Canon 300D RAW & Adobe questions.

[CousinMadness]

Greetings all. After 33 years of SLR photography I have joined the ranks of digital with a Canon Rebel. I am no novice to computers, Adobe, Gimp, Paintshop Pro... but have two basic questions.

I am a stickler for sharpness in an image, spoiled I guess since the first time I used a Hassleblad. This digital CMOS thingy and how the little piece of silicone translates an image into pixel/compression ratio before you compress the already compressed file in Adobe is driving me to distraction.

The primary use of my images is for viewing on the Internet. Maximum size and K factor about 1024x768 - 120K except for the odd wallpaper which would be more. So to my questions...

1) RAW. Am I wasting my time taking RAW images (7 meg average file yikes) for Internet viewing only? Will using the Canon 300D's maximum quality JPEG mode be sufficient for good quality shots?

2) Adobe. I am confused as to how I should manipulate an image that in it's original form is a staggering 3072 x 2048 pixels. Should I manipulate the (copy) in it's original size, then resize it? Or resize it first then fiddle with curves, sharpness saturation etc.

I would be very grateful for any advice as to how I may squeeze every last ounce of quality out of that little thumb sized CMOS sensor.

Thank you very much.


Martin.

[sparker1]

The idea behind using Raw is to allow you to post-process several factors that were traditionally fixed in the camera, such as exposure and white balance. Fixing those things after seeing the image should allow you to do a better job and produce a better image. The size of the file, whether Raw or JPG fine, is much larger than you need for the web. As you say, 1024 and 72-96 ppi is ample for web viewing. Start with the large file, post-process to get results you want, then re-size and Save For Web. BTW, you probably should save your source file (on CD or DVD?) in case you later want to print it, or even re-process.

[scottbergerphoto]

Shooting RAW gives you a 16 bit file to work with that give you lots of latitude in editing. Shooting in Adbe RGB gives you the ability to cature a larger range of colors then sRGB. That being said, if all you want to do is put your pictures on the web which only needs about 72 dpi and is only sRGB, you can shoot jpeg in sRGB. Personally I want the most potential for my images. I do the following:
1. Shoot Raw, Adobe RGB.
2. Open in PS CS. My Working Space is Adobe RGB. I adjust White Balance, Exposure and Color if needed in ACR(Adobe Camera Raw)
3. Convert to 16 bit tiff.
4. Levels or Curves, Hue/Saturation, Sharpening(USM).
5. Save the image as 16 bit tiff.
6. Image>Mode>Convert To Profile>sRGB
7. File>Save For Web>jpeg quality 60, Size-largest side 800>Save
Now I have 3 files, 1. the original Raw file, 2. a 16 bit tiff of my RAW file that I will keep forever and 3.a downsized image for the WEB.
Regards,
Scott

[PhotosGuy]

I would be very grateful for any advice as to how I may squeeze every last ounce of quality out of that little thumb sized CMOS sensor. Raw is the way to go. Before saving for the web, remember to sharpen the image with Unsharp Mask.

[Kennymc]

I always shoot in RAW... For web use I resize before any other work is done to the images because this saves time as you are dealing with a smaller file size... I work in layers so they can be easily altered or deleted... When all is finished the last thing I apply is the unsharp mask and I save as 30 quality jpeg...

[PhotosGuy]

Nice shot, Kenny. For web use I resize before any other work is done to the images because this saves time as you are dealing with a smaller file size... I did that at first. Then when I wanted to make a big print, I was screwed & had to redo all the work. Now I have a 160GB 2nd drive I got for $60 after the rebate. ;-)

[Kennymc]

Hi PhotoGuy...
That's why I wrote for web use... :) Images I need to print I convert to either PSD or Tiff files and store with the layers on DVD along with the DNG file (converted from either CRW or CR2 files)...
Thank for the compliment on the image...

[PhotosGuy]

I got that, but it looks like CousinMadness needs a lot of help & since he said, "The primary use of my images is for viewing on the Internet.", I wanted to point out a potential trap he could fall into. ;-)

[Kennymc]

True and appreciated...

[CousinMadness]

Thank you everyone for replying.

There is quite a bit of interesting information that I will have to print off and digest. The unsharp mask is a new animal for me... I confess I never knew what it was for.

So to summarize:-

* Shoot in RAW.
* Save a "copy" in tiff format (presumably because it doesn't lose any quality after manipulating?)
* Resize first (Smaller file = faster fiddling time)
* Then do the fiddling.
* Unsharp mask last?
* Then save as jpeg for web use.
* Original untouched RAW file still exists.

Incidentally I use Adobe CS.

Thank you everyone for the great advice.... and your patience with what is most probably second nature to you by now.

Now to go and experiment!


Martin.

[scottbergerphoto]

I see no point in editing the same file twice. If you do all your work on the 16 bit tiff, then all you need to do is change the color space to sRGB, resize for the web , and resharpen. Resizing before editing makes absolutely no sense. The destructive nature of editing becomes readily apparant.
Scott

[Kennymc]

But if you are only working for the web why work in large files in 16 bit if you are only going to change them to 8 bit and reduce the size of the image in the end?..

[PhotosGuy]

But if you are only working for the web why work in large files in 16 bit if you are only going to change them to 8 bit and reduce the size of the image in the end?.. It will reduce color banding for one thing, but you do have a point if 99.999% of what you do is for the web. You'll still have the RAW if you decide to make that 1 in a thousand print.

[scottbergerphoto]

But if you are only working for the web why work in large files in 16 bit if you are only going to change them to 8 bit and reduce the size of the image in the end?..
Take a 16 bit tiff and an 8 bit downsized jpeg for the web. Make the usual edits: crop, hue/sat, levels/curves. Then compare the histograms of the two files. Which one has a histogram with missing data?
Scott

[Kennymc]

I never mentioned changing them to 8 bit jpeg before working on them and apologise if I gave that impression... I would change them to 8 bit images from RAW 16 bit, halving the file size to start with... I would then reduce the image to the size I wanted for the web and work on that which would reduce the file size again... When finished I would save it as an 8 bit jpeg, remember if it's for web use you can't save a 16 bit jpeg, well not that I know of, so after all your work on your large files you are throwing away half the information by changing it to 8 bit, then you are throwing more info away by resizing it to the size you want for the wb just like I would have done before I started the manipulation... Below are two versions of the same image one done your way and one done mine...

[scottbergerphoto]

I prefer to throw the information away after I don't need it. In many cases you won't see a difference. But if there are exposure problems or color issues you probably will. It's hard to compare your images as the second one is much sharper.
Scott

[Kennymc]

True... The second one is much sharper because after throwing away all the information I would need to re sharpen it because of the loss of information due to downsizing... I'm not saying your way is wrong and my way is right just that there is more ways than one to do it...

[CousinMadness]

CONFUSION ?


Thanks again everyone for your continued suggestions, both online and in private e-mail. I confess to being even more confused now ... aaaaaargh!

Whether I shoot in RAW or largest JPEG capability of the Canon Rebel, am I correct is understanding that I can make a 16 bit Tiff from either?

Is a Tiff file by proxy a 16 bit file?

Should I fiddle with the appropriate sat/curves/sharpness with this Tiff file after I have resized it to it's maximum size (ie1024x768) - then - save a "copy" as an 8 bit JPEG for the web?

I am still confused as to the two camps of thought here. Some of you prefer shooting RAW then convert to TIFF fiddle, resize - then JPEG. Others shoot in large JPEG, convert to TIFF, resize, then fiddle.

The gentleman who posted the pictures of the Swan... the image on the bottom was much sharper. I believe this was fiddled with *after* resizing. It looked much better than the one fiddled with as a huge file.

Several folks alluded to the destructive nature of resizing... which leads me to another question (At 44 I am beginning to sound like the classroom dunce!).

My pictures will be posted via thumbnails - the viewer can click on it, and an image of 800x600 is sent. A larger image is viewable at 1024x 768 if required.

Now....

Some parts of my site will have features with photos. These pictures will not be a defined size. (Depending on how the article looks/how many words etc). Should I in this instance resize the image to what I want displayed.... eg.. 550x??? in Adobe for best results? Or will simply resizing a copy of my original 1024x768
8 bit JPEG be sufficient?

Thank you again to everyone for your patience. Quality is very important to me and I really appreciate your advice.


Martin

[Jon]

Well, neither of them has 16-bit colour data to begin with. RAW actually has 12 bits of data per colour and JPEG has 8 bits/colour, so your original conversion from RAW to TIFF will be less incomplete than your conversion from JPEG to TIFF. Subsequent editing, smoothing, etc. will be able to take advantage of the full 16-bit colour space, but you'd have to work harder to overcome the steeper steps between hues in the image converted from a JPg. If you want to work with JPG, stay with 8-bit TIFF, which is also an option. But if your intended target is the web, you're probably better off working in 8-bit TIFF all the way through if you start with JPEG. Shifting number of colours up and then back down after editing can cause you problems with shifts as there are likely to be minor shifts in colour as part of the editing process and resizing will collapse them back to what's nearest, not what's closest to what it was.

[jrm]

Like Jon pointed out, RAW is a 12-bit format.

When you use PhotoShop to "process" the RAW file, you can decide to open it into a 16-bit or 8-bit color space. Since your end use is relatively low resolution files for the web, I would agree that opening/working in 8-bit is probably the better choice here. (Although there is no absolute "correct" choice)

As to the Tiff/jpeg/file format question.... When you open (process) the RAW file in PhotoShop, you are not working on a TIFF or jpeg at that point. You are simply working on a file displayed by photoshop.

When you _save_ the file is where file format comes in. You cannot save your changes or edits back to RAW. The RAW file always stays the same. (However, PhotoShop will let you save the import settings as a separate file).

When you save as TIFF, the file is either uncompressed or compressed with a "lossless" method. This means that the next time you open the file, all the pixels will be exactly what you saved.

Jpeg, on the other hand, is a "lossy" compression method. It will create smaller file sizes, but does so at the expense of data. Basically, when you save a jpeg, some data is lost. It may be virtually unnoticeable at first. After several edits/re-saves it can become very noticeable.

"Whether I shoot in RAW or largest JPEG capability of the Canon Rebel, am I correct is understanding that I can make a 16 bit Tiff from either?"

You can open a jpeg in photoshop and then change the setting to 16-bit and save as tiff. However, there is no real reason to do so. The jpeg you opened only has 8-bits of info, and has already been jpeg compressed. There is no way to get back any extra bit-depth or information that may have been in the "original" file.

"Is a Tiff file by proxy a 16 bit file?"

Tiff files can be 8 or 16-bit.


"Some parts of my site will have features with photos. These pictures will not be a defined size. (Depending on how the article looks/how many words etc). Should I in this instance resize the image to what I want displayed.... eg.. 550x in Adobe for best results?"

Not sure what your questions is exactly, but...

If you have a "small" photo on a web page that links to a "large" version... you will need to create two different files - one for each size. You can resize and save all the versions from photoshop. Always start with the "large" version and create the smaller from that. For example:

Original photo size: 1600px x 1200px

1. open file, make edits, save as tiff, psd, etc. (any non-lossy format). This is now your "Original"
2. Resize Image to 800px x 600px (make sure resample is on in Image Size dialog)
3. "Save As" large.JPEG - this is now your large version.
4. Close Window, open "original" from step #1.
5. Resize Image to 160px x 120 (make sure resample is on in Image Size dialog)
6. "Save As" small.JPEG - this is now your small version.

You could eliminate some work by using "save a copy" and "undo." This is just a down and dirty explanation.

--Joe

[PhotosGuy]

4. Close Window, open "original" from step #1. There's a potential trap here. When you close the file after "Save As" large.JPEG, PS may prompt you asking "Do you want to save the file"? The answer is no, since it would save the file you're working with as a jpeg too. 4. Close Window, open "original" from step #1. You could just back up in History to before you resized it.

[jrm]

There's a potential trap here. When you close the file after "Save As" large.JPEG, PS may prompt you asking "Do you want to save the file"? The answer is no, since it would save the file you're working with as a jpeg too. You could just back up in History to before you resized it.


PhotosGuy,

You are right about the trap. I should have accounted for that.

However, there are much simpler ways to do this. History, Save a copy, or even recording a simple script if you are going to repeat the process often. I was trying to keep it as simple as possible, not knowing the original poster's level of expertise in PS.

Again, thanks for pointing this out. You do NOT want to overwrite your "original" with a scaled down or jpeg version.

--Joe

[malla1962]

its the printing with photoshop i cant seem to handle.

[jrm]

its the printing with photoshop i cant seem to handle.

? What do you mean? Can you be more specific (System, OS. printer model, problem details?)

--Joe

[malla1962]

os xp pro
photoshop cs
canon 20d
canon i965 printer
problems i am having are when i print a tiff out of fotoshop the prints are nothing like on
screen,but if i print jpg using easy photoprint theres no problems.i have my camera set to record srgb.

[rdenney]

I see no point in editing the same file twice.

One thing that helped my workflow considerably was drawing the distinction between correction and targeting.

When I convert the image from RAW, I'm correcting. I use the histogram display and try to spread my values out as much as possible in the conversion, to make the best use of my 16 bits, and I make white balance corrections (the best argument for using RAW, in my opinon, other than keeping the file in a 16-bit color space). Then, in PS, I make all the adjustments to make an image that matches my visualization. This includes dodging, burning in, correcting color (to my calibrated screen, and therefore to a standard color definition), spotting, and all the other things we do to make the image look the way we want it to from an artistic perspective. I will also apply a bit of sharpening to counteract the natural softening effect of the sensor, using Fred Miranda's stuff.

Then, I store the image. Now, I have the original CRW file and a 16-bit, full-resolution Adobe PSD file that has been manipulated to create the image I visualized. It might be CRW_4531.CRW and RicksHouse.PSD.

Then, I target the image to the particular output device. If the output device is a printer, I crop to the print dimensions I want, resize if necessary to achieve my usual standard 240 pixels/inch on the print. Then, I sharpen it. Finally, I turn on the "RicksEpsonPrinterColorProfile" preview and adjust the color, contrast, and saturation to achieve what I want, and make test prints. This is not color correction, but rather color targeting for a specific output device. When I'm satisfied with my print, I save the file as RicksHouse8x10Epson.PSD.

Note that sharpening, other than to correct the basic fuzziness of the sensor, is a targeting activity. It's done AFTER you resize the image.

Saving for the web is just like targeting for a printer. I resize the image for web display and sharpen it. I then turn on the "WindowsRGB" preview, and usually adding about 20 points of saturation will adjust that more limited color space to look like my original, which is is always in Adobe1998 color space. The WindowsRGB preview lets me see the image in the limited colorspace used for monitors, without any of the corrections of my color calibration. Again, that's part of targeting to a specific output device. I then save the file as RicksHouseLores.PSD. Then, I use the "Save for Web" feature of Photoshop. This will create a JPEG file with no assigned color space, which is perfect for web display. I save that file as RicksHouseLores.JPG.

Now, I have four files. They include the original raw file and the full-resolution corrected image file, plus two files, one targeted for my Epson printer and the targeted for web display.

By keeping correction and targeting processes separate, I always start with my final image when I decide to send it to a new output device, applying only the corrections needed by that device. That way, the output from all the devices looks as much as possible like the SAME original image. But it does involve two sharpening process, though for different reasons and in different ways. I don't really think of that as editing twice.

Think of it as making an internegative. You might undergo quite a process, including contrast masking and color corrections, to create an internegative of a transparency, but once you have it, you can go into production on prints without having to apply ALL the corrections to EVERY print.

I've worked as a professional photographer quite a bit in my life, but I'm mostly an amateur now and my processes have to be efficient as possible if I'm going to have any fun at all, heh, heh.

http://www.rickdenney.com/photography.htm

Rick "a flow theoretician by training" Denney

[jrm]

os xp pro
photoshop cs
canon 20d
canon i965 printer
problems i am having are when i print a tiff out of fotoshop the prints are nothing like on
screen,but if i print jpg using easy photoprint theres no problems.i have my camera set to record srgb.

I am also using a Canon printer with PSCS. My camera is a DRebel, and OS is Mac OS X, so there may be some differences in our settings. There is no reason that you shouldn't be able to get excellent results with that setup.

The first thing to check is your color settings. PhotoShop is probably set to manage color, and then your Canon driver is doing a second color management. This is the most common cause I have found for PhotoShop prints differeing from the screen/other programs.

Select "print" like you normally do. Then in the "Print Dialog" that appears, there should be several choices and/or a drop down menu. This is where you would normally choose paper settings, adjust color, etc. Somewhere there should be a setting for "Color Correction" (on the Mac it is under the "Color Settings" tab). Set this to "none."

This should immediately improve your situation.

If this is already set to "none" or the output doesn't improve, then you have to look elsewhere. Some things to do:

-calibrate monitor
-check color settings in photoshop (including profiles used, etc.)
-check printer settings (make sure paper type matches paper used, etc)

Hope this helps

--Joe