Cropping in Raw/Jpeg, and Crop ratios? Destructive?

Discussion in 'Canon EOS' started by mike_swardson, Sep 21, 2011.

  1. Ok everybody I have a question on cropping. I was discussing this with another photographer friend of mine about which is better. To do your cropping in photoshop raw, or once you open it in photoshop. Now the way I like to do it is in photoshop raw simply because if your going to crop you are going to sacrifice resolution. I know exactly what megapixels I will have because it tells me right at the bottom along with the new pixel amount per side. Also I don't always open every photo in the full photoshop. If all I need is a little color correction then I just save right from the photoshop raw window.
    Now her way is she likes to do her cropping after photoshop raw while she is in the normal photoshop. She uses the crop tool and sets it to front image so when she does crop it maintains the same HxW pixel count. Does this sound like an extra step on her part, or is it destructive to crop in raw? As mentioned my take is it doesen't matter if it's in raw, or regular photoshop when you crop because you are still discarding resolution. She mentioned something about it reassigning pixels, but I think it may just be a piece of mind thing for her to see the original resolution even if half the photo has been cropped.
    My other question I had about cropping is if a picture needs a little cropping I try to use a 5x7 ratio. My camera (like most) takes pictures in a 2 to 3 ratio, and as we all know this is a standard 4x6. If I take a picture straight from my camera no crop at all, and order an 8x10 for example then the printer will automatically cut off the sides. I had this happen once when I ordered an 8x10, and didn't change the ratio. My theory is this. 8x10 means one side is 80% of the other, and 4x6 is 66.6% of the other. In a 5x7 ratio 5 is 72% of the 7 so it's right in the middle. I feel as though if I hand off pictures to a client and they just order prints on their own then it will most likely come out ok. Yes you will sacrifice something either in height or width, but just tiny bits of the picture either way you go. Thank you in advance for any input.
  2. I follow your crop approach except that I use Lightroom. In most cases I don't take the images into Photoshop. In Photoshop you can crop with or without changing pixel resolution assuming that is what you mean by reassigning pixels. In either case, the cropped image contains less information regardless of pixel resolution. And I don't leave it up to the printer to crop my images. I crop to the aspect ratio that is appropriate for the paper size. For some images, I create differently cropped versions (saved as tiff files then jpg for sending off to a commercial printer) for different aspect ratios. That's all very straightforward in LR especially if you print directly from LR.
  3. Real life doesn't always fit a 2:3 window. I personally think RAW is way too early to crop.
    It might help to think of cropping as you do any other print-specific tweak, particularly when targeting multiple or unknown print sizes as you are. For almost every picture, you'll have a hard time getting any two photographers to agree on the best crop. I often fiddle with it back and forth, as I might any other setting. I avoid having to cut custom mattes, but even that isn't a hard rule.
  4. Mike,
    It depends on how you have ACR configured to an extent. Say you have a 6000 x 4000 pixel image, if you crop that in ACR it throws nothing away, it just writes instructions in an xmp file to display that image cropped to 3000 x 2000, if you throw away, or lose the xmp file you get the entire picture back. ACR does no destructive processing at all to your original RAW file.
    But when you save ACR might be configured to generate a jpeg in addition to writing the xmp file for the untouched RAW file, if it is then that jpeg will only have 3000 x 2000 pixels. Your RAW file will still be full resolution though, ACR cannot change that.
    Your friend is resampling the image to get PS to generate pixels, Photoshop does destroy, change and throw away pixels. There is no magic to it, if she crops 50% she has 50% less pixels, if you then just resample that remaining 50% to get you back to the same number you had Photoshop is just making them up. Unless you need the pixel number for some good reason there is no advantage to doing it her way. Both methods are reducing the file size, she is then just artificially increasing it again.
    I wouldn't do what you are suggesting for your printing files. Crop the image to its optimum shape, or maintain a set ratio for all delivered images in a job. That way the customer can at least have a chance of getting an off the shelf frame to fit correctly.
  5. I virtually never crop in ACR (Adobe Camera Raw, which is what I presume you mean by "Photoshop raw") but instead wait to crop in Photoshop. For one thing, I do not always know for certain exactly how I'll crop as early as the raw conversion stage. If you do crop in ACR, it would be an extremely good idea to bring the image into Photoshop as a smart layer - that way you can return to ACR by doubly clicking the layer in PS if you later need to change the crop or other elements.
    Another reason to not crop so early in the process is that your ideal crop may not be the only crop you'll need. You may have a perfect 3:2 crop in mind, but what happens when client wants a 4:3 ratio?
    And there is no reason at all for your friend to try to keep the pixel dimensions of the image the same when cropping. In fact, the interpolation process that must occur to make this happen can do not good and may do harm. If the pixel dimensions end up smaller when you crop, just leave them that way and don't do any interpolation until you print.
    As to what crop ratio is "best," there are a variety of factors to consider. I do not subscribe to the notion that one ratio is good and others less good - it is largely an aesthetic decision for the photographer... or a practical one when the client asks for a specific size. If you will license your photos for use in various media, there is no single ratio that you can expect to be asked to provide. If you like the 5x7 or 4x6 or whatever, go with it - but realize that some of your images may work better with other ratios.
  6. Thanks for the input everybody, and it sounds like I am not crazy. As I said before if you crop you are discarding data. The way I have my workflow set up is I like to do my cropping in camera raw just because I don't pull everything in photoshop, and may save it from there to jpeg for the client. I have however been looking at a photo, and thought man I wish I had cropped that a little less. In that case I do just go back to the raw, and bam its all right there. I keep the raw files just for that reason, but only save to jpeg for the client. In regards to the print ratio I am with all of you. I for sure do not let a printer decide my fate for the ratio. If a client orders an 8x10, then I can take a second to open it up, and configure it really quick to make sure it will be the right ratio. Perfect for like I just mentioned having the raw to open up, and do what I need to really quick. My point is if I sell the images on a DVD to the client, and say they decide 5 years from now that they want to print a picture I had on there, and I don't have the option to configure it for them. whichever ratio they decide it will probably be ok. Guaranteed to trim a bit of the image if they are not in 5x7 ratio, but tiny trims either way. I know whenI ordered an 8x10 and it was in 2x3 format it cut off A LOT. Just to drastic is all, and I just want to minimize that. Now I have had to configure plenty of pictures for albums too. Like if I do a full page photo, and part of the photo goes into the crease, I will add canvas to that side by about a 1/4-1/2" to make up for that. This way half the picture isn't in the crease, and you don't even notice because it's just a little fine tuning specific to print. Anyways thank you everybody.
  7. In general, depending on how you save the file, many changes made to the RAW image are saved in auxiliary files (e.g., xmp), not in the original.
    However, you can be absolutely sure you're not destroying any data if you just open the file and then SAVE AS in a lossy (jpg) or not-lossy format (TIFF, etc.). Otherwise I'm always afraid to run the risk that you did in film days when you retouched the negative rather than the print, so to speak. SAVE AS is so easy and you absolutely know that you still have the original image to go back to.
  8. I usually crop last. before sharpening. That means, I crop in PS.
    1. What if a photo needs a retouch? If it is cropped first, in ACR or PS, it will need to be developed again. That means, you have to work on the post again.
    2. What if you have to crop differently for an application, such as an exhibition for example? That means you will have to work on the post again!
    3. Do you guarranty that every crop will be exactly the same, so that you won't have to work on the post again?
  9. "I wish I had cropped that a little less. In that case I do just go back to the raw, and bam its all right there."
    It is also all still there if you do the crop in Photoshop. I don't have my copy of PS on the screen right now, but I believe you'll find a "reveal all" or similar command that will remove all previous PS cropping and show the entire original frame.
    I know it is a matter of preference, but with this capability in PS, I can't think of a really good reason to do the crop in ACR if you will need to modify the crop later on.
  10. The methodology that Dan described above works if an image is cropped in PS. But if Genuine Fractals is used, PS probably won't keep the uncrop info, as it does in the former case.
    Regarding to methodology, I normally develop a RAW and save it to a file. This is my master copy. I then flatten, crop, and sharpen the master copy for a print, which will be saved as a 2nd file. This 2nd size tends to be larger than the web's size. Either from the master file or from the 2nd file, I then down-size the image for the web. For the web, I don't care (much) to use an already-sharpened image. I will have to sharpen it again though, because sharpening goes with size.
    Having a full-frame master copy is convenient. You can set the working color space there, and use it to make variations for different color gamut in different devices.
  11. I use Lightroom. I can crop and straighten anywhere in the process and it doesn't make a difference. Cropping might be my first operation after importing or the last thing that I do before exporting a final copy. I can crop and re-crop multiple times until I'm satisfied. it makes absolutely no difference to Lightroom, it's completely nondestructive, and I don't have to save a big TIFF file to keep a copy of the image before cropping. Life is good.
  12. I don't have to save a big TIFF file to keep a copy of the image before cropping.​
    A PS file doesn't have to be saved as a TIFF file. And it is really stupid to do so. A PS file can always be saved as a PSD file. This way, the file size is much smaller. Espcecially if a file that has many layers.
    LR is good, much like DPP of Canon, but they both don't have the masking capability, yet. To do masking , users have to jump to PS. I may be wrong on this though. LR may have the masking capability in the latest version.
    Without the masking capability, local adjustments cannot be achieved. The image thus cannot be optimized.
    With local adjustments, life is good, as always.

Share This Page