CA: Multiple Exports?

jshrager

I'm confused. CA is ... well ... A! But there are some very confusing things about it. (First confusion, is there a separate CA forum? Well, anyway...) So, I have been exporting to my iPhone Camera Roll (hereafter iRoll), and I've probably accidentally done it multiple times, which, of course, gives me multiple copies of the same picture. That's only slightly confusing. What's REALLY confusing is that the two copies sometimes appear to be DIFFERENT SIZES!? I try to remove accidental duplicates by comparing the time stamps and sizes, and, of course, that they look alike. The time stamps (and looks) are identical, but the sizes of the shots that I've multiple-exported to iRoll from CA, that I *think* (from ts and look) are identical have DIFFERENT SIZES?! Uh....is that possible, or am I, as above, confused?

rainforest1155

You can export or share photos as often as you'd like. Once you've exported / shared a photo, it should display a small icon to what service you shared the photo to.
There's no limitation on how often you can export a photo as you can apply different effects, crops or other edits to your various edits.

I've just tried to export a photo multiple times and the 2 resulting photos in the camera roll were exactly the same size both times.

Perhaps you've used the burst shooting mode and that resulted in files that while they look visually identical, are in fact slightly different. Or you've applied some effect that's not really noticable on that particular photo.

Can you give it a try again with a photo and export the same photo twice to see what files you get?

How do you jugde the file size? Can you share your whole workflow on what do you do from the poin that you export the files?

jshrager

More detail
They really are identical, and I managed to replicate it as follows (see replication attached): 1. Take a picture with CA. 2. Export it to the iRoll. 3. Re-import to CA. 4. Re-export to iRoll. Note that the sizes keep changing -- probably due to compression differences. Also, note that it losses the fstop information, which is a little curious. I'm only slightly scared about the compression differences....could be a problem. I can't upload the raw files for investigation due to upload limits.

mbonocore

Hi JS,

I tried to recreate, and did lose some size on the original compression, but none after that. Also, my fstop information stayed intact. Can I ask what you need to do so many imports and exports of the original file for? Maybe we can think of an alternative that won't have this effect.

Michael

jshrager

I don't. It happens via what my son calls "accidental on purpose" :-) That is, it's hard to keep track of the relationship between all these rolls, so I (and probably everyone else in the world) ends up with duplicates. It would actually be fine if they really were exact duplicates, because I (and probably everyone else in the world) have programs that try to guess which pics are dups, and cull them. (Actually, if no one else has this, I do -- maybe I should make it an app because it's EXTREMELY useful! :-) Anyway, the way the duplicate analysis algorithm works is to look at the file size and meta-data. If the meta-date and file size are the same, it declares it a duplicate. In fact, the file size is much more important than the meta-data bcs it's so unlikely that (as you pointed out) even two pics of the same scene taken back-to-back (e.g., rapid shooting) would have the exact same file size. And, to be honest, I'm unclear how it could be possible for these ex/in/ex pics, which really ARE the exact same pic, to have such different file sizes, so ... if an algorithm could be worried, mine would be ... but at least the algorithm's author is very worried! I wonder, is there some way to tell whether they are the same uncompressed? Is there an app for that?

jshrager

Update: Pix ARE (subtly) different!
The pics ARE different! I expanded them to single pixels and found identical areas; see attached. It's subtle, but they ARE different. Okay, so now I'm really confused. Probably the best thing to do is to use the largest -- presumably this losses the least information.

jshrager

comparison from imagemagick
Attached is out of imagemagick's compare command.

rainforest1155

I guess each time you export the photo to the camera roll, it needs to be saved as jpg again. Jpg is a lossy compression and repeat opening and saving a jpg file leads to a quality loss over time. That would also fit with the file size getting smaller. Therefore if possible, it's best to go back to the original file when doing edits for the best possible quality. Of course, this may not be possible, especially if you do edits across different apps on your device. But then you also have to keep in mind that those quality differences are mostly not visible unless you zoom in a lot so they are mostly likely negligible for the intended purposes of the phone camera shots.

Note that I have no insights in iOS and how access (read / write) from and to the camera roll is handled and I'm just going by what I know about jpg in general.

jshrager

Actually, I've solved one major confusion here, although it created a new one. I was wondering why I was getting multiple copies bcs I was pretty sure I'd only exported ONCE. Although that's true, I have my CA set to backup to the iRoll, so whenever I take a picture with CA, I get a copy in my iRoll. Okay, so that explains why I initially had 2 copies (i.e., not realizing that it did this -- I simply forgot I had set that -- I subsequently explicitly exported another). Okay, but, even more weirdly than any of the above, the auto-export (backup to iRoll) copy is significantly larger than the explicit export, even though they are, presumably, both jpeg exports to iRoll. ?!?!?!?

rainforest1155

I do see the same. I'll check with the team if this is intended behavior.

BradfordBenn

I see the same thing as well, and was about to report it as well. (Finally cleaning up the iPhone for the new year....)