Basic Stuff: Low res Jpegs for Website

What I would do in your shoes is to always work with high res images and save the work in PSD or TIFF format. you can then resize the image to whatever is needed for posting on the web as a JPEG file. If you need to re-edit the image, go back to the PSD (or TIFF) file and work on that. Never work directly with the low res image itself.

That's good advice.
How much are you compressing your images for the web?
And, do you know that each time you save (& recompress) a jpeg, you're going to lose even more image information?
Always work on a copy of the background image, not the original layer.
Use Adjustment layers. They don't add to the file size & the changes can be easily reversed.
If you're adding things to the image, put them on they're own layers.
When you're done, save it with the layers.
Then flatten, resize, sharpen & Save For the Web to a different folder. I don't think that PS6 automatically changes the filename, so be sure that you don't overwrite the original pic.

Do a search on "jpg +degradation" and see what turns up.

Compression is when a program (could be a graphic program or a program like WinZip) takes a file and tries to create some sort of code to represent the data using fewer bytes than the original. With text files, it's easy because there's so much repetition of characters. With other types of data files it becomes increasingly more difficult as the data itself becomes less and less repetitive. Some file types don't compress much at all. RAW picture files have a compression rate of less than 10%.

JPG as a graphic compression routine was developped to take care of that problem, but the much higher compression (a picture may be around 8MB as a RAW file and end up as only 3MB or so as a JPG file) comes at a price. That price is to lose some of the data. The same thing happens with MP3 files compared with WAVE file or CDA files.

When the JPG standard was being worked on, the developpers had to decide what data was "expendable" -- in other words, what parts of the picture would the typical human eye/brain not really notice if it was removed. I'm not an expert, so this part may be a bit incorrect, but the theory is correct -- the transition pixels between different neighboring colors were the first bits of data that they realized they could live without. The human eye/brain does a great job of filtering out unwanted information (how many times have you taken a picture of a beautiful landscape, only to look at the picture you took and notice nothing but telephone wires?) and also of providing data that is missing in the original. When the transition bits are removed, the human eye/brain puts them back in when the picture is viewed.

However, each time you save and resave the picture as a JPG, more and more of those transition bits are removed each time until a point is reached where the picture really degrades, and looks blotchy, blocky and in a word, terrible.

The reason the picture from the cheap, lower megapixel camera looks better at the same size as your edited picture is that it only took a limited number of pixels originally so very little is lost since the picture is saved only once, while you have to get rid of a lot of data to get to that size. And if you have edited/saved, edited/saved over and over, each time you saved you lost valuable data.

My advice would be to use the ReSize function to get the picture to the size you want, after having done all the work you want (resaving as often as you want as zero-compression TIFF files). In PaintShopPro, I can take a 2.8MB jpg file and simply by resizing it to 50% and saving it once as a jpg (with compression factor of 20) it shrunk to 359KB with no discernible change in quality. But I saved it only ONCE.

Unfortunately PaintShopPro9 doesn't support the RAW file format from my new 20D so I have to wait before I can save in the camera as RAW files and then work on them in that format in PaintShopPro and do my final JPG save only once for a decent size and a minimal degradation. If PhotoShop supports the RAW format from the 10D, you might consider saving as RAW/jpg-Large-Fine to get two different pictures to work with, and then do the PhotoShop work on the RAW file, saving to jpg only once, after you have the picture just as you want it. But if you want to shoot only as jpg, make sure it's the Large-Fine format.

I hope this explanation helps get you on the right path.

In brief:
1) Open the picture in PhotoShop and immediately save as TIFF.
2) Open the TIFF and do all the editing, resaving each time as TIFF.
3) When all is as you want it to be, resize the picture to the size you need for the web-site.
4) Save the FINAL TIME ONLY as jpg in the type you want (interlaced, non-interlaced, whatever) to minimize degradation.

I apologize for being so wordy -- I hope I haven't offended anybody by the length of this post.

Do a search on "jpg +degradation" and see what turns up.

Compression is when a program (could be a graphic program or a program like WinZip) takes a file and tries to create some sort of code to represent the data using fewer bytes than the original. With text files, it's easy because there's so much repetition of characters. With other types of data files it becomes increasingly more difficult as the data itself becomes less and less repetitive. Some file types don't compress much at all. RAW picture files have a compression rate of less than 10%.

JPG as a graphic compression routine was developped to take care of that problem, but the much higher compression (a picture may be around 8MB as a RAW file and end up as only 3MB or so as a JPG file) comes at a price. That price is to lose some of the data. The same thing happens with MP3 files compared with WAVE file or CDA files.

When the JPG standard was being worked on, the developpers had to decide what data was "expendable" -- in other words, what parts of the picture would the typical human eye/brain not really notice if it was removed. I'm not an expert, so this part may be a bit incorrect, but the theory is correct -- the transition pixels between different neighboring colors were the first bits of data that they realized they could live without. The human eye/brain does a great job of filtering out unwanted information (how many times have you taken a picture of a beautiful landscape, only to look at the picture you took and notice nothing but telephone wires?) and also of providing data that is missing in the original. When the transition bits are removed, the human eye/brain puts them back in when the picture is viewed.

However, each time you save and resave the picture as a JPG, more and more of those transition bits are removed each time until a point is reached where the picture really degrades, and looks blotchy, blocky and in a word, terrible.

The reason the picture from the cheap, lower megapixel camera looks better at the same size as your edited picture is that it only took a limited number of pixels originally so very little is lost since the picture is saved only once, while you have to get rid of a lot of data to get to that size. And if you have edited/saved, edited/saved over and over, each time you saved you lost valuable data.

My advice would be to use the ReSize function to get the picture to the size you want, after having done all the work you want (resaving as often as you want as zero-compression TIFF files). In PaintShopPro, I can take a 2.8MB jpg file and simply by resizing it to 50% and saving it once as a jpg (with compression factor of 20) it shrunk to 359KB with no discernible change in quality. But I saved it only ONCE.

Unfortunately PaintShopPro9 doesn't support the RAW file format from my new 20D so I have to wait before I can save in the camera as RAW files and then work on them in that format in PaintShopPro and do my final JPG save only once for a decent size and a minimal degradation. If PhotoShop supports the RAW format from the 10D, you might consider saving as RAW/jpg-Large-Fine to get two different pictures to work with, and then do the PhotoShop work on the RAW file, saving to jpg only once, after you have the picture just as you want it. But if you want to shoot only as jpg, make sure it's the Large-Fine format.

I hope this explanation helps get you on the right path.

In brief (not knowing how PhotoShop would work with the RAW file, this outline if based on your saving as jpg):
1) Open the picture in PhotoShop and immediately save as TIFF.
2) Open the TIFF and do all the editing, resaving each time as TIFF.
3) When all is as you want it to be, resize the picture to the size you need for the web-site.
4) Save the FINAL TIME ONLY as jpg in the type you want (interlaced, non-interlaced, whatever) to minimize degradation.

I apologize for being so wordy -- I hope I haven't offended anybody by the length of this post.

My most humble apologies for sending my lengthy post twice. It didn't look as if it was going through, so I hit my browser's STOP button and then decided to reword something at the end and then hit Submit again. To my horror, even though it didn't appear to do so, it went through the first time as well.

Sorry.

It depends. It could depend on the webmaster, web engine used, etc. I think the best thing is to find out the exact dimensions of the full-sized image to be displayed on the website. Also, check if there are any file size limitations. Then, save your pictures so they meet that criteria. That way, your pictures will be untouched and you retain full control.

Example, the photo gallery on my website is configured to automatically resize any photo that's more than 640 pixels on its longest side. It will also automatically compress the files. What I do is upload photos that are already 640 pixels on the longest side so the files are left alone.

I don't mind you copy/paste if it helps you.

As for web-sites compressing photos further, that would depend on how the web-site is set up. But repeated viewings shouldn't further degrade the photo -- once it is placed on the web-site the compression should already be done, and any repeated viewings should all show the same image. I can't imagine a situation where viewing the photo once would allow it to be resaved so it would further degrade.

Or perhaps I misunderstand your question?