MicrostockGroup Sponsors


Author Topic: Finally starting with AI Generated images, but have a few process questions?  (Read 2391 times)

0 Members and 1 Guest are viewing this topic.

« Reply #25 on: May 29, 2024, 04:52 »
+1
I have upload more than 100000 AI files to adobe stock. And I have developed some tools to help me to get it.
Such as
keywording automatically.
upscaling automatically.
remove the background automatically.
Upload automatically.

SO if you have some questions, You can ask me.

Just one question: How many $$$ per month do you make with that strategy?


« Reply #26 on: May 29, 2024, 06:13 »
0
I have upload more than 100000 AI files to adobe stock. And I have developed some tools to help me to get it.
Such as
keywording automatically.
upscaling automatically.
remove the background automatically.
Upload automatically.

SO if you have some questions, You can ask me.

Just one question: How many $$$ per month do you make with that strategy?

If you have 100 AI files on adobe stock, it is 1000 times.

« Reply #27 on: May 29, 2024, 06:23 »
+2
I remove background by some technical process.
I have upload more than 100000 AI files to adobe stock. And I have developed some tools to help me to get it.
Such as
keywording automatically.
upscaling automatically.
remove the background automatically.
Upload automatically.

SO if you have some questions, You can ask me.

Ok, I'll bite.

How do you remove the background automatically to produce good pngs?

I always have to refine every single file, so I find producing pngs quite tedious.

Do you have a good acceptance rate (over 90%)?

Because uploading many files just to have them declined is not very efficient.

Are you having good sales with your port?

50 dollars a week, 500 a week, 5000??? - just a sample range would be interesting for many I am sure.

100k files is what many people dream to have over a lifetime.

You are already there after 18 months.



If the PNG is not good, I will delete it.
If the fingers are not good, I will delete it.
If the face is ugly, I will delete it.
If the .... , JUST DELETE IT.
The stock image markets are not seeking refined single files. They are seeking huge number files with it-is-ok-standard.
So I am an images-producer now, not a photographer or artist.

I have a 70%-95% acceptance rate for different batches. Images with people, acceptance is 70%.
I think it depends on whether the reviewer is in a good mood.

« Last Edit: May 29, 2024, 06:26 by whatwolf »

« Reply #28 on: May 29, 2024, 07:18 »
0
Thank you for responding.

We obviously have different strategies, but maybe my workstyle is for the old days.

Don't know if it is still true today, but some agencies have an "efficiency" factor. I.e. if you have a small port with a very good sales rate, your port gets promoted much better by the algos than a port with a very high fluff factor, or just dead files that don't sell.

So in addition to wanting to have a high acceptance rate, I hope my port does not have too much "fluff".

That is a lot easier to in themes that I understand well and much more difficult in genres where I am new, especially illustrations.

But since you are uploading such a very high volume and probably also very diverse content, not just abstract background waves...you can obviously make it work for you.

I keep trying to upload a higher volume, but I really, really struggle with that.

I have thousands of files that are ok, but not stellar, but need little processing.

Will try to upload them in a larger batch at some point, but perhaps not to Adobe, to keep Adobe as my prime port for ai.

Best of your luck with your journey, you already have more files than I will ever have on Adobe in my entire life.

« Reply #29 on: May 29, 2024, 07:35 »
+1
My educated guess would be probably he 'made' between $10-$15k usd with that strategy (over 1-1.5 years, or since whenever he started doing the AI stuff)... Now, whether that is 'worth' it, I dunno... There's people with 1/20th the size in portfolio (or even smaller) who do exactly the same revenue, or better.

It depends on:
a) Is it in demand/useful content?
b) Does it get discovered (i.e., 'unique', good promotion/etc)
c) Then some images just tend to get really lucky and go on wild selling streaks, doesn't seem to be any reason why, it just happens (i.e., say 1k+ sales for a single image)...

So if he is just spamming the crap out of the images, hoping something sticks - while that is certainly 'one' strategy... not sure if it is the most effective strategy...
Now of course, if he has really top notch images, that is a different thing... but not sure whether that is the case from the sounds of it...

« Reply #30 on: May 29, 2024, 07:57 »
+4
I have upload more than 100000 AI files to adobe stock. And I have developed some tools to help me to get it.
Such as
keywording automatically.
upscaling automatically.
remove the background automatically.
Upload automatically.
Submit automatically. (It is not work yet.) :-[

SO if you have some questions, You can ask me.

Yes, there is a question.
Why do you only speak about quantity?

Success in microstock is a good mix between quality and quantity, and this is a well known rule for any contributor.
Speaking only about quantity is not good way. 100K images... images of what? you have to find what to create. Or you could produce 100k similar files, that is a nonsense.

A good and quick workflow is for sure a good idea, but in terms of revenue it has 50% of weight.
So why do you speak about quantity only?

As someone else already told, there are a lot of contributors that do same or better results with 1/20, or maybe 1/50 of that numbers

« Reply #31 on: May 29, 2024, 09:12 »
+1
I have upload more than 100000 AI files to adobe stock. And I have developed some tools to help me to get it.
Such as
keywording automatically.
upscaling automatically.
remove the background automatically.
Upload automatically.
Submit automatically. (It is not work yet.) :-[

SO if you have some questions, You can ask me.

Yes, there is a question.
Why do you only speak about quantity?

Success in microstock is a good mix between quality and quantity, and this is a well known rule for any contributor.
Speaking only about quantity is not good way. 100K images... images of what? you have to find what to create. Or you could produce 100k similar files, that is a nonsense.

A good and quick workflow is for sure a good idea, but in terms of revenue it has 50% of weight.
So why do you speak about quantity only?

As someone else already told, there are a lot of contributors that do same or better results with 1/20, or maybe 1/50 of that numbers

The first of all, everyone who use Midjourney, could get images with THE SAME QUALITY.
WHY I produce 100k simlilar files?
I produce 4 images in one prompt. I dont use any repeated prompts unless the images are really good at selling.

When you have more than 100000 or 200000 or more ai images, WHAT TO CREATE is meaningless. I CREATE EVERY THING.

« Reply #32 on: May 29, 2024, 09:41 »
+1
Thank you for responding.
We obviously have different strategies, but maybe my workstyle is for the old days.
Don't know if it is still true today, but some agencies have an "efficiency" factor. I.e. if you have a small port with a very good sales rate, your port gets promoted much better by the algos than a port with a very high fluff factor, or just dead files that don't sell.
So in addition to wanting to have a high acceptance rate, I hope my port does not have too much "fluff".
That is a lot easier to in themes that I understand well and much more difficult in genres where I am new, especially illustrations.
But since you are uploading such a very high volume and probably also very diverse content, not just abstract background waves...you can obviously make it work for you.
I keep trying to upload a higher volume, but I really, really struggle with that.
I have thousands of files that are ok, but not stellar, but need little processing.
Will try to upload them in a larger batch at some point, but perhaps not to Adobe, to keep Adobe as my prime port for ai.
Best of your luck with your journey, you already have more files than I will ever have on Adobe in my entire life.

Think about the majority of customers of ADOBE, shutterstock and freepik. Do they care you have done the little processing or not? Most of them CAN NOT TELL what they want, until they see it on the pages. Most of them have no aesthetic ability.

AND, AI is killing stock image businesses.

As far as I know, some news websites produce AI images and put them on stockimage websites to sell. THESE NEWS WEBSITE were customers of gettyimages.

The Customers become Contributors!

Therefore, what I can do is to make more money by producing a large number of pictures before the business completely disappears and before customers completely abandon the stock images industry. Maybe you will say that there will be high-end customers who choose high-end contributors, but why don't high-end customers pay for commercial customization services? Who of us can say that we are high-end enough to avoid being eliminated?

« Reply #33 on: May 29, 2024, 10:28 »
+1

The first of all, everyone who use Midjourney, could get images with THE SAME QUALITY.
WHY I produce 100k simlilar files?
I produce 4 images in one prompt. I dont use any repeated prompts unless the images are really good at selling.

When you have more than 100000 or 200000 or more ai images, WHAT TO CREATE is meaningless. I CREATE EVERY THING.

Well, yes, not only you but everyone can create anything, this is no more a breaking news, it's more than one year that Midjourney is on its way.
And in one year we saw boith traditional photography and AI images that sells well.
So I'm not agree with you, what you're talking aboout is quite overtaked... There are actually billions of AI images in Adobestock and, you guess, traditional photography is still selling.
Sorry but I'm not agree with you, you seems to be a little outdated. Actually the best sellers are quality images not quantity, as quantity is something that anyone can have access to.
My opinion, of course

« Reply #34 on: May 29, 2024, 11:05 »
+2

The first of all, everyone who use Midjourney, could get images with THE SAME QUALITY.
WHY I produce 100k simlilar files?
I produce 4 images in one prompt. I dont use any repeated prompts unless the images are really good at selling.

When you have more than 100000 or 200000 or more ai images, WHAT TO CREATE is meaningless. I CREATE EVERY THING.

Well, yes, not only you but everyone can create anything, this is no more a breaking news, it's more than one year that Midjourney is on its way.
And in one year we saw boith traditional photography and AI images that sells well.
So I'm not agree with you, what you're talking aboout is quite overtaked... There are actually billions of AI images in Adobestock and, you guess, traditional photography is still selling.
Sorry but I'm not agree with you, you seems to be a little outdated. Actually the best sellers are quality images not quantity, as quantity is something that anyone can have access to.
My opinion, of course

One BIG thing is... people should be focusing on getting midjourney to pay authors RETROACTIVELY for the STOLEN works they made their tool based on, as well as set up a system for RECURRING PERPETUAL works going forward. discuss that. that is effective.

« Reply #35 on: May 29, 2024, 12:09 »
0


Therefore, what I can do is to make more money by producing a large number of pictures before the business completely disappears and before customers completely abandon the stock images industry. Maybe you will say that there will be high-end customers who choose high-end contributors, but why don't high-end customers pay for commercial customization services? Who of us can say that we are high-end enough to avoid being eliminated?

At this point I am not worried about being eliminated. On the contrary, I feel set free to experiement with any genre I am interested in.

In my case especially to experiment with illustrations, without having to learn illustrator, oil paintings or ink drawings.

You keep saying that customers don't care about quality, but 20 years experience with photos shows me that customers only really buy the absolute best files. And these then outsell all the mediocre ones by the thousands.

While customers often only have a vague idea what are looking for, if you browse thousands of files on an agency, you can quickly detect quality content.

Otherwise customers would source all their images from the free sites or use the creative common license on flickr.

There are already billions of files you can use commercially for free.

Customers pay agencies to save time.

And so far working with ai costs a lot more time, than just browsing millions of files on Adobe.

I sincerly wish you all the best. Please kepp us updated how it goes.

You will probably have a million files in 5 years.


« Reply #36 on: May 30, 2024, 05:21 »
+1

The first of all, everyone who use Midjourney, could get images with THE SAME QUALITY.
WHY I produce 100k simlilar files?
I produce 4 images in one prompt. I dont use any repeated prompts unless the images are really good at selling.

When you have more than 100000 or 200000 or more ai images, WHAT TO CREATE is meaningless. I CREATE EVERY THING.

Well, yes, not only you but everyone can create anything, this is no more a breaking news, it's more than one year that Midjourney is on its way.
And in one year we saw boith traditional photography and AI images that sells well.
So I'm not agree with you, what you're talking aboout is quite overtaked... There are actually billions of AI images in Adobestock and, you guess, traditional photography is still selling.
Sorry but I'm not agree with you, you seems to be a little outdated. Actually the best sellers are quality images not quantity, as quantity is something that anyone can have access to.
My opinion, of course

One BIG thing is... people should be focusing on getting midjourney to pay authors RETROACTIVELY for the STOLEN works they made their tool based on, as well as set up a system for RECURRING PERPETUAL works going forward. discuss that. that is effective.


I agree with you but unfortunately this will never happen.

regarding AI content,quantity is certainly the winning strategy.

better to generate many more than you need and then carefully select the best ones,rather than waste too much time in postproduction.

I have recently developed a new system in postproduction,which allows me to speed up my work considerably,whilst maintaining an excellent standard of quality.

It took me several tries and several steps before I achieved this strategy,but it works,my approval rate is very high,and the content is of excellent quality,but above all I can go much faster now.  :)

« Reply #37 on: May 30, 2024, 08:36 »
+1

The first of all, everyone who use Midjourney, could get images with THE SAME QUALITY.
WHY I produce 100k simlilar files?
I produce 4 images in one prompt. I dont use any repeated prompts unless the images are really good at selling.

When you have more than 100000 or 200000 or more ai images, WHAT TO CREATE is meaningless. I CREATE EVERY THING.

Well, yes, not only you but everyone can create anything, this is no more a breaking news, it's more than one year that Midjourney is on its way.
And in one year we saw boith traditional photography and AI images that sells well.
So I'm not agree with you, what you're talking aboout is quite overtaked... There are actually billions of AI images in Adobestock and, you guess, traditional photography is still selling.
Sorry but I'm not agree with you, you seems to be a little outdated. Actually the best sellers are quality images not quantity, as quantity is something that anyone can have access to.
My opinion, of course

One BIG thing is... people should be focusing on getting midjourney to pay authors RETROACTIVELY for the STOLEN works they made their tool based on, as well as set up a system for RECURRING PERPETUAL works going forward. discuss that. that is effective.


I agree with you but unfortunately this will never happen.

regarding AI content,quantity is certainly the winning strategy.

better to generate many more than you need and then carefully select the best ones,rather than waste too much time in postproduction.

I have recently developed a new system in postproduction,which allows me to speed up my work considerably,whilst maintaining an excellent standard of quality.

It took me several tries and several steps before I achieved this strategy,but it works,my approval rate is very high,and the content is of excellent quality,but above all I can go much faster now.  :)

Why do you believe that, and believe you could speak on behalf of an entire industry?
It would only 'never happen' if no one did anything and just whined and complained. Of course it can happen and it will happen, providing people first become aware of it as a possibility (in process), and then people taking coordinated action. Then it will happen. And authors will be compensated retroactively for their stolen works, as well as receive perpetual revenue in the future should they choose to license those works going forward.
« Last Edit: May 30, 2024, 08:38 by SuperPhoto »

Uncle Pete

  • Great Place by a Great Lake - My Home Port
« Reply #38 on: May 30, 2024, 13:20 »
+2

One BIG thing is... people should be focusing on getting midjourney to pay authors RETROACTIVELY for the STOLEN works they made their tool based on, as well as set up a system for RECURRING PERPETUAL works going forward. discuss that. that is effective.

I am not a fan of AI, but your request is flawed.

When you read a dictionary and learn what a word means, do you pay a royalty, every time you speak or write that word? When you look on the web for an answer and learn how to bake a cake, do you pay for that knowledge, every time you bake a cake?

When a machine learning program, looks at an image and learns what a cat looks like, should every image the software looked at to learn, get a commission, every time the AI creates a cat image? The software is not "using" the original image, it isn't referencing the original image, it has learned elements of what a cat looks like.

Second flaw, the companies don't know where they got the images or who made them. There's no source association, no one to one way to identify, which image or images were used to train or which images were used to create the output.

When rain falls from the sky, do you know where that water came from, originally through evaporation?

If 100 million images were used to train the AI, do you know which tiny bits were used to create each part of the specific image?

The only agencies that seem to be blocking AI and not allowing people to upload, are the ones that have their own pay system that produces the same kind of images. They just say "no AI" because they want to block our competition.


 

Related Topics

  Subject / Started by Replies Last post
60 Replies
19804 Views
Last post January 24, 2009, 17:16
by bittersweet
19 Replies
8618 Views
Last post September 22, 2009, 20:08
by click_click
12 Replies
10667 Views
Last post January 06, 2014, 16:27
by Uncle Pete
4 Replies
4248 Views
Last post February 13, 2015, 02:29
by Niakris
9 Replies
2105 Views
Last post November 29, 2023, 10:56
by Evaristo tenscadisto

Sponsors

Mega Bundle of 5,900+ Professional Lightroom Presets

Microstock Poll Results

Sponsors