Analysis of RAW samples at Fred Miranda show weak DR

[M42]

why would the 6D2 be different?

Agree 100%. I believe most of us here were assuming we'd see the 13+ EV DR (+1.5-ish EV bump over the prior model) that the 1DX2 / 5D4 / 80D all demonstrated, so no, this makes no sense at face value.

Nobody is claiming it does, and I especially can't believe people saying Canon deliberately crippled this model to protect the 5D4 sales. But it is is also unlikely that they would distribute pre-production samples which are extremely different from the end model, particularly if this difference is on the sensor which is the core component of the camera. As a potential buyer, I just want to hold off for now and wait for better testing.

 

[Loswr]

I have a real hard time blaming people reporting on things for the chicken littles of the world

I have a little 'cheat sheet' for scientific publications that helps lay folks interpret what standard phrasing really means. For example, 'It is belived that...' really means 'I think'; 'It is widely believed that...' really means 'Me and a few colleagues think'; and 'A representative example is shown' really means 'the best example is shown'.

So just for fun, let's apply similar translations to Bill's post:

[quote author=Bill Claff on DPR]
The preliminary results are based on the same CR2 files that have been discussed elsewhere.
[/quote]
...means: 'I downloaded files from the internet, from unverified sources, and given that Canon generally doesn't permit dissemination of RAW files from preproduction units, those files are most likely illicit, and I then proceeded to analyze those files and I will call the results preliminary.

[quote author=Bill Claff on DPR]
It's not unusual for my initial estimates to be low particularly for low ISO Settings.
[/quote]
...means: My preliminary results are often wrong.

[quote author=Bill Claff on DPR]
That said, as has been discussed elsewhere, the 6D Mark II does not appear to have significantly better PDR than the 6D.
[/quote]
...means: Other people on the internet are saying the 6DII is no better than the 6D, and my preliminary results (which are often wrong) based on those same leaked images show the same thing, so it must be true.


[quote author=Bill Claff on DPR]
Bill ( Your trusted source for independent sensor data at http://www.PhotonsToPhotos.net )
[/quote]

Well, I'm not sure that I trust someone who analyzes unverified source material, admits those preliminary analyses are often incorrect, then proceeds to draw conclusions from those data.

All in all, I think it's rather sad that he's stooped to this level.

 

[unfocused]

We all should be thankful to those controversial pre-productional RAWs and to the folks who did the test. Now we have so much discuss, otherwise the forum was getting boring

No, I think a 7D3 in 2018 rumor is handling that just fine.

- A

Yes, but Dynamic Range is so much more fun. Primarily because it is so much more pointless. The arguments are so bitter because the stakes are so low.

 

[unfocused]

Just a thought.

Let's assume for a millisecond that the new 6DII does have slightly less dynamic range than the 5DIV. Here is the conclusion I would draw from that.

The 5DIV sensor is turning out to be something very special. We already knew that it controls noise at high ISO at a level that rivals the lower megapixel and more expensive 1DX II. Not quite as good, but much better than one would expect given its higher pixel density.

Isn't it reasonable to assume that Canon put some extra effort and expense into designing and manufacturing this new sensor so that it would be the best possible tool available to event and wedding photographers (one of its primary target audiences). After having produced an original 6D that "beat" the 5DIII in low light (marginally at best), it is logical that they decided they really needed to step things up for the 5DIV.

The result could, plausibly, be a sensor design and manufacturing process that requires more care and expense than is typical. Perhaps some costly tweaking of the photo cells, I don't know, I'm not a sensor designer.

But, my point is simply this: If (and it is a big "if" that is currently fueled only by internet rumors) the 5DIV sensor outperforms the 6DII sensor, why on earth would anyone find it surprising? After all, the 5DIV is the more expensive camera. Would anyone expect a car manufacturer to put a superior engine in its budget model and a weaker engine in its premium model?

Oh, I can hear the whines right now: "But they did it last time. Therefore I am entitled to Canon continuing that same mistake in all future generations." Sorry, but no. Grow up.

Now, if it turns out that for some reason the 6DII underperforms the original 6D that would be surprising and concerning, but I doubt that will be the case.

 

[M42]

Well, I'm not sure that I trust someone who analyzes unverified source material, admits those preliminary analyses are often incorrect, then proceeds to draw conclusions from those data.

True. I was in the same position a few days ago, so I downloaded the data and I ran my own tests. For the moment, I have no reason to believe that the CR2s were doctored in any way : the distribution of the read noise (in ADU) is consistent in shape with other camera models (a simple Gaussian), including the 6D and the 5D4. It is just much more spread out (actually more spread out than even my old T3). The files come from a well-known source with a sample "pre-production" unit, not some kind of engineering sample. So there is still a slim chance that these units use a lower sensor grade than the final product, but it seems very unlikely that Canon would share a botched product with external entities for preview.

 

[Loswr]

Did anyone else get the pre-order email advertisement from Adorama? If so, did you read it carefully?

 

[ahsanford]

Did anyone else get the pre-order email advertisement from Adorama? If so, did you read it carefully?

Hahahahahaha well done

- A

 

[tr573]

Well, I'm not sure that I trust someone who analyzes unverified source material, admits those preliminary analyses are often incorrect, then proceeds to draw conclusions from those data.

All in all, I think it's rather sad that he's stooped to this level.

I guess I'm not disappointed or sad because this isn't a new thing. He's done it before, with the same caveats provided (or lack thereof if you translate them with your guide ) I don't expect him to wait until the camera ships because he doesn't if he has the chance to publish numbers early. It's hard to be disappointed if you already know Santa isn't real.

And I certainly don't blame him for people running wild with it. The world is full reactionary doom and gloomers. The whole business model of cable news is built on it.

 

[tr573]

Did anyone else get the pre-order email advertisement from Adorama? If so, did you read it carefully?

LOL

The sad thing here is that I really starting to enjoy forums free of people have post processing pissing matches with each other over who could torture a bland photo the most to "prove" how sufficient Canon's DR was. I hope this is all wrong just so I don't have to go back to those days.

 

[ahsanford]

But, my point is simply this: If (and it is a big "if" that is currently fueled only by internet rumors) the 5DIV sensor outperforms the 6DII sensor, why on earth would anyone find it surprising? After all, the 5DIV is the more expensive camera. Would anyone expect a car manufacturer to put a superior engine in its budget model and a weaker engine in its premium model?

You make perfect sense from the good-better-best market segmentation perspective: the 5D4 logically should have a better sensor than a 6D2 in general terms.

But on base ISO DR, Canon of late has simply been putting in the latest gen of tech in there -- not everywhere, but certainly once you leave Rebel territory. The 1DX2 / 5D4 / 80D (and I believe EOS M5/M6?) all got the 'on chip ADC bump' over its predecessor. Why wouldn't the 6D2 be afforded that same courtesy?

I don't see that as 'the 6D1 sensor outperformed the 5D3 sensor so that always should be true' sort of entitlement, I see it as Canon has upped it's base ISO DR game and every camera should eventually reap that benefit. So -- if this somehow turns out to be true with production 6D2s (and we are not there yet) -- the news would be much more about a 6D2 not getting the 'on chip bump' over the 6D1 than anything to do with the 5D4.

- A

 

[Jopa]

Did anyone else get the pre-order email advertisement from Adorama? If so, did you read it carefully?

Weird, mine came slightly different...

 

[Loswr]

;D

 

[ahsanford]

Mine was altogether different 8)

- A

 

[Jopa]

Seems like Neuro got us a really great idea to exploit ;D

ND grands are awesome btw. Tried them first time last Sat and was very impressed.

 

[privatebydesign]

Seems like Neuro got us a really great idea to exploit ;D

ND grands are awesome btw. Tried them first time last Sat and was very impressed.

A lot of stuff is on first use, wait until you have used them more and hit their limitations regularly. They suck 95% of the time they are inflexible and have several severe limitations, put the fact that blending gives a much better result 95% of the time and they look like a very poor investment.

People interested in knockout landscape images are far better spending $20-40 on a blending program/plugin than three to ten times that on ND grads.

 

[Loswr]

Seems like Neuro got us a really great idea to exploit ;D

ND grands are awesome btw. Tried them first time last Sat and was very impressed.

A lot of stuff is on first use, wait until you have used them more and hit their limitations regularly. They suck 95% of the time they are inflexible and have several severe limitations, put the fact that blending gives a much better result 95% of the time and they look like a very poor investment.

People interested in knockout landscape images are far better spending $20-40 on a blending program/plugin than three to ten times that on ND grads.

I have and use NDs (non-grad) frequently, up to 10-stop and all round/screw-in. I had a Lee setup for use with grad NDs for a while, used it a few times initially, sold it a while back. The guy who bought the kit from me brought his camera to try it out – it was a Sony a7R II. ;D

 

[unfocused]

But, my point is simply this: If (and it is a big "if" that is currently fueled only by internet rumors) the 5DIV sensor outperforms the 6DII sensor, why on earth would anyone find it surprising? After all, the 5DIV is the more expensive camera. Would anyone expect a car manufacturer to put a superior engine in its budget model and a weaker engine in its premium model?

You make perfect sense from the good-better-best market segmentation perspective: the 5D4 logically should have a better sensor than a 6D2 in general terms.

But on base ISO DR, Canon of late has simply been putting in the latest gen of tech in there -- not everywhere, but certainly once you leave Rebel territory. The 1DX2 / 5D4 / 80D (and I believe EOS M5/M6?) all got the 'on chip ADC bump' over its predecessor. Why wouldn't the 6D2 be afforded that same courtesy?

I don't see that as 'the 6D1 sensor outperformed the 5D3 sensor so that always should be true' sort of entitlement, I see it as Canon has upped it's base ISO DR game and every camera should eventually reap that benefit. So -- if this somehow turns out to be true with production 6D2s (and we are not there yet) -- the news would be much more about a 6D2 not getting the 'on chip bump' over the 6D1 than anything to do with the 5D4.

- A

No disagreement. The 6DII should (and I expect that it does) have on chip ADC and the resultant improvements over its predecessor. As I stated, if it should turn out that it doesn't have some improvement over the original 6D, that would be a logical basis for complaints or as least disappointment. Even the models prior to on-chip ADC improved from generation to generation (7d vs. 7D II is a good example). So, I would be sympathetic to the complainers if that turned out to be the case. I'm just dismissing that possibility as I don't find it credible in the least and instead focusing on those who will cry "crippling" if they don't get a better sensor in the 6DII than that in the 5DIV. Those are the whiners that will find no sympathy from me.

 

[CanonCams]

Seems like Neuro got us a really great idea to exploit ;D

ND grands are awesome btw. Tried them first time last Sat and was very impressed.

A lot of stuff is on first use, wait until you have used them more and hit their limitations regularly. They suck 95% of the time they are inflexible and have several severe limitations, put the fact that blending gives a much better result 95% of the time and they look like a very poor investment.

People interested in knockout landscape images are far better spending $20-40 on a blending program/plugin than three to ten times that on ND grads.

What program would you recommend for that price?

 

[privatebydesign]

Seems like Neuro got us a really great idea to exploit ;D

ND grands are awesome btw. Tried them first time last Sat and was very impressed.

A lot of stuff is on first use, wait until you have used them more and hit their limitations regularly. They suck 95% of the time they are inflexible and have several severe limitations, put the fact that blending gives a much better result 95% of the time and they look like a very poor investment.

People interested in knockout landscape images are far better spending $20-40 on a blending program/plugin than three to ten times that on ND grads.

What program would you recommend for that price?

I use RayaPro, it costs $44.99 now and is a PS plugin.

http://www.shutterevolve.com/raya-pro-the-ultimate-digital-blending-workflow-panel-for-photoshop/

Mind you, there isn't a thing it does you can't do 'longhand' in PS, but it is a real timesaver. Look at some of his tutorials on YouTube, he normally explains how todo it with the plugin and without.

 

[CanonCams]

Seems like Neuro got us a really great idea to exploit ;D

ND grands are awesome btw. Tried them first time last Sat and was very impressed.

A lot of stuff is on first use, wait until you have used them more and hit their limitations regularly. They suck 95% of the time they are inflexible and have several severe limitations, put the fact that blending gives a much better result 95% of the time and they look like a very poor investment.

People interested in knockout landscape images are far better spending $20-40 on a blending program/plugin than three to ten times that on ND grads.

What program would you recommend for that price?

I use RayaPro, it costs $44.99 now and is a PS plugin.

http://www.shutterevolve.com/raya-pro-the-ultimate-digital-blending-workflow-panel-for-photoshop/

Mind you, there isn't a thing it does you can't do 'longhand' in PS, but it is a real timesaver. Look at some of his tutorials on YouTube, he normally explains how todo it with the plugin and without.

I am not a fan of the 'HDR' cartoony look. His pictures look nice.

I'll look into the plugin.