Photoshop image resizing -- great loss of detail

bogleric

TPF Noob!
Joined
Oct 23, 2003
Messages
932
Reaction score
1
Location
North Carolina, USA
I have encountered an issue with photoshop during the resizing of a photo that I did not previously encounter... maybe I just didn't notice.

I have an image that is RAW format and I am using Adobe Bridge to open it in photoshop. I need to resize it to a smaller image and when I do I lose a great amount of detail in the image. I am editing in 16 bit, sRGB. The purpose of the resize is so I can send it via email.

I can even save the original size to JPG and not lose any detail, and of course if I resize and don't re-sample the image (like I would do for printing) I don't lose any detail.

Does anyone have any tips for reducing this effect. I know that some is normal, but this is beyond normal. Unfortunately I cannot post the images that I am working with, there are some minors in the picture and the parents are very explicit about not posting to the web. That is why I need to email.
 
bogleric said:
I have encountered an issue with photoshop during the resizing of a photo that I did not previously encounter... maybe I just didn't notice.

I can even save the original size to JPG and not lose any detail, and of course if I resize and don't re-sample the image (like I would do for printing) I don't lose any detail.

Does anyone have any tips for reducing this effect. I know that some is normal, but this is beyond normal. Unfortunately I cannot post the images that I am working with, there are some minors in the picture and the parents are very explicit about not posting to the web. That is why I need to email.

well, loss of detail can be caused by several things. one is that you downsampled it too much, or you maybe use too agressive jpeg compression. do you save the final image as jpeg?

very hard to judge what happened if one cannot see the image. maybe you can reproduce the effect with a different image and post it?
 
its hard to say without seeing, but I would say make sure you only save your changes as serperate files, dont save any changes in size to your original raw file.... i dont even know if that is the problem, but it could be........
 
When I downsize in CS from raw for the web I do it in stages, from 100% to 60% in image size, then apply us mask, resize again pixels this time, usmask again, then again, down to about 5x4 inch, this usually gives a sharp image at 72 ppi, good enough for onscreen, less pixels than this tend to look rubbish. SAVE AS SOMETHING ELSE OR YOU'LL LOSE THE ORIGINAL.
 
Resizing for web/e-mail is one of the last steps I do.

I change the image size to less than 800 pixels and then save as JPEG at 5-8 quality. I don't loose any noticeable detail.
 
You have to resharpen the image. Of course it is losing detail, because photoshop is removing the pixels that it deems unnessecary to the image. A good unsharp mask will bring back and appearance of detail, and sharpness.
 
usually when I resize for web, I take my print image and change the setting to 72ppi. The save as, use the original filename and tack -web on the end of it to let me know it is the lower resolution file.
 
$naps! said:
First, never edit in the smaller sRGB colour space. Convert and/or capture your files to the larger adobe 1998. Then convert to sRGB only for web viewing(thats what sRGB is for). If its just going to print Adobe 1998 is what to use. Editing in 16 bit and sRGB kinda defeats the purpose of 16 bit.
That's not completely accurate as I know it. Adobe RGB has a larger color gammut, but since sRGB is a smaller color space, it has higher definition of the colors it does cover. The steps between the colors are smaller. If your end result is going to be in sRGB, you will have better results if you capture in sRGB, because you won't lose color resolution in the conversion*. Some people print using online services that require JPGs. I personally print my own, so I use Adobe RGB all the way and convert to sRGB only for a web JPG.

Also, no matter what color space you use, 16-bit will give you greater color resolution than 8-bit. You won't lose that by capturing in sRGB.


*It's similar to capturing a frame with the focal length/crop you intend to use compared to using a wider angle and then cropping in. Sure, the wider angle captured more info at the start, but when you enlarge to get the crop you wanted at the same size, you end up losing data.
 
It looks like you are confusing color space and color resolution. Color space is the colors that are represented. Color resolution is how finely detailed those color colors are represented. Raw is 12-bit color, which gives you 4096 levels of each color. 8-bit is 256 levels and 16-bit is 65,536 levels. Color space tells you what colors those levels are spread over.

When you convert AdobeRGB to sRGB, you don't usually compress the info; that would cause color shift and is what causes the image to look dull when you do this by accident. If you have a certain red, you want it to be the same red after you convert. The only way to do this is to chop off the extra color space that doesn't fit in the new color space. If you try to keep the extra range, you won't be seeing the correct color after the conversion.

These pages explain it better than I can:
http://www.smugmug.com/help/srgb-versus-adobe-rgb-1998
http://www.cambridgeincolour.com/tutorials/sRGB-AdobeRGB1998.htm
Which echo the link Matt posted.

And true, if you capture in JPG it's probably going to be 8-bit. But with my camera, I can choose the color space I capture in, which means I can get 8-bit in either, or 12-bit raw in either. Since there is no 12-bit editing in Photoshop, this gets converted to 16-bit, which mitigates the color resolution issue when converting (as long as you do it before converting from 16-bit to 8-bit).

So basically, my suggestion is, if you shoot in JPG, use sRGB. If you shoot in RAW, use Adobe RGB.
 

Most reactions

New Topics

Back
Top