Register · Search · Software · Join Upload & Sell · Hosting

Moderated by: Fred Miranda
Username   Password

FM Forum Rules
Canon SLRs, primes, and zooms lenses reviews
FM Forums | Canon-mount SLRs | Join Upload & Sell   
Search Used
1   2  
3
   4   end
   

Archive 2009 · 7D & 'Mazing' (serious question)
  
 
cameron12x
Offline
Upload & Sell: Off
p.3 #1 · 7D & 'Mazing' (serious question)


I wonder if it is possible that Canon actually purposefully introduced the "maze artifacting" as a mechanism to reduce or eliminate the "noise banding" which plagued the 50D (and to a lesser extent, the 5D2) at high ISO settings?

Or perhaps the maze artifacting is a byproduct of getting rid of the noise banding at high ISO?

If so, that's good for high ISO shooting but, as some have reported, LOW ISO images "seem" noisier than expected when compared to other crop bodies.

Nov 18, 2009 at 02:39 AM
jorkata
Offline
Upload & Sell: Off
p.3 #2 · 7D & 'Mazing' (serious question)


msalvetti wrote:
Compared to other examples, the DPP results look pretty good, don't they?

Yup.
But look at the black & white dots.
This is the salt & pepper noise that only the 7D has of all DSLRs.


Edited on Nov 18, 2009 at 03:32 AM · View previous versions

Nov 18, 2009 at 03:10 AM
jorkata
Offline
Upload & Sell: Off
p.3 #3 · 7D & 'Mazing' (serious question)


cameron12x wrote:
I wonder if it is possible that Canon actually purposefully introduced the "maze artifacting" as a mechanism to reduce or eliminate the "noise banding" which plagued the 50D (and to a lesser extent, the 5D2) at high ISO settings?


Canon purposefully tweaked the sensor's color filter, so that more light can reach the sensor.

This is to minimize noise in general, not just banding.

Don't know exactly what they've done but it's something along the lines of Kodak's TrueSense sensor:


This image is copyrighted by the owner



With this arrangement, the RGBG color data in the raw file no longer contains accurate red, green, and blue colors.

The mazing is created by the raw-demosaicing algorithm - unless it's aware of the tweaked color filter, so that compensatation can be applied.

The problem is that even with compensation, there are still color artifacts - the salt & pepper noise.

Nov 18, 2009 at 03:25 AM
jorkata
Offline
Upload & Sell: Off
p.3 #4 · 7D & 'Mazing' (serious question)


M Vers wrote:
I believe that's because the pink area is the lightest of the four colors...the pattern shows up because of the greater level of contrast.
Nope.
See my post ^^^.
The pattern shows up in colors that are not accurately recorded in the raw file.

Nov 18, 2009 at 03:28 AM
NCAndy
Offline
Upload & Sell: On
p.3 #5 · 7D & 'Mazing' (serious question)


jorkata wrote:
cameron12x wrote:
I wonder if it is possible that Canon actually purposefully introduced the "maze artifacting" as a mechanism to reduce or eliminate the "noise banding" which plagued the 50D (and to a lesser extent, the 5D2) at high ISO settings?


Canon purposefully tweaked the sensor's color filter, so that more light can reach the sensor.

This is to minimize noise in general, not just banding.

Don't know exactly what they've done but it's something along the lines of Kodak's TrueSense sensor:


This image is copyrighted by the owner



With this arrangement, the RGBG color data in the raw file no longer contains accurate red, green, and blue colors.

The mazing is created by the raw-demosaicing algorithm - unless it's aware of the tweaked color filter, so that compensatation can be applied.

The problem is that even with compensation, there are still color artifacts - the salt & pepper noise.
So can these artifacts be reduced or eliminated with a software tweak in the next upgrade of DPP?

Nov 18, 2009 at 03:56 AM
jorkata
Offline
Upload & Sell: Off
p.3 #6 · 7D & 'Mazing' (serious question)


I don't know. If I have to guess, DPP is already doing the best it can.
The hope is that Adobe will come up with a smarter raw converter that will do a better job of eliminating the artifacts.

Nov 18, 2009 at 04:23 AM
douglasf13
Offline
Dedicated FM
Upload & Sell: Off
p.3 #7 · 7D & 'Mazing' (serious question)


jorkata wrote:
cameron12x wrote:
I wonder if it is possible that Canon actually purposefully introduced the "maze artifacting" as a mechanism to reduce or eliminate the "noise banding" which plagued the 50D (and to a lesser extent, the 5D2) at high ISO settings?


Canon purposefully tweaked the sensor's color filter, so that more light can reach the sensor.

This is to minimize noise in general, not just banding.

Don't know exactly what they've done but it's something along the lines of Kodak's TrueSense sensor:


This image is copyrighted by the owner



With this arrangement, the RGBG color data in the raw file no longer contains accurate red, green, and blue colors.

The mazing is created by the raw-demosaicing algorithm - unless it's aware of the tweaked color filter, so that compensatation can be applied.

The problem is that even with compensation, there are still color artifacts - the salt & pepper noise.

That sensor uses Kodak's high ISO color array, which could be similar to what Canon is doing. Interestingly, in the comments from an article about this new array in 2007, Pete Myers had this to say about it and artifacts:

"Perhaps I am the only one in North America that finds the news of this new sensor to be dreadful, but I do. The original Bayer patent was based on a cell of two luminance pixels, and one red and one blue pixel---not two green, one red and one blue. It was a scheme originally developed for single tube vidicon television cameras. So in essence Kodak is returning back to the original Bayer patent, but since they could not get a new patent on old work (and collect royalties), the green element has been retained.

Part monochrome camera, remainder bayer camera. While the monochrome portion is going to be two stops faster, not so for the bayer portion. So to suppress noise, the monochrome information will be superimposed on the bayer info.

All I think its going to do is create even more artifacts in images.

I cant wait until some company finally comes to the same conclusion that the broadcast industry did a long time ago---make a 3-CCD camera and get beyond matrix based sensors.

I really think this is more about patent rights and Kodak trying to show something "progressive" in a failing company."

Here is the link

Nov 18, 2009 at 04:31 AM
dcains
Offline
Dedicated FM
Upload & Sell: On
p.3 #8 · 7D & 'Mazing' (serious question)


You've run out of things to bitch about while "testing" your gear, so now you're just making up new issues and even new words to identify them. That's OK, though, because there will always be a 7DII or 8D to keep you all busy for years to come. Any of you ever make prints and hang them on your walls?

Nov 18, 2009 at 05:48 AM
garycoleman
Offline
Upload & Sell: Off
p.3 #9 · 7D & 'Mazing' (serious question)


Tim Ashton wrote:
garycoleman wrote:
i think they forgot to put the letter A on mazing

OK. Here is yet another Nikon troll getting into the act

From what I have seen and heard from a couple of owners and a retailer I reckon Gary is very close to the truth

What I see on various FM sites only supports that. If I had Canon glass and were looking to / could afford to upgrade, I would be in there like a shot,

Tim


I love my 7d. IQ and AF are amazing. I don't need to perform these ridiculous tests to find artificially inflated problems.

Nov 18, 2009 at 07:00 AM
Fred Tedsen
Offline
Upload & Sell: Off
p.3 #10 · 7D & 'Mazing' (serious question)


jorkata wrote:
The mazing is created by the raw-demosaicing algorithm - unless it's aware of the tweaked color filter, so that compensatation can be applied.

There seems to be more to than that. If you process a RAW file in DPP, which shows no mazing, and then import the TIFF to Lightroom, the TIFF will show the mazing exactly like the RAW, even though Lightroom is not doing a RAW conversion on the TIFF.

Nov 18, 2009 at 02:14 PM
 



kewlcanon
Online
Dedicated FM
Upload & Sell: On
p.3 #11 · 7D & 'Mazing' (serious question)


Save some amunitions/critiques for 60D, 70D, 7D II, will you guys.. . Can some of you move to Nikon so you can "destroy" their cameras .

Nov 18, 2009 at 02:21 PM
Stoffer
Offline
Upload & Sell: Off
p.3 #12 · 7D & 'Mazing' (serious question)


The mazing can be reduced a lot i ACR when you use some Masking (which you should alomost always use anyway).

Note that I'm not saying that my 7D shouldn't be more clean in low ISO shots - after all my 5D certainly was - but you can use strong pre-sharpening without having to pay up with hefty artifacts in smooth tonal areas.

Anyway, it'll be interesting to see how ACR 5.6 will handle the 7D raw-files. I hope the best.

Nov 18, 2009 at 02:38 PM
Karl Witt
Offline
Dedicated FM
Upload & Sell: On
p.3 #13 · 7D & 'Mazing' (serious question)


I'm a 'mazed' what I learn when I drop in on this forum Thanks for the fine examples and comparisons guys!

Karl

Nov 18, 2009 at 02:54 PM
Daan B
Offline
Dedicated FM
Upload & Sell: Off
p.3 #14 · 7D & 'Mazing' (serious question)


Maybe a stupid question, but do you see the "mazing" when viewing at 100% (or smaller)? Or is 200% or greater needed?

And what about JPEG vs RAW?

Nov 18, 2009 at 02:57 PM
paulfeng
Offline
Upload & Sell: Off
p.3 #15 · 7D & 'Mazing' (serious question)


Stoffer - thanks for contributing constructively to the conversation, rather than thread crapping.

For the record: I don't know how significant (or not) "mazing" is for real world images, but pushing sliders to the max makes it clear that there are real differences between 7D bodies. True, many things viewable at 100% are unlikely to affect most prints.

However, when trying to eke out every last bit of image quality out of a capture, this mazing stuff may well make a difference to real images. What happens when you want to seriously crop a wildlife photo and print it big? At full rez, this crop is 3072 pixels wide from a 50D file (4752 pixels wide). If I want to print it 20" wide, that's about 150 pixels per inch. I hope that all 7Ds have the ability to squeeze out every last bit of potential quality from their pixels, for the best possible final printed image quality.

Why do some people feel compelled to come in here and tell others that they are wasting time and energy on this? This is an investigation in progress. May end up being significant, may not add up to a hill of beans. In my view, the people serious about investigating "mazing" are not jumping to the conclusion that there is some awful problem. But the people coming in here and mocking those who are curious about how the 7D works appear to have jumped to the conclusion that there can be no problem, or they refuse to admit, for whatever reason, the possibility of one.

Nov 18, 2009 at 03:31 PM
willis
Offline
Upload & Sell: On
p.3 #16 · 7D & 'Mazing' (serious question)


I'm pleased to say that after biting the bullet and getting a 7D I don't seem to have the mazing problem. my 7D (at 200%, upped contrast and a good dose of sharpening) looks very like the 50D crops that have been posted.
Overall IQ is as I expected, overall better than my old 30D but not in the 5D/1DIII league. AF seems faster than the 30D but noticeably inferior to the 1DIII so far (but I'm not fully familiar with it yet.)
I like the feel of the camera and it's got plenty of new gizmos to learn about. I dont get the feeling that ISO 3200 will print any larger than the 5D/1DIII with decent results depite the pixel advantage. I see myself using it mainly for bird phography when FL limited and when light enough to use 800 ISO or less (800 being a bit marginal as NR is already reducing resolution at that point).
I don't see any problems with image sharpness so far except for the progressive NR smearing from 800ISO+. 400ISO looks pretty sharp but a bit more grainy at the 100% level that my 30D.
Overall the 18MP IQ is better than previous crop cameras I've used but don't expect the 7D high iso viewed at 100% to look as clean as the 30D (of course there's an extra 10MP of detail to make up for it). I think in print it'll be better that previous 1.6 croppers but the increase in usable print size will be smaller than the increase in resolution suggests (ie we are seeing diminishing returns as resolution increases). Canon are valiantly battling noise but don't expect miracles.

Nov 18, 2009 at 04:24 PM
paulfeng
Offline
Upload & Sell: Off
p.3 #17 · 7D & 'Mazing' (serious question)


I apologize in advance, but it has to be said:

Gary Coleman and Willis in the same thread!

Sorry again. US-centric humor for persons of a certain age.

P.S. Glad to hear that you're both enjoying your 7Ds.

Nov 18, 2009 at 05:01 PM
willis
Offline
Upload & Sell: On
p.3 #18 · 7D & 'Mazing' (serious question)


What'chu talkin' 'bout?

Nov 18, 2009 at 06:20 PM
skibum5
Offline
Dedicated FM
Upload & Sell: Off
p.3 #19 · 7D & 'Mazing' (serious question)


willis wrote:
What'chu talkin' 'bout?

finally this thread has gotten somewhere

Nov 18, 2009 at 07:14 PM
Fred Relaix
Offline
Upload & Sell: Off
p.3 #20 · 7D & 'Mazing' (serious question)


After 6000 shots with my 7D, all processed with LR 2.5, I really enjoyed the picture quality. Beside capturing a lot of details, the noise has a true "film-like" quality that I really find very pleasing. I think that this is this specific noise that produces the mazing effect when viewed at 200% and pushing the sharpness all the way to the right (at 150, wheresas in real life I would almost never go over 50). Printed or even seen at 100% but with normal sharpness, the 7D files are just beautiful. Despite this mazing hidden within the pixels.

Edited on Nov 18, 2009 at 07:40 PM · View previous versions

Nov 18, 2009 at 07:32 PM
1   2  
3
   4   end
FM Forums | Canon-mount SLRs | Join Upload & Sell

1   2  
3
   4   end
    
 

You are not logged in. Login or Register

  Username   Password  
Lost your password?