More on RAW / Poll

[defordphoto]

Originally Posted by maderito

Jim - Thanks for getting that right - I did mean silent majority. BTW - I like your stats. They make more sense than your average scientific poll

HAHA! Too funny!!

[rstobbe]

I do 100% RAW with my D30. If you are going to adjust levels/contrast later the tonal range of the output is richer when doing this adjustment in 12bit instead of 8bit per channel. If I want to have Jpegs, hey BreezeBrowser or FVU will give me the same result that the Camera would have given me when shooting Jpeg, so why should I go for one when I can have both

Well, except for speed ... but I am not a sports photographer.

ciao
-- Ruediger

[msnow]

Originally Posted by RFMSports

Ilya: Interesting results on your test. The CS shot is world's noisier than the FVU and JPEG. The JPEG and FVU shot are nearly identical. Thanks for posting those.

I too am very interested in seeing the results from others. I attented a seminar a couple of months ago where the Canon representative told the group that he had a private conversation with Thomas Knoll (author of Photoshop) where Knoll confided in him that Adobe's camera RAW converter was "not as good" as Canon's "yet". Now, I have no way of knowing if this "conversation" really took place or not but the rep went on to say that most people inside Canon believe Knoll's next version (update to PSCS) will most likely exceed their own conversion application. He indicated that Knoll sort of rushed the PSCS RAW convertor for marketing reasons.

If the above is true, however, I would expect that the test results of those convertors using Canon's SDK (e.g. FVU and Breezebrowser) would have better results than PSCS. Having said that, I still use C1 because I "percieve" the results to be better than any of the above and I like the extra cptions and workflow.

[Vegas Poboy]

This is a very hot topic & so far I've learned alot even though I voted I shoot 90% of the time. It's hard to get a true count when people don't give there input but overall I think that alot of Canon owners have the DSLR camera but not the photography background. Meaning had money to spend and it make them looks good when they pull it out of the bag, so they are learning from our post & not inputing. In fact I know of a few who have the camera & treat it as a point & shoot & they have know clue of software editing or how to edit the RAW formats. It also may be a case of if they only own PS2 they can't afford extra software to edit the RAW format.

I did due some researching last night and found a script that is used in PSCS and it has changed my mind to shooting more RAW and I'm only using two software programs. ACDsee for quick browisng and emailing to private friends & PSCS, I tried Breezebrowser but was not comfortable with it.
Anyway go to Adobe forum http://share.studio.adobe.com/Default.asp or
http://www.russellbrown.com/body.html. Dr. Brown's Image Processor, I loaded it up and it works great, view the AVI file first but this is working for me. It will save time and give I feel most of us what we want out of our RAW files.
Let me what you think

[UK_Terry]

RAW 100% when useing 10D

JPG 90% RAW 10% when useing S45.

[CyberDyneSystems]

Vegas,..

I use the same tools 90% of the time,. ACDsee and then PS (originally 7.01 and ACR,. now CS)

I will try out that script,. that was a great link to Adobe's site,. I had not seen that page.

FYI. Two of the first three "scripts" listed were created by our man Roger Cavanagh! 8)

[CyberDyneSystems]

I just posted this in another thread which has turned into yet another RAW Vs. jpeg dissection... After posting it I felt it would fit here better than it did in the other thread

...So, I hope no minds the repetitiveness.

I have yet to shoot anything like a ruler,. or do any side by side tests...

I go strictly on what works for me while doing the photography I enjoy (mostly wildlife and the occasional theatrical event)

I'm pretty sure that in all the times I have championed shooting in RAW, or any peice of equipment or software,. I have allways done so from the point of view of what either does work for me,. or what I suspect may work for the individual style of the person asking advice...

I do not beleive that shooting a side by side comparison will reveal flaws in the jpeg,. rather it is in the ease of processing the RAW file to a more complete image than the processing the camera applies that will reveal the fundamental difference of the two file types.

It is a strange contradiciton that as every time Canon realeases a new less costly DSLR we get dozens of posts from individuals who ask "why are my pictures soft? why are they flat? where's the color saturation? where's the contrast...etc.."

......that we in unison respond,. "The DSLR is not like the point and shoots,. it doesn't do it all for you." "this is in fact an advantage,. the processing has been left to you to decide in you editing software"... etc

To offer up this principle to the proper use of the Canon DSLR and then backpedal on this argument when the perenial jpeg Vs. RAW question rears it's ever so ugly head is a contradiction.

If you like the way the camera processes the image,. then jpeg is indeed just as good as RAW,. in fact in that case, the jpeg benifits DO outweigh those of RAW.

But if you think you can do a better job of post processing than the camera can (like Canon itself recomends here: http://www.cps.canon-europe.com/arti...articleId=1240 )
Then RAW is by far the best choice.

So the only side by side I can offer is this.

Any shot I have taken that I am in any way proud of,. if I take the RAW file and do my own post processing work to create the final image,. that final image WILL be a much more pleasing photograph than the embedded jpeg,. on it's own or after post processing on the jpeg. The jpeg simply will not offer the same flexibility to create the final peice of artwork that the RAW fill will.

If anyone is interested,. this is a test I can easily post the results of,. but the problem is it won't alter any ones perceptions and the results really only matter to me!

But they do mater to me a great deal,.

[imago57]

When I was shooting film (positive) in 35mm, 2 1/4 X 2 1/4 and 4x5 I always bracketed, because I know I am not perfect, neither is my camera or my lightmeter. So now I shoot RAW, for the same reason. It is a great tool when used properly, and since I never shoot sport I really don't have to be concerned about buffer and speed. Just my 2 cents

[PacAce]

Originally Posted by PacAce

Yes, very interesting, Ilya. When I first tried out RAW processing in CS, I thought there was something "different" about the converted images which I couldn't quite place my finger on. Maybe you just shed some light on that. I'm going to do my own tests to see for myself.

Ha! Maybe "Go Canon!" doesn't just apply to lenses, eh?

OK, folks, check this out. I got one of my RAW images that was shot at 1600 during the day when I forgot to reset it back down to 100 or 200. Here's a 100% crop of a small portion of it, converted using FVU as shot. As you can see, the noise is very evident.




Now, here is another crop of the same area of one that was converted using PS CS, also converted as "camera default". The Color Noise Reduction was set to 0. As you can see, the noise level is comparable to FVU's.




However, CNR=0 is not the default value for PSCS (at least not one mine). The next one was converted, again with PSCS, but with Color Noise Reduction set to it's default value 25. If there's any noise, it's really hard to tell compared to the previous images.





And lastly, one that was converted with CNR set to 100. I can't tell the difference here between the 25 and the 100.



With the CNR set to 25 in PSCS, the color noise level on the image is a lot less than of that from FVU. However, I did notice that, in "Camera Default" mode, the contrast of the images are not as well defined as that of FVU but I guess that can be fixed in PS itself.

I guess I'm back to using PSCS for RAW conversion again.

[Guillermo Freige]

I'm shooting only in RAW since day one with my S50 (and RAW capability was one of the main reason for buying that camera), and of course also use RAW with the dRebel. Besides all the already said here (WB correction, 12-bit linear mapped in 16-bit nonlinear colordepth instead of 8-bit one, exp.compensation, lack of compression artifacs if saved as TIFF), one of the advantages of RAW is you can "improve" the camera just using different RAW converters. For example PS CS adds AdobeRGB and exp.compensation to my S50 RAW files (canon soft doesn't support them), and C1Rebel and PS CS add different noise reduction, exp.compensation and ICC profiles to the dRebel. You can check some of my comparisons between different converters in my faq. The 30min exposure time examples are pretty impressive.
The faq is in
http://gfreige.homelinux.org/faq/

[ilya]

Leo, can you post a small crop, say the ear or something that gets you up close. I was using CNR at 25 in my test, and even then it was still noisy. What are your exposure, shadow, brightness, contrast sat, sharpness and lum settings?

[PacAce]

Originally Posted by ilya

Leo, can you post a small crop, say the ear or something that gets you up close. I was using CNR at 25 in my test, and even then it was still noisy. What are your exposure, shadow, brightness, contrast sat, sharpness and lum settings?

Ilya, that was the small, tight crop. I can't get any tighter than that without exceeding 100% and going into artificial zoom.

All my settings were set to whatever the default settings are when using "Camera Default". The only setting I changed was the CNR to show the effects at different settings. But here are the numbers anyway:

Adjust:
WB=As Shot
Temp=5050
Tint=+25

Exposure=0.00
Shadows=0
Brightness=50
Contrast=+25
Saturation=0

Detail:
Sharpness=25
Luminance Smoothing=0
Color Noise Reduction=25

And all Calibration settings are 0.

And here's the resized full frame of the test picture (with some levels, saturation and USM applied for presentation).

[ilya]

Leo,

That's a cute picture by the way, with the kid looking up at the horse like that.

On my screen, to my eye, your PS CS crop looked a bit noisier. It helps to get in more then 100% to see the diff. I'm not sure what you mean by artificial zoom? PS doesn't do any kind of interpolation to a cropped image, it is what it is, but closer.

In any case, the bottom line is that I'm trying to figure out if the PS CS noise factor is because something I'm doing wrong, and this can be corrected by playing around with settings, or that the upgrade itself is at fault - as compared to say C1 or BB. If anyone else has done comparisons from CS to C1 or BB, pls let me know what you found.

Thanks
Ilya

[PacAce]

Originally Posted by ilya

Leo,

That's a cute picture by the way, with the kid looking up at the horse like that.

On my screen, to my eye, your PS CS crop looked a bit noisier. It helps to get in more then 100% to see the diff. I'm not sure what you mean by artificial zoom? PS doesn't do any kind of interpolation to a cropped image, it is what it is, but closer.

In any case, the bottom line is that I'm trying to figure out if the PS CS noise factor is because something I'm doing wrong, and this can be corrected by playing around with settings, or that the upgrade itself is at fault - as compared to say C1 or BB. If anyone else has done comparisons from CS to C1 or BB, pls let me know what you found.

Thanks
Ilya

Ilya, thanks for the compliment. I thought it was cute, too. That little kid just happened to walk into the picture and look at the horse like that as I was taking pictures of the horse. It's what I call serendipity. It doesn't happen often but when it does, it "makes" the picture, if you know what I mean.

Now, getting back to the crops, I guess what I was trying to say is that what you see there are crops of the actual pixels of the original, not resized crops.

I had 3 samples of the PSCS images, with CNR=0, CNR=25 and CNR=100. Which one were you referring to when you said that the PSCS image looked noisier than the FVU on your screen? Or were you referring to all of them?

[CyberDyneSystems]

Well,. all I can say,. is damn fine looking grain for 1600!!!!!!