pancakes

MicrostockGroup Sponsors


Author Topic: adobe massive batch rejections? (actual photos + gen ai)  (Read 7816 times)

0 Members and 1 Guest are viewing this topic.

« on: December 15, 2023, 10:41 »
0
So...

a) For actual photos (not genAI)... Do you "need" keep the meta data in the images? I submitted a batch of high quality photographs - and "all" were rejected... (I had cleaned out the meta data, i.e., what camera was used, and other details)... was that just laziness on the reviewers part (i.e., did they just "assume" it was genAI because of no meta data, so just rejected it), or what was going on - do I "need" to leave that data in? Extremely frustrating, as I had waited quite some time for them to be processed...

b) When I do some of the genAI,I do take the time to remove extra fingers, logos, make sure the composition is correct, etc... I realize there are probably many that don't (seeking 'genAI' riches with no work/editing/etc)... HOWEVER... it's also frustrating when it seems you get a lazy reviewer - that rejects 90%-95% of a batch that required a lot of time consuming editing to make sure it looked good... Matt, could you please fix that?

Thanks very much!
« Last Edit: December 15, 2023, 10:43 by SuperPhoto »


« Reply #1 on: December 15, 2023, 13:07 »
+3
So...

a) For actual photos (not genAI)... Do you "need" keep the meta data in the images? I submitted a batch of high quality photographs - and "all" were rejected... (I had cleaned out the meta data, i.e., what camera was used, and other details)... was that just laziness on the reviewers part (i.e., did they just "assume" it was genAI because of no meta data, so just rejected it), or what was going on - do I "need" to leave that data in? Extremely frustrating, as I had waited quite some time for them to be processed...

b) When I do some of the genAI,I do take the time to remove extra fingers, logos, make sure the composition is correct, etc... I realize there are probably many that don't (seeking 'genAI' riches with no work/editing/etc)... HOWEVER... it's also frustrating when it seems you get a lazy reviewer - that rejects 90%-95% of a batch that required a lot of time consuming editing to make sure it looked good... Matt, could you please fix that?

Thanks very much!

OOhhhh... prompting for genAI is Time consuming... A pity...  :-[ :-[ :-[ :'(

« Reply #2 on: December 15, 2023, 13:14 »
+1
So...


Do you have an example, one or two rejected images? full size
« Last Edit: December 15, 2023, 13:17 by hellou »

« Reply #3 on: December 15, 2023, 14:08 »
0
So...

a) For actual photos (not genAI)... Do you "need" keep the meta data in the images? I submitted a batch of high quality photographs - and "all" were rejected... (I had cleaned out the meta data, i.e., what camera was used, and other details)... was that just laziness on the reviewers part (i.e., did they just "assume" it was genAI because of no meta data, so just rejected it), or what was going on - do I "need" to leave that data in? Extremely frustrating, as I had waited quite some time for them to be processed...

b) When I do some of the genAI,I do take the time to remove extra fingers, logos, make sure the composition is correct, etc... I realize there are probably many that don't (seeking 'genAI' riches with no work/editing/etc)... HOWEVER... it's also frustrating when it seems you get a lazy reviewer - that rejects 90%-95% of a batch that required a lot of time consuming editing to make sure it looked good... Matt, could you please fix that?

Thanks very much!

OOhhhh... prompting for genAI is Time consuming... A pity...  :-[ :-[ :-[ :'(

Haha, glad you feel the sadness :)

It's not the prompting itself... It's the actual editing to make sure it is a useable image. I realize of course not everyone does that - but, I do take the time to do that which actually makes it quite time consuming... Like brushing stuff out, 'adding' an extra finger where there should be one, etc... Quite time consuming, such that lol - almost seems faster to take regular photos.

BUT - if you read my post - it was ALSO for regular photography that got rejected, that was frustrating. High qualty cameras, good lighting, good subject focus, good subject, unique content, desireable/commercial value, and then just batch 'quality issues' rejections...

« Reply #4 on: December 15, 2023, 14:18 »
0
So...

a) For actual photos (not genAI)... Do you "need" keep the meta data in the images? I submitted a batch of high quality photographs - and "all" were rejected... (I had cleaned out the meta data, i.e., what camera was used, and other details)... was that just laziness on the reviewers part (i.e., did they just "assume" it was genAI because of no meta data, so just rejected it), or what was going on - do I "need" to leave that data in? Extremely frustrating, as I had waited quite some time for them to be processed...

b) When I do some of the genAI,I do take the time to remove extra fingers, logos, make sure the composition is correct, etc... I realize there are probably many that don't (seeking 'genAI' riches with no work/editing/etc)... HOWEVER... it's also frustrating when it seems you get a lazy reviewer - that rejects 90%-95% of a batch that required a lot of time consuming editing to make sure it looked good... Matt, could you please fix that?

Thanks very much!

OOhhhh... prompting for genAI is Time consuming... A pity...  :-[ :-[ :-[ :'(

Haha, glad you feel the sadness :)

It's not the prompting itself... It's the actual editing to make sure it is a useable image. I realize of course not everyone does that - but, I do take the time to do that which actually makes it quite time consuming... Like brushing stuff out, 'adding' an extra finger where there should be one, etc... Quite time consuming, such that lol - almost seems faster to take regular photos.

BUT - if you read my post - it was ALSO for regular photography that got rejected, that was frustrating. High qualty cameras, good lighting, good subject focus, good subject, unique content, desireable/commercial value, and then just batch 'quality issues' rejections...

Maybe they reject all your "regular photographies" because they get tons of AI generated that they judge better quality...
Just think that your AI generated do the same for regular photographies from others...
Be sure Adobe makes no difference for both. They are unable. Brave new world...

« Reply #5 on: December 15, 2023, 14:19 »
+2
mass rejects have been reported for some time now in previous threads

« Reply #6 on: December 15, 2023, 14:39 »
0
mass rejects have been reported for some time now in previous threads
Mass AI submissions also have been reported  ::) coincidence?  ;D

« Reply #7 on: December 15, 2023, 16:56 »
+4
So...

a) For actual photos (not genAI)... Do you "need" keep the meta data in the images? I submitted a batch of high quality photographs - and "all" were rejected... (I had cleaned out the meta data, i.e., what camera was used, and other details)... was that just laziness on the reviewers part (i.e., did they just "assume" it was genAI because of no meta data, so just rejected it), or what was going on - do I "need" to leave that data in? Extremely frustrating, as I had waited quite some time for them to be processed...

b) When I do some of the genAI,I do take the time to remove extra fingers, logos, make sure the composition is correct, etc... I realize there are probably many that don't (seeking 'genAI' riches with no work/editing/etc)... HOWEVER... it's also frustrating when it seems you get a lazy reviewer - that rejects 90%-95% of a batch that required a lot of time consuming editing to make sure it looked good... Matt, could you please fix that?

Thanks very much!

Removing metadata such as camera details would have zero impact on moderation results. The quality of the asset is the top criteria when it comes to image reviews. Without seeing the content you are referencing, I can only assume there were issues with the files and that the moderators got it right. I'm certainly open to being proven wrong and would encourage you to share a couple of examples here.

-Mat Hayward

« Reply #8 on: December 16, 2023, 02:00 »
+2
So...

a) For actual photos (not genAI)... Do you "need" keep the meta data in the images? I submitted a batch of high quality photographs - and "all" were rejected... (I had cleaned out the meta data, i.e., what camera was used, and other details)... was that just laziness on the reviewers part (i.e., did they just "assume" it was genAI because of no meta data, so just rejected it), or what was going on - do I "need" to leave that data in? Extremely frustrating, as I had waited quite some time for them to be processed...

b) When I do some of the genAI,I do take the time to remove extra fingers, logos, make sure the composition is correct, etc... I realize there are probably many that don't (seeking 'genAI' riches with no work/editing/etc)... HOWEVER... it's also frustrating when it seems you get a lazy reviewer - that rejects 90%-95% of a batch that required a lot of time consuming editing to make sure it looked good... Matt, could you please fix that?

Thanks very much!

Removing metadata such as camera details would have zero impact on moderation results. The quality of the asset is the top criteria when it comes to image reviews. Without seeing the content you are referencing, I can only assume there were issues with the files and that the moderators got it right. I'm certainly open to being proven wrong and would encourage you to share a couple of examples here.

-Mat Hayward

I totally agree with you Mat, In my 15 years experience in stock photography I also many times got frustrated about rejections, but looking at those images 5-10 years later I see the reviewers were right in most of the cases.Without examples this topic is useless.
I see overall people thinking on quality only on technical aspect, but to me in one stock photo quality means also what value it brings to the collection and from there the quality of the Adobe Stock collection as a whole. Every image is like a small pixel who create the whole collection image. From this point I think with AI we have to think even more on this level of "hidden" quality and to think what our brain can create as idea and bring it to the image's heart, because a couple of years from now there will be an option in the AI tools to create millions of images in a couple of hours in bulk by group of criteria and prompts templates. I'm very excited to see our new role as stock photographers. Sorry, got a bit off topic.


« Reply #9 on: December 16, 2023, 13:13 »
+5
...
I see overall people thinking on quality only on technical aspect, but to me in one stock photo quality means also what value it brings to the collection and from there the quality of the Adobe Stock collection as a whole. Every image is like a small pixel who create the whole collection image....

perhaps, but we can't tell what the reason is when the rejection just says 'quality' - rejection because it doesn't fit is the old very subjective "low commercial value". While annoying (since no one really kows what will sell), a separate rejection for LCV at least tells us there's nothing we can do.  otherwise its's just a big guessing game

« Reply #10 on: December 16, 2023, 19:00 »
+7
Adobe rejecting by batch and not by individual image has become the norm, it started about 6 months ago. They need to stop using AI to review images, train their reviewers correctly or employ people who can do their job.

« Reply #11 on: December 18, 2023, 04:14 »
+7
Since the past two weeks, everything seems to be getting rejected.

JaenStock

  • Bad images can sell.
« Reply #12 on: December 18, 2023, 06:59 »
+1
Adobe debe limitar la subida masiva de IA.

« Reply #13 on: December 18, 2023, 13:39 »
+2
I can relate to the frustration of having images rejected for "quality" with no hints as to why.

I had 65 AI images accepted with no problem followed by 14 rejected for "quality."

It's from 4 different batches/concepts.

The largest batch includes full-sized (i.e. 4-6MP) carefully edited illustrations, sets, banners and seamless backgrounds I designed from those edited elements, where the components were carefully shrunk down from the much larger elements (after they were already edited to remove defects at the largest size). I've looked at them at 200% and don't see any flaws, other than the type of "imperfection" you'd expect from hand-drawn images - with the lines still all smooth. I used my stylus and paintbrushes when correcting defects and I'm not pretending they are vectors. They've rejected everything - the individual components, the sets, the banners, and the seamless patterns.

I didn't rely on AI to create seamless patterns from the two rejected sets. Instead, I carefully planned each pattern out and worked with the offset filter, checking how each looked at different offsets to be sure each one worked well whether at +50 or +700 e.g. Then, for each one, I made a huge document so I could cut and paste copies of the final pattern to assure that it worked as a seamless pattern in every direction and that there were no stray lines appearing when the edges were combined.

Ironically, the AI images that were part of an even larger batch, where my edits were more minimal, were all accepted quickly and a couple have already sold a few times.

It feels like the less human intervention the better. Kinda depressing since I feel like it's important to add some human artistry - and in earlier batches heavily edited AI images were in fact all accepted.

There really is no way that every single component and group of images each batch had quality issues. I'm reluctant to put any more time or effort into it since I feel like they didn't even look at any of my images, but just hit "reject all."

Another reject was the final image of a themed group of different animals where the other 6 were all accepted weeks earlier.

A handful from the last set I uploaded are still waiting for review days after the others were rejected. It will be interesting to see how they do.

Meanwhile, dt accepted all of them. I'm tempted to make these themed sets exclusive there since exclusive images there tend to do well, despite the site's otherwise lackluster performance. I'd prefer to have them on Adobe, but I already spent way too much time on these AI images and having to spend hours trying to figure out if I missed something after already putting in a lot of work defeats the concept of quickly adding AI to my portfolio.

Requiring an explanation for why the quality is bad would help stop (perhaps lazy?) reviewers from just rejecting an entire batch. The 14 rejected included a cat, plants with leaves, flowers, cacti, drawings of feathers, a dreamcatcher, and abstract patterns. Hard to believe I got them all wrong.

I've given all the images another look and honestly I can't see what's wrong with any of them. Hard to correct when you don't know what you've done wrong.

Time to get back to my camera.
« Last Edit: December 18, 2023, 13:52 by wordplanet »

« Reply #14 on: December 18, 2023, 19:10 »
+2
Yep. Same here. I inspect them after upsizing, remove artifacts if any, etc, etc. And it just seems to be a lazy reviewer - or - just meeting a 'quota' of rejections.

« Reply #15 on: December 18, 2023, 20:21 »
+5
Yep. Same here. I inspect them after upsizing, remove artifacts if any, etc, etc. And it just seems to be a lazy reviewer - or - just meeting a 'quota' of rejections.

Why dont you share some photos like Matt suggested? This thread without actual photos is not beneficial to anyone.

« Reply #16 on: December 19, 2023, 00:07 »
0
Yep. Same here. I inspect them after upsizing, remove artifacts if any, etc, etc. And it just seems to be a lazy reviewer - or - just meeting a 'quota' of rejections.

Why dont you share some photos like Matt suggested? This thread without actual photos is not beneficial to anyone.

Because I don't believe it would result in anything different, based on past threads I've seen. If I felt they'd receive an honest assessment, and get a response of something like 'golly gee! you are correct! We'' get that batch put back up', then I probably would. I spend a LOT of time preparing them, to make sure there is no pixellation, chromatic abberation, good subject(s), good in demand content, unique saleable content/perspectives, etc, etc. I do a LOT of that. even though I know it seems many don't. So to see batches like that get rejected to me just says it seems the reviewer is probably just trying to meet a quota, so he/she can "work" 10 minutes for the day while punching in an "8-hour workday" (I suspect/believe many of those types of jobs are 'outsourced'), then go outside and enjoy the sun. Of course, I wouldn't say that issue is unique to Adobe, but it is frustrating when you get such a reviewer.


« Reply #17 on: December 19, 2023, 03:06 »
+5
So, what is the point of creating this post if not to show the rejected images and find why you got such rejections? It's a frivolous to assume Adobe, the biggest company in the photography area don't know what they do, but you know. Many experienced people here can help you to find some weak points in your job and to resolve this rejection problems.

« Reply #18 on: December 19, 2023, 03:41 »
+7
So, what is the point of creating this post if not to show the rejected images and find why you got such rejections? It's a frivolous to assume Adobe, the biggest company in the photography area don't know what they do, but you know. Many experienced people here can help you to find some weak points in your job and to resolve this rejection problems.

This has been going on and on and on for months now. The point is to complain and maybe to finally get Matt to aknowledge that there is something wrong. Rejections have become CRAZY on Adobe. I used to have a 95% acceptance rate, it went to below 40% from one day to the next and I stopped submitting real photos to Adobe completely, because the reject so much (at the same time they accept almost all my AI images, even though the full size quality doesn't even come close to the quality of my real photos) and there have been multiple threads by contributors reporting the same issue.
 Yet Matt claims "everything was fine and nothing changed" when people keep telling him over and over and over again that this is not the case.

Now, if this were just posts from new contributors who do not understand the quality requirements for submitting this would be one thing, but the complains come from experienced contributors who have been doing this for years and when the acceptance rate changes so drastically for so many people from one day to the next, then it seems right to assume that a bunch of people not suddenly and simultaneously lost their abilities to take good quality photos and the problem is with Adobe instead. But to this day Matt refuses to aknowledge that.

What is the point in showing Matt individual photos? I've seen  the extreme level of nitpicking he goes to to justify rejections (like "The photo shows different kind of plates!"). This is not an individual problem, but a large structural problem on Adobe. Nitpicking single photos will not solve this problem.
« Last Edit: December 19, 2023, 04:12 by Her Ugliness »

« Reply #19 on: December 19, 2023, 04:14 »
+3
To me this topic is useless in the forum if there is no discussion on the particular problem with the problematic examples, it looks like personal problem between contributor and Adobe, but not professional. I can't believe in this, by my experience Adobe do it's job professional. But maybe there is some private case and better contact Mat or Adobe directly, if you don't want to share your images, something I understand.

About the AI quality, the agencies evaluate the quality based on the current state of the technology, just like istock in 2002 for example with the image sensor quality compared to the film photography, this is normal.

« Reply #20 on: December 19, 2023, 05:12 »
0
I cannot comment on the mentioned cases without seeing images, but one big issue is that the ai quality keeps improving.

The majority of content I prompted in the early days, and that I paid a lot of money for, are now unusable, because even after careful upsizing the quality is too low.

I am currently in the process of deleting thousands of older files, just keeping the highlights as a reference. They will be reprompted and developed newly with better ai engines.

At the moment my acceptance rate is still very high. Thanks to very helpful insights from the discord community I now also have a much better understanding why so many of my beloved little illustrations were declined.

As a photographer I simply didn't have a proper understanding what a good quality illustration is supposed to look like at 100%.

If my photo ai are declined, which is rare, then usually when I look at the file again a few weeks later, I can usually see the problem.

If not...then I reprocess from scratch and send it in again and then usually it gets accepted.

I would recommend to post examples in the Adobe discord for review and discussion, I find them very, very helpful.


The Adobe reviewers are not out to get us. If a file has stunning quality it will be approved.


Personally, I now downsize much more than I used to. Often down to 3500*2000. Better to have a small file online, then a larger one declined.

I also often do that for normal photos, I have always had the issue that Adobe is the only agency that often declines my files at full resolution.

So many files have two sizes - normal XXL size for everyone and a small size for Adobe.

Customers can easily upsize, so it does not get in the way of sales.

I also make small mixed batches from different series, usually 2-4 files. Instead of one large batch to submit end of day.

Actually I believe this is also good for file discovery. They will go live at different times, sometimes spread out over several months. So if customers sort their search by "new" they can catch one or two files from a useful series. If they want to see them all, they can do a search in my port.

It also means my port is not clogged up with batches of 100 files from a series. Nobody will just browse a port page by page if all they see is 5 series after 10 pages.

Whatever the problem...I hope the affected contributors can make it work.

Adobe is an amazing agency that can give very reliable and strong financial results.

My biggest regret is ignoring them for 10 years.

ETA:

Obviously...I will be in here whining and complaining when my beautiful perfect batches get declined....


« Reply #21 on: December 19, 2023, 05:40 »
0
ETA2:

I like idea of giving Dreamstime some exclusive ai content. They are a small agency but they have been good to us and I appreciate that we can see our files going live in real time.

I have so many test series, it might be a good idea to give them an exclusive home on Dreamstime.

« Reply #22 on: December 19, 2023, 06:34 »
+2
To me this topic is useless in the forum if there is no discussion on the particular problem with the problematic examples, it looks like personal problem between contributor and Adobe, but not professional.

I can only repeat myself. It has become a structural systematic problem on Adobe, not an individual one.

« Reply #23 on: December 19, 2023, 09:00 »
+1
Adobe doesnt give a crap about our footage and images... if they would give a crap, there would be a contact email at each rejection to get a more useful information why it was rejected.

« Reply #24 on: December 19, 2023, 11:08 »
+3
Adobe doesnt give a crap about our footage and images... if they would give a crap, there would be a contact email at each rejection to get a more useful information why it was rejected.

You get that with the premium agencies - there you can talk to an editor and ask for advice. They will also help you create shooting briefs or might even share costs with you for in demand content production.

Macrostock is a different world.

micro platforms have several hundred thousand producers. How do you think are they supposed to deal with millions of complaints every week???

There are plenty of options to get qualified feedback from other producers and sometimes also reviewers chip in.

Have a look at the Adobe discord channel. The community there is extremely helpful.



 

Related Topics

  Subject / Started by Replies Last post
25 Replies
19842 Views
Last post June 18, 2010, 18:06
by luissantos84
0 Replies
5523 Views
Last post September 01, 2011, 11:18
by Morphart
26 Replies
17865 Views
Last post November 28, 2011, 19:00
by sponner
14 Replies
6560 Views
Last post November 10, 2014, 12:34
by BaldricksTrousers
3 Replies
2971 Views
Last post August 29, 2017, 14:21
by vladimir

Sponsors

Mega Bundle of 5,900+ Professional Lightroom Presets

Microstock Poll Results

Sponsors