Barbara, the title is a little misleading. What these tools do is prevent *additional* loss after the initial lossy JPEG compression had occurred (by not re-compressing the file). While there are some forms of true lossless JPEG, they are not in wide use.
Guy
Posted by Jim Bullard on September 4, 2011 at 11:07 AM
There is a JPG 2000 format that is less lossy than traditional JPG but for whatever reason, it never caught on. The way JPG makes smaller files is by throwing out some information and then reconstructing it by reversing the formula when the file is reopened. For example (simplified) it compares the difference of tone & hue across a range of pixels, drops the actual data that describes each pixel in the middle of the range and instead inserts data for recreating a similar pattern when the file is reopened. The higher the compression (to get smaller files) the broader the range of difference. The reconstruction is not precisely what the original pixels were but an approximation. The higher the compression, the less true to the original.
Any time you reprocess a JPG you are already working with a file that has had some approximated reconstruction when you reopened it. Anything you do to the file to edit it (editing, by definition, is removing data and/or replacing it with data that was not in the original) "loses" original data. You cannot crop an image and still have all the original data in any format any more than you can trim an 8x10 print to 5x7 and still have an 8x10. If you crop telling the software to maintain the original pixel dimensions the software has to "make up" pixel data to fill in the gaps. The same is true of other edits which alter parts of the original image unless you save it in layers with the original as an unaltered background layer (cropping alters even the background of a layered image) and JPG doesn't do layers.
The JPG problem is that when you or your camera saves the image in JPG you have lost some of the original information in the original image that the sensor recorded. When you open the JPG file the software "reconstructs" the missing information according to an algorithm. If you change that reconstructed information, save it and open it again, you are reconstructing a reconstruction. Keep doing that and you get the same effect as making photocopies of photocopies of photocopies, the image starts to break up. Your best bet is to only use JPG as a format for final copies of images after you are done editing and never, ever directly edit your original files regardless of what file format you use, only copies. That way you can always go back and start over.
Guy and Jim are correct. What you need to do is create a workflow that maintains the original capture file in an unaltered state, and also prevents compounding jpeg compression losses. Here's an example of what is called 'targeted' workflow for a Web image...
1. Create a *.PSD or *.TIF file from your original camera file. This file will be your 'Master' file.
2. Edit your master file to taste, and save it.
3. Duplicate your master file and save under new name, e.g. myfile0-web.psd
4. Size the file, convert] the profile to sRGB, sharpen, and save
(If you have any selections saved as channels, be sure to delete them, as they increase file size)
5. Now, save the file as a jpeg.
Let's say you want to create another jpeg that's a little smaller than [i]myfile0-web.psd Here's what to do.
1. Open your 'master' file and duplicate it.
2. Save this file as myfile0small-web.psd
3. Resize the file to suit, convert the profile, and sharpen as before.
4. Save this new file as a jpeg.
Notice that in the above steps we have not touched the original camera file, or the 'master' file. Also, we have not compounded losses due to repeated compression of your first jpeg image.
I use this workflow every day for both Web jpegs and *.PSD files for printing, always starting from my 'master' file. It actually takes longer to write the steps than to do them. The bottom line is you should develop a set of steps that maintains the camera file and your 'master' file in a pristine state. Oh, yeah--backup, backup, backup!
--P
Posted by Michael Elenko on September 5, 2011 at 2:50 PM
Barbara,
If you are interested in lossless processing of images including camera-generated jpegs, I would recommend using Lightroom. If you already have Photoshop, then the included Camera Raw will suffice.
ME
Posted by David Bostedo on September 6, 2011 at 8:54 AM
Barbara - You may want to take a look at this thread on saving JPGs multiple times. Basically, if you're saving at the highest quality level, you can open, edit, and re-save a JPG several times and you'll likely never notice the difference. If you start to slide the quality level down though, things can change pretty quickly.
Discussion
5 replieshttp://softwaretopic.informer.com/lossless-crop-photoshop/
Guy
Any time you reprocess a JPG you are already working with a file that has had some approximated reconstruction when you reopened it. Anything you do to the file to edit it (editing, by definition, is removing data and/or replacing it with data that was not in the original) "loses" original data. You cannot crop an image and still have all the original data in any format any more than you can trim an 8x10 print to 5x7 and still have an 8x10. If you crop telling the software to maintain the original pixel dimensions the software has to "make up" pixel data to fill in the gaps. The same is true of other edits which alter parts of the original image unless you save it in layers with the original as an unaltered background layer (cropping alters even the background of a layered image) and JPG doesn't do layers.
The JPG problem is that when you or your camera saves the image in JPG you have lost some of the original information in the original image that the sensor recorded. When you open the JPG file the software "reconstructs" the missing information according to an algorithm. If you change that reconstructed information, save it and open it again, you are reconstructing a reconstruction. Keep doing that and you get the same effect as making photocopies of photocopies of photocopies, the image starts to break up. Your best bet is to only use JPG as a format for final copies of images after you are done editing and never, ever directly edit your original files regardless of what file format you use, only copies. That way you can always go back and start over.
Guy and Jim are correct. What you need to do is create a workflow that maintains the original capture file in an unaltered state, and also prevents compounding jpeg compression losses. Here's an example of what is called 'targeted' workflow for a Web image...
1. Create a *.PSD or *.TIF file from your original camera file. This file will be your 'Master' file.
2. Edit your master file to taste, and save it.
3. Duplicate your master file and save under new name, e.g. myfile0-web.psd
4. Size the file, convert] the profile to sRGB, sharpen, and save
(If you have any selections saved as channels, be sure to delete them, as they increase file size)
5. Now, save the file as a jpeg.
Let's say you want to create another jpeg that's a little smaller than [i]myfile0-web.psd Here's what to do.
1. Open your 'master' file and duplicate it.
2. Save this file as myfile0small-web.psd
3. Resize the file to suit, convert the profile, and sharpen as before.
4. Save this new file as a jpeg.
Notice that in the above steps we have not touched the original camera file, or the 'master' file. Also, we have not compounded losses due to repeated compression of your first jpeg image.
I use this workflow every day for both Web jpegs and *.PSD files for printing, always starting from my 'master' file. It actually takes longer to write the steps than to do them. The bottom line is you should develop a set of steps that maintains the camera file and your 'master' file in a pristine state. Oh, yeah--backup, backup, backup!
--P
If you are interested in lossless processing of images including camera-generated jpegs, I would recommend using Lightroom. If you already have Photoshop, then the included Camera Raw will suffice.
ME