don't have to think about ETTR when making the shot, I don't have to worry about clipping highlights, I don't have to worry about fixing color blotch or removing bands or recovering DR. With the Sony's I've used, it's just take the shot, adjust exposure and maybe color to taste and style...that's it! :P

The Sony cams outside of the A6000 haven't performed as well on the AF front, I still generally prefer my 5D III for AF critical applications. Very interested in the A7r II + Metabones + EF 600/4 though....really wonder how that will perform.
DR difference between latest rebel and Sony a6000 is 1stop. Do you think, one going to see this much difference between sensors with only 1 stop extra DR.

With the A6000, it is over stop (PrintDR 11.96 -> 13.14; ScreenDR 11.17 -> 12.34), plus significantly higher quantum efficiency, plus 11fps, plus a much better AF system, plus smaller and lighter (mirrorless), pocket portable (I can drop the camera and a couple lenses into the pockets of my birding slacks, which have some extra pockets on the legs).

The AF system and frame rate are the real big bonuses with the A6000. Were talking 11fps vs. 5fps, and 25/179 point AF system vs. 19 point AF system.

So, for all of you "It's the full package that counts" guys out there...the A6000 trounces the Canon options in the same price range. In every category, not just dynamic range, it delivers better. The way I account for dynamic range...at full RAW image size in Lightroom (no downsampling here, it is no longer RAW if you downsample and edit), the difference in DR is 1.17 stops.

The only thing I don't like about the A6000 is that it doesn't work as well with my Canon lenses when adapted via Metabones. I'm hoping the A6100/A7000 will resolve that issue. But it's the SOLE issue, and I only have it because of my existing Canon lens collection.
Difference in DR is 1.1 (13.1 vs 12) stops. A6000 is going to write 12bit raw with 11 fps shooting. You are going to end up lower DR than even Canon sensor. But my question is, are we going to see all those advantages (The signal is so clean SO deep, you just don't ever have to think about having to fix an issue with the data. It's clean, pristine data, stops and stops deeper than any Canon could ever go) you mentioned between d6000 vs 760d sensor with approximately 1 stop DR difference.

I feel this is just digging for negatives in the Sony to complain about. :P If you actually use these Sony cameras, and dig around in the utter depths of their signals, you would understand what I am referring to when I talk about the clean data.

Sony data can be lifted far more than the stops difference in DR between a Canon. I've lifted A7r data as much as seven stops. That is far beyond the literal differences in dynamic range between any two cameras. The A6000 is still an Exmor. It's the Exmor technology that delivers clean, low noise data. The dynamic range increases because the STDev of noise drops. But there is a lot more to it than that.

The characteristic of the noise changes as well. I don't like lifting my Canon data more than two stops...not because the random noise looks bad (it looks fine)...but because of all the other kinds of noise that show up in Canon data. The blotching and the banding and everything else. This is the area that DXO doesn't cover with their tests, and while some people seem to think that if DXO DID cover it, the noise characteristic realm, that it would paint Canon's in better light. I think it would actually do exactly the opposite, because the noise characteristics of Canon cameras are rather poor.

You might get away with a a two stop lift in a DR-heavy scene (where you've ETTRed heavily) without any hint of banding. If you are really good about debanding, you MIGHT be able to pull the data up three stops. You could easily pull up four, five, maybe even six stops with an A6000, and easily six stops or more with an A7r/A7s, not a hint of banding, not a hint of blotch, not a hint of hot pixels, nothing. Just low STDev random noise.

Oh, as for 12 stops of DR. It doesn't matter as much in practice, not that I've seen. Canon cameras don't even support 12 bits of data in the RAW information coming off the sensor (downsampled images are not telling you about the per-pixel signal actually being read off the sensor and converted to digital numbers...you have to reference the Screen DR/SNR measures for that.) At least the A6000 is fully utilizing every single bit their ADCs support with extremely clean information. Additionally, the data isn't even really 12-bit. It's 11+7 bit encoded. The lossy 11+7 bit encoding doesn't bother me much either (I would certainly prefer true RAW, and it sounds like Sony is working on that.) On rare occasions I've seen situations where that results in some slight artifacting. The article that used RAW Digger to demonstrate the problems with the compression algorithm used the worst possible kind of data there is for it, star trails. In all my time using Alpha series cameras or working with Alpha series RAW data, I've NEVER seen that level of artifacts. Most of the time I work with astro images from Alpha cameras, and I've never seen compression artifacts. I've seen some slight amounts of posterization at the bottom of the signal in dark black background skies in a few cases. Stacking usually cleans most of that up as well. Overall, I'm happy to trade off a couple bits of output precision for the very clean data and all the other features. I have no complaints about the IQ coming from Alpha series cameras. (And, as most people here know, I have a lot of complaints about the data I get from Canon cameras. :\)

I like the pocketable size in particular...I love my 5D III and 600mm f/4 lens. I hate that I cannot safely bring them to work every day without worrying about someone damaging it, or stealing it, because I miss too many opportunities. A pocketable, ultra light weight camera with a high frame rate and excellent IQ is currently at the very top of my value list. Canon doesn't have anything that competes with all of those features at the moment. And I no longer have any interest in waiting for Canon to catch up...I don't even know if they are trying, or are even interested in making a camera that could compete with the A6000. If I knew Canon was interested and trying, that's one thing...but the EOS M isn't even being brought to the states, and the specs have been fairly lackluster. Taken as a whole package, the A6000 is a phenomenal deal, and it STILL has more DR and the ability to recover shadows much, much deeper than any Canon camera on the market.
 
Upvote 0
Sporgon said:
Guys, thanks for the comments you have been making. I wouldn't leave CR, simply because the cool guys here easily out weight the bad. Unfortunate that in this instance one of the bad happens to be an administrator. I'll leave the site owner to deal with that, and perhaps increase this administrators dose of evening primrose oil.

However it has reminded me that I have been drawn into posting too much, and I've trimmed some of my stuff out. And sanj, love your response ;D

Sorry for jumping the gun Sporgon. Glad you are still around. I was enjoying your practical example of the image quality of the Canon vs the Exmor.
 
Upvote 0
jrista said:
ritholtz said:
jrista said:
ritholtz said:
jrista said:
I think that pretty much sums up a large part of my interest in Exmor in a nutshell right there! :P There isn't ever anything to fix in the first place with an Exmor. The signal is so clean SO deep, you just don't ever have to think about having to fix an issue with the data. It's clean, pristine data, stops and stops deeper than any Canon could ever go.

I don't have to think about ETTR when making the shot, I don't have to worry about clipping highlights, I don't have to worry about fixing color blotch or removing bands or recovering DR. With the Sony's I've used, it's just take the shot, adjust exposure and maybe color to taste and style...that's it! :P

The Sony cams outside of the A6000 haven't performed as well on the AF front, I still generally prefer my 5D III for AF critical applications. Very interested in the A7r II + Metabones + EF 600/4 though....really wonder how that will perform.
DR difference between latest rebel and Sony a6000 is 1stop. Do you think, one going to see this much difference between sensors with only 1 stop extra DR.

With the A6000, it is over stop (PrintDR 11.96 -> 13.14; ScreenDR 11.17 -> 12.34), plus significantly higher quantum efficiency, plus 11fps, plus a much better AF system, plus smaller and lighter (mirrorless), pocket portable (I can drop the camera and a couple lenses into the pockets of my birding slacks, which have some extra pockets on the legs).

The AF system and frame rate are the real big bonuses with the A6000. Were talking 11fps vs. 5fps, and 25/179 point AF system vs. 19 point AF system.

So, for all of you "It's the full package that counts" guys out there...the A6000 trounces the Canon options in the same price range. In every category, not just dynamic range, it delivers better. The way I account for dynamic range...at full RAW image size in Lightroom (no downsampling here, it is no longer RAW if you downsample and edit), the difference in DR is 1.17 stops.

The only thing I don't like about the A6000 is that it doesn't work as well with my Canon lenses when adapted via Metabones. I'm hoping the A6100/A7000 will resolve that issue. But it's the SOLE issue, and I only have it because of my existing Canon lens collection.
Difference in DR is 1.1 (13.1 vs 12) stops. A6000 is going to write 12bit raw with 11 fps shooting. You are going to end up lower DR than even Canon sensor. But my question is, are we going to see all those advantages (The signal is so clean SO deep, you just don't ever have to think about having to fix an issue with the data. It's clean, pristine data, stops and stops deeper than any Canon could ever go) you mentioned between d6000 vs 760d sensor with approximately 1 stop DR difference.

I feel this is just digging for negatives in the Sony to complain about. :P If you actually use these Sony cameras, and dig around in the utter depths of their signals, you would understand what I am referring to when I talk about the clean data.

Sony data can be lifted far more than the stops difference in DR between a Canon. I've lifted A7r data as much as seven stops. That is far beyond the literal differences in dynamic range between any two cameras. The A6000 is still an Exmor. It's the Exmor technology that delivers clean, low noise data. The dynamic range increases because the STDev of noise drops. But there is a lot more to it than that.

The characteristic of the noise changes as well. I don't like lifting my Canon data more than two stops...not because the random noise looks bad (it looks fine)...but because of all the other kinds of noise that show up in Canon data. The blotching and the banding and everything else. This is the area that DXO doesn't cover with their tests, and while some people seem to think that if DXO DID cover it, the noise characteristic realm, that it would paint Canon's in better light. I think it would actually do exactly the opposite, because the noise characteristics of Canon cameras are rather poor.

You might get away with a a two stop lift in a DR-heavy scene (where you've ETTRed heavily) without any hint of banding. If you are really good about debanding, you MIGHT be able to pull the data up three stops. You could easily pull up four, five, maybe even six stops with an A6000, and easily six stops or more with an A7r/A7s, not a hint of banding, not a hint of blotch, not a hint of hot pixels, nothing. Just low STDev random noise.

Oh, as for 12 stops of DR. It doesn't matter as much in practice, not that I've seen. Canon cameras don't even support 12 bits of data in the RAW information coming off the sensor (downsampled images are not telling you about the per-pixel signal actually being read off the sensor and converted to digital numbers...you have to reference the Screen DR/SNR measures for that.) At least the A6000 is fully utilizing every single bit their ADCs support with extremely clean information. Additionally, the data isn't even really 12-bit. It's 11+7 bit encoded. The lossy 11+7 bit encoding doesn't bother me much either (I would certainly prefer true RAW, and it sounds like Sony is working on that.) On rare occasions I've seen situations where that results in some slight artifacting. The article that used RAW Digger to demonstrate the problems with the compression algorithm used the worst possible kind of data there is for it, star trails. In all my time using Alpha series cameras or working with Alpha series RAW data, I've NEVER seen that level of artifacts. Most of the time I work with astro images from Alpha cameras, and I've never seen compression artifacts. I've seen some slight amounts of posterization at the bottom of the signal in dark black background skies in a few cases. Stacking usually cleans most of that up as well. Overall, I'm happy to trade off a couple bits of output precision for the very clean data and all the other features. I have no complaints about the IQ coming from Alpha series cameras. (And, as most people here know, I have a lot of complaints about the data I get from Canon cameras. :\)

I like the pocketable size in particular...I love my 5D III and 600mm f/4 lens. I hate that I cannot safely bring them to work every day without worrying about someone damaging it, or stealing it, because I miss too many opportunities. A pocketable, ultra light weight camera with a high frame rate and excellent IQ is currently at the very top of my value list. Canon doesn't have anything that competes with all of those features at the moment. And I no longer have any interest in waiting for Canon to catch up...I don't even know if they are trying, or are even interested in making a camera that could compete with the A6000. If I knew Canon was interested and trying, that's one thing...but the EOS M isn't even being brought to the states, and the specs have been fairly lackluster. Taken as a whole package, the A6000 is a phenomenal deal, and it STILL has more DR and the ability to recover shadows much, much deeper than any Canon camera on the market.

Think you nailed it (them). Love your sientific way of arguing based on facts. But there one more thing, video in Sony camras, crunching Canon. That what made me start to investigate into Sony.
 
Upvote 0
Sporgon said:
Maximilian said:
Eldar said:
zim said:
benperrin said:
Ok I just noticed that Sporgon is now gone. Sad to see someone that was being so helpful leave the community.

+1000
it was because of a disgraceful post by someone he thought was a mod
If someone has his true identity, he should en ouraged to come back. He truly adds value.
He's (fortunately - up to now) still a forum member, just deleted his posts on this thread.

Guys, thanks for the comments you have been making. I wouldn't leave CR, simply because the cool guys here easily out weight the bad. Unfortunate that in this instance one of the bad happens to be an administrator. I'll leave the site owner to deal with that, and perhaps increase this administrators dose of evening primrose oil.

However it has reminded me that I have been drawn into posting too much, and I've trimmed some of my stuff out. And sanj, love your response ;D

Hahaha. Thank you. But perhaps I was too harsh. Maybe the Mod was going through a tough day. Nice to see that you are still around. :)
 
Upvote 0
sanj said:
Hahaha. Thank you. But perhaps I was too harsh. Maybe the Mod was going through a tough day. Nice to see that you are still around. :)

Actually I don't think you were too harsh. When I have a bad day I don't dump on people if I'm in a senior position to them. This guy has actually got involved in inappropriate postings for an admin before anyway.

I'm a great believer in letting someone take enough rope.........
 
Upvote 0
Sporgon said:
sanj said:
Hahaha. Thank you. But perhaps I was too harsh. Maybe the Mod was going through a tough day. Nice to see that you are still around. :)

Actually I don't think you were too harsh. When I have a bad day I don't dump on people if I'm in a senior position to them. This guy has actually got involved in inappropriate postings for an admin before anyway.

I'm a great believer in letting someone take enough rope.........

I agree here...having seen that post for myself now, that was unbecoming of a mod. Bit surprising...
 
Upvote 0
benperrin said:
Sporgon said:
Guys, thanks for the comments you have been making. I wouldn't leave CR, simply because the cool guys here easily out weight the bad. Unfortunate that in this instance one of the bad happens to be an administrator. I'll leave the site owner to deal with that, and perhaps increase this administrators dose of evening primrose oil.

However it has reminded me that I have been drawn into posting too much, and I've trimmed some of my stuff out. And sanj, love your response ;D

Sorry for jumping the gun Sporgon. Glad you are still around. I was enjoying your practical example of the image quality of the Canon vs the Exmor.

As I said Ben, more good guys than bad here. Thanks for your support.
 
Upvote 0
jrista said:
Sporgon said:
sanj said:
Hahaha. Thank you. But perhaps I was too harsh. Maybe the Mod was going through a tough day. Nice to see that you are still around. :)

Actually I don't think you were too harsh. When I have a bad day I don't dump on people if I'm in a senior position to them. This guy has actually got involved in inappropriate postings for an admin before anyway.

I'm a great believer in letting someone take enough rope.........

I agree here...having seen that post for myself now, that was unbecoming of a mod. Bit surprising...

Thanks for taking the time to comment Jon. I know we don't always see eye to eye on the Exmor / Canon tech but arguments must be balanced on each side.
 
Upvote 0
KeithBreazeal said:
Speaking of DXO...
I tried to load a 5DS raw file into DXO Pro 9 Elite. IT WON'T LOAD. Went to their chat and found I'll need Pro 10. NOT HAPPY! The chat guy was good, but the company failed.
DXO flaming...

Try it with an older major version of ACR...then contact Adobe for help. :o
 
Upvote 0
neuroanatomist said:
KeithBreazeal said:
Speaking of DXO...
I tried to load a 5DS raw file into DXO Pro 9 Elite. IT WON'T LOAD. Went to their chat and found I'll need Pro 10. NOT HAPPY! The chat guy was good, but the company failed.
DXO flaming...

Try it with an older major version of ACR...then contact Adobe for help. :o
DXO issue. Lightroom is perfect.
 
Upvote 0
jrista said:
Oh, as for 12 stops of DR. It doesn't matter as much in practice, not that I've seen. Canon cameras don't even support 12 bits of data in the RAW information coming off the sensor (downsampled images are not telling you about the per-pixel signal actually being read off the sensor and converted to digital numbers...you have to reference the Screen DR/SNR measures for that.) At least the A6000 is fully utilizing every single bit their ADCs support with extremely clean information. Additionally, the data isn't even really 12-bit. It's 11+7 bit encoded. The lossy 11+7 bit encoding doesn't bother me much either (I would certainly prefer true RAW, and it sounds like Sony is working on that.) On rare occasions I've seen situations where that results in some slight artifacting. The article that used RAW Digger to demonstrate the problems with the compression algorithm used the worst possible kind of data there is for it, star trails. In all my time using Alpha series cameras or working with Alpha series RAW data, I've NEVER seen that level of artifacts. Most of the time I work with astro images from Alpha cameras, and I've never seen compression artifacts. I've seen some slight amounts of posterization at the bottom of the signal in dark black background skies in a few cases. Stacking usually cleans most of that up as well. Overall, I'm happy to trade off a couple bits of output precision for the very clean data and all the other features. I have no complaints about the IQ coming from Alpha series cameras. (And, as most people here know, I have a lot of complaints about the data I get from Canon cameras. :\)

what does lossy 11 + 7 bits encoding mean.

I understand lossy and bits.
 
Upvote 0
KeithBreazeal said:
neuroanatomist said:
KeithBreazeal said:
Speaking of DXO...
I tried to load a 5DS raw file into DXO Pro 9 Elite. IT WON'T LOAD. Went to their chat and found I'll need Pro 10. NOT HAPPY! The chat guy was good, but the company failed.
DXO flaming...

Try it with an older major version of ACR...then contact Adobe for help. :o

DXO issue. Lightroom is perfect.

Lightroom 5 is perfect for opening 5Ds files?
 
Upvote 0
RGF said:
jrista said:
Oh, as for 12 stops of DR. It doesn't matter as much in practice, not that I've seen. Canon cameras don't even support 12 bits of data in the RAW information coming off the sensor (downsampled images are not telling you about the per-pixel signal actually being read off the sensor and converted to digital numbers...you have to reference the Screen DR/SNR measures for that.) At least the A6000 is fully utilizing every single bit their ADCs support with extremely clean information. Additionally, the data isn't even really 12-bit. It's 11+7 bit encoded. The lossy 11+7 bit encoding doesn't bother me much either (I would certainly prefer true RAW, and it sounds like Sony is working on that.) On rare occasions I've seen situations where that results in some slight artifacting. The article that used RAW Digger to demonstrate the problems with the compression algorithm used the worst possible kind of data there is for it, star trails. In all my time using Alpha series cameras or working with Alpha series RAW data, I've NEVER seen that level of artifacts. Most of the time I work with astro images from Alpha cameras, and I've never seen compression artifacts. I've seen some slight amounts of posterization at the bottom of the signal in dark black background skies in a few cases. Stacking usually cleans most of that up as well. Overall, I'm happy to trade off a couple bits of output precision for the very clean data and all the other features. I have no complaints about the IQ coming from Alpha series cameras. (And, as most people here know, I have a lot of complaints about the data I get from Canon cameras. :\)

what does lossy 11 + 7 bits encoding mean.

I understand lossy and bits.

It just refers to how the compressed data is actually encoded into the ARW file. The information is "encoded"...in other words, it is not simply "stored" as is from the sensor.

I honestly don't know why Sony is doing this. It's hurting the maximum potential of their cameras. It's called cRAW, Compressed RAW, and it is currently the only "RAW" option on all current Sony cameras. Because the data is LOSSY compressed, technically speaking, the data is most definitely NOT RAW. It's much more akin to JPEG than RAW.

That said, outside of certain kinds of images, such as star trails (which seem to exhibit compression artifacts worst by far), and maybe potentially around thin strait lines in images...like power lines...the compression artifacts are extremely difficult to see in most cases. Generally speaking, the differential in compressed pixel values is lower than photon shot noise (although not always), so the artifacts end up just getting drowned out. Under very low signal strengths, and (at least according to reports) when using ETTR in very high signal strengths, you can sometimes see artifacting. It's called posterization by most, but I think that is the wrong term. Compression artifacts is the right term.

Anyway. To demonstrate how little it seems to matter:

FcZkz0O.jpg


A7r and 5D III. Which is which is obvious. This is about a seven stop or so shadow push between the exposure and the shadows sliders in LR. The A7r data, despite the most precise pixel data being encoded with only 11 bits, clearly has loads more dynamic range than the 5D III data, which is encoded with a full uncompressed 14 bits.

The cRAW algorithm does apply an S-log like curve to the 14 bit data coming off the sensor first. Then it applies black point clipping. This effectively preserves the full dynamic range of the original data, but compresses it into a 12-bit range. Then it encodes two pixels in a block (I think it's a 32 pixel block) at the 11 bit precision, the rest are stored as 7 bit deltas from those first two. So, 11+7 bit encoding. Why they don't encode the first two with the full 12 bits is beyond me, but, that's how it works.

Anyway...it's lossy compressed. That is certainly not ideal. Sony seems to have heard the complaints about it, and is apparently working on a firmware fix...although who knows when that will arrive. However in practice, the results are still better than Canon's 14-bit RAW images, where around three bits of data are just lost to noise anyway. The difference is that the Sony data starts out with significantly more precision and usable information than the Canon data. Compression costs you some precision...however it does not cost you the usability of the information (except in a few extreme circumstances...i.e. don't buy a Sony camera if you want to do star trails! :P)
 
Upvote 0
neuroanatomist said:
KeithBreazeal said:
neuroanatomist said:
KeithBreazeal said:
Speaking of DXO...
I tried to load a 5DS raw file into DXO Pro 9 Elite. IT WON'T LOAD. Went to their chat and found I'll need Pro 10. NOT HAPPY! The chat guy was good, but the company failed.
DXO flaming...

Try it with an older major version of ACR...then contact Adobe for help. :o

DXO issue. Lightroom is perfect.

Lightroom 5 is perfect for opening 5Ds files?

Lightroom 6 opens 5DsR files if that helps.
 
Upvote 0
scyrene said:
neuroanatomist said:
KeithBreazeal said:
neuroanatomist said:
KeithBreazeal said:
Speaking of DXO...
I tried to load a 5DS raw file into DXO Pro 9 Elite. IT WON'T LOAD. Went to their chat and found I'll need Pro 10. NOT HAPPY! The chat guy was good, but the company failed.
DXO flaming...
Try it with an older major version of ACR...then contact Adobe for help. :o
DXO issue. Lightroom is perfect.
Lightroom 5 is perfect for opening 5Ds files?
Lightroom 6 opens 5DsR files if that helps.

Of course it does – v6 is the current version of LR, it supports the newly-released 5Ds. For DxO, the current version is v10, which also supports the 5Ds. Older versions (LR v5 and DxO v9) do not support the 5Ds.

Yet according to Keith, 'Lightroom is perfect,' while 'DxO failed'. Something failed alright, but not software or a software company. Keith's problem in this case is purely a wetware failure.
 
Upvote 0