Reading mode: Light Dark

Yes or no: 7d mRaw = 40d Raw natively (same iso, same conditions)

Go to last post
M

merlin2375

Active member
Messages
99
Reaction score
0
Location
US
The 7D manual says that a mRaw is about 10.1 megapixels. The 40d is 10.1 megapixels natively. If you take the same picture, under the same conditions, using the same high iso (as high as is available in both)?

Will the noise be the same?

If yes , then why are people upset over 18mp.

If no , then what is actually happening in the camera and why is mRaw not an acceptable solution for some high iso shooting.
 
merlin2375 said:
The 7D manual says that a mRaw is about 10.1 megapixels. The 40d is 10.1 megapixels natively. If you take the same picture, under the same conditions, using the same high iso (as high as is available in both)?

Will the noise be the same?

If yes , then why are people upset over 18mp.

If no , then what is actually happening in the camera and why is mRaw not an acceptable solution for some high iso shooting.
Click to expand...
No. MRAW isn't raw, it's a demosaiced file, thus you can't use your own demosaicing algorithm with it. Thus, it's not the same as a 40D raw.

Performance might be the same or similar, but the above is a key difference you didn't mention.

--
Lee Jay
(see profile for equipment)
 
but in reality 7D could be evne better because it has two years newer technology. People who are upsetting over more pixels are the same kind of people who are scared of the the "death panel" in medical care reform.
merlin2375 said:
The 7D manual says that a mRaw is about 10.1 megapixels. The 40d is 10.1 megapixels natively. If you take the same picture, under the same conditions, using the same high iso (as high as is available in both)?

Will the noise be the same?

If yes , then why are people upset over 18mp.

If no , then what is actually happening in the camera and why is mRaw not an acceptable solution for some high iso shooting.
Click to expand...
 
I was also wondering what exactly is Canon doing inside the camera when you switch the resolution from large to medium or small? I always thought they used in camera down-sampling. But I am not sure.
 
7D mRAW will be slightly better than 40D RAW, SNR-wise.
merlin2375 said:
The 7D manual says that a mRaw is about 10.1 megapixels. The 40d is 10.1 megapixels natively. If you take the same picture, under the same conditions, using the same high iso (as high as is available in both)?

Will the noise be the same?

If yes , then why are people upset over 18mp.

If no , then what is actually happening in the camera and why is mRaw not an acceptable solution for some high iso shooting.
Click to expand...
--

. 。o O o 。 . 。o O o 。 . 。o O o 。 .
 
I'm very interested to find out the exact same thing.

The 50D and 5DII mRAW files are slightly cleaner (and of course have less detail) than the full RAW files.

The 7D mRAW size is identical to the 40D full RAW size. Considering that at full resolution the 7D is basically as clean as the 40D, it's not unrealistic to expect that the mRAW will be slightly better - both for noise and detail.
We'll see. Waiting for tets.
 
The most straightforward explanation is “binning”, e.g. combining signals from several pixels into one. The simplest way is just to add up two (or four) adjacent pixels but I doubt it’s all that Canon is doing. The result and processing efficiency are the two primary concerns. It is not unthinkable that some noise messaging can be done at the same time of binning.
 
carlk said:
The most straightforward explanation is “binning”, e.g. combining signals from several pixels into one. The simplest way is just to add up two (or four) adjacent pixels but I doubt it’s all that Canon is doing. The result and processing efficiency are the two primary concerns. It is not unthinkable that some noise messaging can be done at the same time of binning.
Click to expand...
That's not what they do. They demosaic and downsample. I don't know if any other processing is done in the noise or sharpening areas.

--
Lee Jay
(see profile for equipment)
 
ljfinger said:
carlk said:
The most straightforward explanation is “binning”, e.g. combining signals from several pixels into one. The simplest way is just to add up two (or four) adjacent pixels but I doubt it’s all that Canon is doing. The result and processing efficiency are the two primary concerns. It is not unthinkable that some noise messaging can be done at the same time of binning.
Click to expand...
That's not what they do. They demosaic and downsample. I don't know if any other processing is done in the noise or sharpening areas.
Click to expand...
Are you sure about that?

Seem logical to me that they'd try 'readout binning' - that is, combining the output of multiple pixels after the signal has been read.

If they indeed do in-camera downsampling, then there are no SNR gains in using mRAW.
 
That is just binning except not straight adding up of signals but use certain algorithm to put weight of nearest neighbors and perhaps beyond. There are a lot of tricks can be played in signal processing the only limitation is processing power.
ljfinger said:
carlk said:
The most straightforward explanation is “binning”, e.g. combining signals from several pixels into one. The simplest way is just to add up two (or four) adjacent pixels but I doubt it’s all that Canon is doing. The result and processing efficiency are the two primary concerns. It is not unthinkable that some noise messaging can be done at the same time of binning.
Click to expand...
That's not what they do. They demosaic and downsample. I don't know if any other processing is done in the noise or sharpening areas.

--
Lee Jay
(see profile for equipment)
Click to expand...
 
GeorgeML said:
ljfinger said:
carlk said:
The most straightforward explanation is “binning”, e.g. combining signals from several pixels into one. The simplest way is just to add up two (or four) adjacent pixels but I doubt it’s all that Canon is doing. The result and processing efficiency are the two primary concerns. It is not unthinkable that some noise messaging can be done at the same time of binning.
Click to expand...
That's not what they do. They demosaic and downsample. I don't know if any other processing is done in the noise or sharpening areas.
Click to expand...
Are you sure about that?
Click to expand...
Pretty sure. Second hand, not first, but from an ultra-reliable source.

--
Lee Jay
(see profile for equipment)
 
Canon's material seems to be very coy on this subject. This is a big question that it would be nice to see an official answer on.
ljfinger said:
GeorgeML said:
ljfinger said:
carlk said:
The most straightforward explanation is “binning”, e.g. combining signals from several pixels into one. The simplest way is just to add up two (or four) adjacent pixels but I doubt it’s all that Canon is doing. The result and processing efficiency are the two primary concerns. It is not unthinkable that some noise messaging can be done at the same time of binning.
Click to expand...
That's not what they do. They demosaic and downsample. I don't know if any other processing is done in the noise or sharpening areas.
Click to expand...
Are you sure about that?
Click to expand...
Pretty sure. Second hand, not first, but from an ultra-reliable source.

--
Lee Jay
(see profile for equipment)
Click to expand...
 
GeorgeML said:
ljfinger said:
carlk said:
The most straightforward explanation is “binning”, e.g. combining signals from several pixels into one. The simplest way is just to add up two (or four) adjacent pixels but I doubt it’s all that Canon is doing. The result and processing efficiency are the two primary concerns. It is not unthinkable that some noise messaging can be done at the same time of binning.
Click to expand...
That's not what they do. They demosaic and downsample. I don't know if any other processing is done in the noise or sharpening areas.
Click to expand...
Are you sure about that?

Seem logical to me that they'd try 'readout binning' - that is, combining the output of multiple pixels after the signal has been read.

If they indeed do in-camera downsampling, then there are no SNR gains in using mRAW.
Click to expand...
With the Bayer pattern filters, they can't do true on-chip 2x2 pixel binning, which would yield only one read noise hit per 4 pixels worth of signals. It would wipe out all color information. It would also go from 18mpix to 4.5 mpix.

By demosaicing and downsampling, they are doing nothing more than what can be done post processing (other than having smaller raw files).

--
Rick Krejci
http://www.ricksastro.com
 
Please see my post below. Downsampling and binning are probably just two different way of describe the same thing. The downsampling mentioned should be done after read out otherwise it does not make sense at all.
GeorgeML said:
ljfinger said:
carlk said:
The most straightforward explanation is “binning”, e.g. combining signals from several pixels into one. The simplest way is just to add up two (or four) adjacent pixels but I doubt it’s all that Canon is doing. The result and processing efficiency are the two primary concerns. It is not unthinkable that some noise messaging can be done at the same time of binning.
Click to expand...
That's not what they do. They demosaic and downsample. I don't know if any other processing is done in the noise or sharpening areas.
Click to expand...
Are you sure about that?

Seem logical to me that they'd try 'readout binning' - that is, combining the output of multiple pixels after the signal has been read.

If they indeed do in-camera downsampling, then there are no SNR gains in using mRAW.
Click to expand...
 
carlk said:
Please see my post below. Downsampling and binning are probably just two different way of describe the same thing.
Click to expand...
Binning is a horribly poor way of averaging and interpolating. Downsampling can use far more sophisticated methods. You can't really bit at the pixel level with Bayer sensors and still have useful information.

--
Lee Jay
(see profile for equipment)
 
I recently found this:
http://forums.dpreview.com/forums/read.asp?forum=1019&message=26234685
and
http://lclevy.free.fr/cr2/#sraw
and
http://luminous-landscape.com/forum/index.php?showtopic=37192&hl=sRAW

Summary by digitaldog in that last link:

In short, every 2 sRAW pixels contain: 2 Y, 1 Cb and 1Cr samples, while the corresponding RAW image area (8 pixels) contains 4 Green, 2 Blue and 2 Red samples. All samples are in 14bit. This explains why 1/4 the picture area carries 1/2 the data volume, and why sRAW has very high per-pixel sharpness. However, we still have no idea how this down-sampling is performed
--
Robert
http://www.pbase.com/robert_michael
 
This is a general discussion without measuring the image quality of the two cameras at the same detail level (frequency of signal and noise) practically.

The 7D image down sampled to 10MP will have higher detail level then the native 10MP 40D image.

Don't forget that the Bayer CFA approach cost image quality. Even though different amount for different kind of details in the image.

Therefore the 7D 10MP image will contain higher frequency noise too, embedded with the higher frequency details.

If you down sample the 7D image to 5-8MP then you will have about the same detail level as the 40D (dependent of the characteristics of the subject), then you will also have the same noise level.

In the end you are pretty safe to get the 7D if you want "40D-image quality" ... or better ...

--
Henrik
 
Did you read my other posts? Canon does say "binning" in the literatures which is essentailly downsampling AFTER read out. There are a lot of sophisticateion algorithms for data bining I have personal experience with doing some bining of astronomical data with the most advanced computer (at that time ;-) ) many years ago.
ljfinger said:
carlk said:
Please see my post below. Downsampling and binning are probably just two different way of describe the same thing.
Click to expand...
Binning is a horribly poor way of averaging and interpolating. Downsampling can use far more sophisticated methods. You can't really bit at the pixel level with Bayer sensors and still have useful information.

--
Lee Jay
(see profile for equipment)
Click to expand...
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top