August 23, 2014, 07:47:11 AM

Author Topic: 6D pattern noise compared to 5D2, the info you won't get from DxO  (Read 10327 times)

DanielW

  • EOS M2
  • ****
  • Posts: 249
    • View Profile
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #15 on: February 07, 2013, 04:40:48 PM »
What if Canon simply ditched the 5D3's sensor and put the 6D's sensor in it?

From the reviews I read and the raw samples I looked at myself you're absolutely overestimating the 5d3/6d difference - the 6d might have 1/3 stop less iso noise, but the 5d3 also has improved banding, has more resolution and is sharper (even with a much stronger aa filter!) than the 6d. Except maybe for iso12800 (favors 6d) or low iso studio shots (favors 5d3) the differences won't show.

I agree about the overestimation thing -- I'm thinking of selling my 60D and getting a 7D (mark 1!) because of AFMA. A FF would be great, indeed much better than my 60D or my "new" 7D, but what's keeping my photos from being better is definitely not my camera.
Going FF with Canon would mean losing the built-in flash as a commander (I have only one flash) and becoming restricted to on-camera flash, when what I need is just the opposite: quit being such a lazy ass and get the flash off-camera more often. Sure, I could just buy another flash, but going FF, buying new lenses and accessories is expensive and probably wouldn't make my photos much better, if any. I'm the limiting factor... :)
And, as jrista pointed out recently, there are softwares with incredible NR capabilities available nowadays. I've found that ISO 1600 and 3200 are really good after adequate NR on the RAW files of my 60D, and for what I shoot it's really enough.
Daniel

TWI by Dustin Abbott

  • 1D X
  • *******
  • Posts: 1555
    • View Profile
    • dustinabbott.net
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #16 on: February 07, 2013, 04:52:56 PM »
What if Canon simply ditched the 5D3's sensor and put the 6D's sensor in it?

From the reviews I read and the raw samples I looked at myself you're absolutely overestimating the 5d3/6d difference - the 6d might have 1/3 stop less iso noise, but the 5d3 also has improved banding, has more resolution and is sharper (even with a much stronger aa filter!) than the 6d. Except maybe for iso12800 (favors 6d) or low iso studio shots (favors 5d3) the differences won't show.

With all due respect, you seem to be repeatedly making this point in a variety of threads, but without substantiating it.  You also don't seem to own any full frame Canon camera body according to your signature, so where is the source for your information?

I don't own and have not used the MKIII, but I did extensive research before purchasing the 6D and have processed many thousands of images with the MKII (which I continue to own).  I can tell you from my research that many professional reviewers when doing testing were surprised to find the improvement in dynamic range, high ISO performance, and reduced noise in the shadows in the 6D.  They weren't looking for it; they were shocked to find it!  I can tell you from my own experience that the difference between the RAW files in these qualities from the 5D2 to the 6D is huge.  I am sure it is far less obvious when compared to the 5D3, but I have not really heard any professional assert that still image quality is better (or even 100% as good).  The 5D3 is unquestionably the better camera; it probably doesn't produce better images.
6D x 2 | EOS-M w/22mm f/2 + 18-55 STM + EF Adapter| Rokinon 14mm f/2.8 | Tamron 24-70 f/2.8 VC | 35mm f/2 IS | 40mm f/2.8 | 100L | 135L | 70-300L -----OLD SCHOOL----- SMC Takumar 28mm f/3.5, Super Takumar 35mm f/3.5, SMC Takumar 55mm f/1.8, Helios 44-2 and 44-4, Super Takumar 150mm f/4

Brock

  • Power Shot G16
  • **
  • Posts: 15
    • View Profile
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #17 on: February 07, 2013, 04:56:50 PM »
Canon can not be equal with theirs old  sensor lay out and and  and signal  path way.
Thera are no indications that Canon has gone from the   old 180nm  tech to Sonys, Panasonic Aptina, Toshiba tech which are down at 65nm

http://www.chipworks.com/blog/technologyblog/2012/10/24/full-frame-dslr-cameras-canon-stays-the-course/

Canon is at 500nm.

Nikon 350nm & 250nm

Sony at 250nm & 180nm (though technically the 180 is in the Nikon D800)

Where are you getting 65nm from?

Marsu42

  • Canon EF 400mm f/2.8L IS II
  • *******
  • Posts: 4391
  • ML-66d / 100L / 70-300L / 17-40L / 600rts
    • View Profile
    • 6D positive spec list
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #18 on: February 07, 2013, 05:15:00 PM »
With all due respect, you seem to be repeatedly making this point in a variety of threads, but without substantiating it.  You also don't seem to own any full frame Canon camera body according to your signature, so where is the source for your information?

Asking for sources is perfectly fine - though since this is not a scientific article, skipping them unless asked to is imho also ok. While looking for the adequate ff camera (5d2/5d3/6d) I did a lot of research and downloaded multiple raw comparisons and had a look for myself in Lightroom. What exact source are you asking for - that the 5d2 is sharpest at low iso, or that the 6d is less sharp than the 5d3? If I can I'll try to post where I got the information from, though it were a lot of articles so I'll really have to look. Not to be misunderstood: I also think the 6d iq is overall much better than the 5d2, that's why I'll buy the 6d.

Btw: I'm not a big fan of the "you don't own that gear, you can't tell anything about it" argument - I for one never told it to anyone commenting on, bashing,  praising or comparing the 60d w/o owning it. While it is certainly true that ownership or extended rental is required for evaluating gear handling over longer periods of time, your 6d probably has the same sensor as any reviewer's 6d, so I'm confident I can come to valid conclusion when using other people's raw files. If reading sources would be no viably way to acquire information, scientific or journalistic work would be confronted with quite a problem.

CarlTN

  • Canon EF 300mm f/2.8L IS II
  • *******
  • Posts: 2227
    • View Profile
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #19 on: February 07, 2013, 05:43:43 PM »
Daniel, AFMA is a must, and the 60D was "rebel-ized" to appeal to Rebel fans.  I've never liked it, which is why I still use my 50D.  Let's face it, the noise is not that much better in the 60D and 7D, nor is the resolution.  By contrast, the resolution was increased a whopping 50% going from 40D, to 50D.  Sure the 40D had less noise at the pixel level, but not when you downsample the 50D's images to match the 40D's 10MP image size.

So the 5D3 has a stronger anti-aliasing filter than the 6D?  I did not know dat...maybe that partially explains why the 6D's video has more moire?

As for 5D3 owners going crazy, how about all those who bought the first production run of the 5D3 with the light leak under the top lcd screen?  Were they not angry too?  hahahaha....

And how about all those Nikon D4 owners who have to contend with the green tinting of both the LCD and the files themselves, with no acknowledgement of a problem from Nikon?  A lot of them just sold everything and bought a 1Dx and the far superior Canon lens line.  Not sure that problem was ever fixed, and don't really care either.

Marsu42

  • Canon EF 400mm f/2.8L IS II
  • *******
  • Posts: 4391
  • ML-66d / 100L / 70-300L / 17-40L / 600rts
    • View Profile
    • 6D positive spec list
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #20 on: February 07, 2013, 06:06:39 PM »
So the 5D3 has a stronger anti-aliasing filter than the 6D?  I did not know dat...maybe that partially explains why the 6D's video has more moire?

Wups, sorry, I read my own post and have to admit that the strenght of the aa filter is just an assumption (unlike the 5d2/5d3 sharpness vs. the 6d) because of the stronger moire, it can also be due to another sampling algorithm vs. the 5d3. So sorry for that, I corrected it above but would also like to have more information on this.

TWI by Dustin Abbott

  • 1D X
  • *******
  • Posts: 1555
    • View Profile
    • dustinabbott.net
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #21 on: February 07, 2013, 06:07:27 PM »
With all due respect, you seem to be repeatedly making this point in a variety of threads, but without substantiating it.  You also don't seem to own any full frame Canon camera body according to your signature, so where is the source for your information?

Asking for sources is perfectly fine - though since this is not a scientific article, skipping them unless asked to is imho also ok. While looking for the adequate ff camera (5d2/5d3/6d) I did a lot of research and downloaded multiple raw comparisons and had a look for myself in Lightroom. What exact source are you asking for - that the 5d2 is sharpest at low iso, or that the 6d is less sharp than the 5d3? If I can I'll try to post where I got the information from, though it were a lot of articles so I'll really have to look. Not to be misunderstood: I also think the 6d iq is overall much better than the 5d2, that's why I'll buy the 6d.

Btw: I'm not a big fan of the "you don't own that gear, you can't tell anything about it" argument - while this is certainly true for saying how some gear handles over longer periods of time, your 6d probably has the same sensor as any reviewer's 6d, so I'm confident I can come to valid conclusion when evaluating other people's raw files. If reading sources would be no viably way to acquire information, scientific or journalistic work would be confronted with quite a problem.

I don't think that I ever said that you could "not tell anything about it".  I do question trying to make such a strong point that seems to contradict prevailing wisdom without more evidence to back it up...and I don't recall you ever actually quoting a source. 

I think your policy of using other people's RAW files to get a sense of the camera's ability is a good one.  I think that doing research is very smart.  But I don't think just having RAW files is going to tell you the whole story about all the potential variables at capture or give you a real sense of the operation/workflow of the camera.  I'm not looking for an argument; I look forward to hearing your thoughts once you have had a chance to use the camera for yourself.
6D x 2 | EOS-M w/22mm f/2 + 18-55 STM + EF Adapter| Rokinon 14mm f/2.8 | Tamron 24-70 f/2.8 VC | 35mm f/2 IS | 40mm f/2.8 | 100L | 135L | 70-300L -----OLD SCHOOL----- SMC Takumar 28mm f/3.5, Super Takumar 35mm f/3.5, SMC Takumar 55mm f/1.8, Helios 44-2 and 44-4, Super Takumar 150mm f/4

Marsu42

  • Canon EF 400mm f/2.8L IS II
  • *******
  • Posts: 4391
  • ML-66d / 100L / 70-300L / 17-40L / 600rts
    • View Profile
    • 6D positive spec list
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #22 on: February 07, 2013, 06:12:47 PM »
I don't think that I ever said that you could "not tell anything about it".  I do question trying to make such a strong point that seems to contradict prevailing wisdom without more evidence to back it up...and I don't recall you ever actually quoting a source.

Well, you didn't ask for a source on any specific item yet :-) ... and sorry if I seemed to generalize your post, I was also commenting on various other past threads where the idea that you have to own something to be able to comment on it is often mentioned - so sorry for being too strong on this.

I'm not looking for an argument; I look forward to hearing your thoughts once you have had a chance to use the camera for yourself.

Me neither, and actually I really value your posts because of you I now decided to go for the Tamron and 6d combination, saving €2000 (that's a lot of money to me) - if you can shoot with it like you do, I certainly don't need the 5d3+Canon mk2 combination. And I'll certainly tell about my experiences here, though I'll wait some for the 6d to drop some more in price.

TWI by Dustin Abbott

  • 1D X
  • *******
  • Posts: 1555
    • View Profile
    • dustinabbott.net
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #23 on: February 07, 2013, 06:18:21 PM »
Good enough  ;) I hope the combo works as well for you as it has for me so far.
6D x 2 | EOS-M w/22mm f/2 + 18-55 STM + EF Adapter| Rokinon 14mm f/2.8 | Tamron 24-70 f/2.8 VC | 35mm f/2 IS | 40mm f/2.8 | 100L | 135L | 70-300L -----OLD SCHOOL----- SMC Takumar 28mm f/3.5, Super Takumar 35mm f/3.5, SMC Takumar 55mm f/1.8, Helios 44-2 and 44-4, Super Takumar 150mm f/4

LetTheRightLensIn

  • Canon EF 300mm f/2.8L IS II
  • *******
  • Posts: 3465
    • View Profile
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #24 on: February 07, 2013, 06:21:23 PM »
LetThe...
Your comment seems to say the 1DsMkIII had no low ISO noise/banding. As someone considering purchasing one I wanted to confirm that positive endorsement (if that's what it was!)
The 5DMkIII and 6D are a confusing pair of cameras. Like the earlier poster I'd like to mix the features differently, some from each... my 5DMkIII a poor copy, I believe, but wished for more improvement in IQ over the 5DMkII.

jonathan7007

Well they all have read noise at low ISO, all a lot more than the exmor stuff, but the 1Ds3 did have among the least low ISO banding that Canon had ever released and some of the better ISO100 read noise for a Canon.

From what I hear and have seen and in some, but not all, cases tested, it seems like for the FF cams from Canon that 1Ds3,1DX,6D all have noticeably less banding than the 5D3/5D2 (5D3 completely cures it in one direction but since it still has it so strongly in the other direction it doesn't really do much anything to help). From what I recall 1Ds3 has no horiz banding and a little bit of vertical and the 5D2 has a lot of both and the 5D3 has no almost vertical but a lot of horiz. Those three (1Ds3/6D/1DX) also all have a bit less read noise than the 5D3/5D2 as well at ISO100 (5D3 is actually the worst of them all for ISO100 read noise, although the degree that it is worse than the 5D2 in that regard is so minor that I don't it is possible to notice real world). Some say the 6D has the least banding of these Canon FF cams, I didn't check that out carefully myself yet.

1DX/6D and even 5D3 (and even 5D2) all have better high ISO DR than the 1Ds3 though (even starting by just ISO800 or even 400 really, not even talking high). And at say ISO3200 the 1Ds3 DR actually does fall quite far behind the 5D3 and bit farther still than the 6D and even a touch yet than the 1DX. Even compared to the 5D2 it doesn't even fair so well at ISO3200.

1DX/6D/5D3 all have better SNR than the 1Ds3 across the range (although SNR is so good that at ISO100 it doesn't matter that much and with the better color filters and less banding and a bit less read noise I'd say 1Ds3 does ISO100 better than the 5D3 even if the SNR is worse outside of the darkest tones).

1Ds3 appears to have the least color-blind color array filter when shooting under outdoor lighting of all those cameras. 6D may be the most color blind. In what ways and what the difference means exactly is very complex and hard to say.

« Last Edit: February 07, 2013, 06:29:05 PM by LetTheRightLensIn »

LetTheRightLensIn

  • Canon EF 300mm f/2.8L IS II
  • *******
  • Posts: 3465
    • View Profile
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #25 on: February 07, 2013, 06:24:34 PM »
What if Canon simply ditched the 5D3's sensor and put the 6D's sensor in it? I mean, same body, same AF system, everything, but made all 5D3s with the 6D's sensor from now on... Same sensor size, same company... Is it feasible, on the technical side?
Daniel

It would mess up the video being only 20MP instead of 22MP. Not 100% sure it could drive the 6fps either, might need more readouts added (might be require only a modest little fix in the grand scheme of things though).

I don't get why they didn't just make the 6D sensor 22MP instead of 20MP and just use the same one in both cameras. Very bizarre.

TWI by Dustin Abbott

  • 1D X
  • *******
  • Posts: 1555
    • View Profile
    • dustinabbott.net
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #26 on: February 07, 2013, 06:36:49 PM »
What if Canon simply ditched the 5D3's sensor and put the 6D's sensor in it? I mean, same body, same AF system, everything, but made all 5D3s with the 6D's sensor from now on... Same sensor size, same company... Is it feasible, on the technical side?
Daniel

It would mess up the video being only 20MP instead of 22MP. Not 100% sure it could drive the 6fps either, might need more readouts added (might be require only a modest little fix in the grand scheme of things though).

I don't get why they didn't just make the 6D sensor 22MP instead of 20MP and just use the same one in both cameras. Very bizarre.

And that is at the heart of what got me when doing research.  The 6D wasn't even on my radar, but as I actually researched the key things that matter to me (IQ being #1), there wasn't really a clear pecking order between the 5D3 and the 6D.  The 5D3 is the better camera...but not in every way.  The 6D was not just a "dumbed down" 5DIII.  In some ways it is actually a superior camera; in others it is considerably inferior.  I was dead set on buying a 5D3 until I did the research and discovered that (at least for my actual purposes) the 6D might actually be the preferred camera.  While I have not yet used a 5DIII, I can safely say that I am very happy with my purchase of the 6D.

But my point is that Canon has really muddied the waters for potential customers like myself.  If the 5D3 was clearly superior in every way, I would be purchasing one right now.  It's not, so I didn't.  Was that in Canon's best interest?
6D x 2 | EOS-M w/22mm f/2 + 18-55 STM + EF Adapter| Rokinon 14mm f/2.8 | Tamron 24-70 f/2.8 VC | 35mm f/2 IS | 40mm f/2.8 | 100L | 135L | 70-300L -----OLD SCHOOL----- SMC Takumar 28mm f/3.5, Super Takumar 35mm f/3.5, SMC Takumar 55mm f/1.8, Helios 44-2 and 44-4, Super Takumar 150mm f/4

Chosenbydestiny

  • EOS M2
  • ****
  • Posts: 238
    • View Profile
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #27 on: February 07, 2013, 06:51:22 PM »
With all due respect, you seem to be repeatedly making this point in a variety of threads, but without substantiating it.  You also don't seem to own any full frame Canon camera body according to your signature, so where is the source for your information?

Asking for sources is perfectly fine - though since this is not a scientific article, skipping them unless asked to is imho also ok. While looking for the adequate ff camera (5d2/5d3/6d) I did a lot of research and downloaded multiple raw comparisons and had a look for myself in Lightroom. What exact source are you asking for - that the 5d2 is sharpest at low iso, or that the 6d is less sharp than the 5d3? If I can I'll try to post where I got the information from, though it were a lot of articles so I'll really have to look. Not to be misunderstood: I also think the 6d iq is overall much better than the 5d2, that's why I'll buy the 6d.

Btw: I'm not a big fan of the "you don't own that gear, you can't tell anything about it" argument - while this is certainly true for saying how some gear handles over longer periods of time, your 6d probably has the same sensor as any reviewer's 6d, so I'm confident I can come to valid conclusion when evaluating other people's raw files. If reading sources would be no viably way to acquire information, scientific or journalistic work would be confronted with quite a problem.

I don't think that I ever said that you could "not tell anything about it".  I do question trying to make such a strong point that seems to contradict prevailing wisdom without more evidence to back it up...and I don't recall you ever actually quoting a source. 

I think your policy of using other people's RAW files to get a sense of the camera's ability is a good one.  I think that doing research is very smart.  But I don't think just having RAW files is going to tell you the whole story about all the potential variables at capture or give you a real sense of the operation/workflow of the camera.  I'm not looking for an argument; I look forward to hearing your thoughts once you have had a chance to use the camera for yourself.

+1 it's a completely different perspective to own the actual product. I bashed the 6D when it was announced and bought one for my wife because she needed "just" the better ISO performance. Ended up buying one for myself as well because of the compact delivery of IQ nearly identical to 5D3. For still subjects at least. It's easily a level above the 5D2 for image quality and AF which matter the most, and several other features. The weaknesses are easily outweighed. I bashed the 60D when it was announced as well, but last year we ended up buying 3 of them instead of the somewhat noisier 50D. IQ just wins in the end over other features, especially since Canon is beginning to trail behind in sensor technology. Not saying it's the only thing that matters, but if you want to make the most out of your lenses I'm sure it's a good place to start.
Nikon electric fan, gas stove, and slippers. Canon Elan 7, 1D Mark III, 5D mark III, and 2x6D. Canon 24mm Samyang/Rokinon, 85L, 135L, and many other lenses. 2x Canon 580ex II, third party speedlites, studio strobes

Ricku

  • 6D
  • *****
  • Posts: 494
    • View Profile
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #28 on: February 07, 2013, 06:58:33 PM »
What if Canon simply ditched the 5D3's sensor and put the 6D's sensor in it? I mean, same body, same AF system, everything, but made all 5D3s with the 6D's sensor from now on... Same sensor size, same company... Is it feasible, on the technical side?
Daniel

It would mess up the video being only 20MP instead of 22MP. Not 100% sure it could drive the 6fps either, might need more readouts added (might be require only a modest little fix in the grand scheme of things though).

I don't get why they didn't just make the 6D sensor 22MP instead of 20MP and just use the same one in both cameras. Very bizarre.

And that is at the heart of what got me when doing research.  The 6D wasn't even on my radar, but as I actually researched the key things that matter to me (IQ being #1), there wasn't really a clear pecking order between the 5D3 and the 6D.  The 5D3 is the better camera...but not in every way.  The 6D was not just a "dumbed down" 5DIII.  In some ways it is actually a superior camera; in others it is considerably inferior.  I was dead set on buying a 5D3 until I did the research and discovered that (at least for my actual purposes) the 6D might actually be the preferred camera.  While I have not yet used a 5DIII, I can safely say that I am very happy with my purchase of the 6D.

But my point is that Canon has really muddied the waters for potential customers like myself.  If the 5D3 was clearly superior in every way, I would be purchasing one right now.  It's not, so I didn't.  Was that in Canon's best interest?
That is why I believe the 5D3 was rushed, in order to go up against the D800, and thus "stealing some of the spotlight" from Nikon.

It is an ugly theory, but it also makes perfect sense.

I also believe that the 5D3 will have a much shorter product cycle than the previous 5D-cameras. A replacement (the 5DX?) will most likely come sooner than later.
« Last Edit: February 07, 2013, 07:04:28 PM by Ricku »

Aglet

  • 1D Mark IV
  • ******
  • Posts: 930
    • View Profile
Re: 6D pattern noise compared to 5D2, the info you won't get from DxO
« Reply #29 on: February 07, 2013, 07:07:26 PM »
Is it better to shoot with the 5d2 at ISO 400 than ISO 100 to avoid the pattern noise? Thanks.
that is a viable workaround for the problem which I should have used more often myself.
dynamic range is almost the same from iso 100 to 800 or more but the effective read noise drops as you move up the iso scale and the signal to noise ratio also gets worse.  BUT, the SNR is still pretty good overall so some minor NR in post will clean up a 400 or 800 ISO raw file to be nearly as good as a 100 iso should be.  IF you have enough shutter speed and other exposure latitude to do this (careful when using flash fill)
after all that, yes, if you can move to iso 400 or more without affecting your composition elements from flash fill then use it to avoid the strong banding because the increased random noise/pattern noise is a benefit in this case.
Of course,all this only applies if you're lifting shadows or raising the overall exposure in post.  If you're not doing that, there's no need to.  the 5d2's shadow banding is usually barely noticeable in shadows that are not lifted at all.  It sort of depends on what you do with the image.