Quote from: wds on May 15, 2024, 11:25

Only because in other previous posts (by other stills photogs), there seemed to be a trend where US based Shutterstock contributors seemed to be complaining about dropping SS sales and contributors outside the US seemed to be happier with their SS results.

Just to be clear, I only sell video.  Stills is a waste of time.
Nevertheless, I am also experiencing fewer sales and lower revenue at Shutterstock.  On the other hand, Adobe is slowly rising, but still not as good as Shutterstock by total dollar volume or per-clip revenue.

Interesting. I primarily sell stills....could be that at SS video (vs. other agencies) is holding it's own and it is on the stills side where SS seems to be doing seriously worse than AS.

So far this month, I've had almost exactly the same number of downloads at Shutterstock as at Adobe.
The average sale at Adobe has been $11.60
The average sale at Shutterstock has been $16.47

That's about 40% more income for the same number of sales. Shutterstock isn't what it used to be, but it is still king.

Just curious, are you in the USA?

Slow week so far, but weekly ranking is above average.  Must be a slow week for everybody.  It's possible more buyers started generating their own AI images via Midjourney etc instead of buying our images.  That's my concern about near future.

A valid concern, but I don't think you would see a sudden drop at this point because of AI competition. I would guess it would be more gradual.

It would be interesting to see some kind of cost comparison analysis. It's got to take up quite a bit more time for an employee to generate their own AI images. If they use AI images generated by others on one of the AI sites, they are still paying a subscription cost to use the AI site...and are their usage charges when they used AI images generated by other people?....if not yet, I would think the AI sites will "wise up" and start charging for usage scope?

Looks like any problems with plan overlaps have been resolved.

My prepaid bonus Photography plan from last year ended on April 28, 2024. I didn't wan't to forget so instead of worrying about losing a few days, I used this year's bonus code on April 24 and the extra days were added automatically, with my plan ending on April 28, 2025 now.

Thanks again Mat & Adobe!

Interesting thanks!...after initially having the "parallel plans" or "overlap" problem and then getting a new code, I have been holding off on trying it. Maybe I will give it a go!

I still see variability in Firefly Version 3 Beta quality or results. For example, sometimes when rendering people you might see quite realistic results including skin and hair and other times, a person's hair may be approach a "blurry mess". It can be frustrating. I have to wonder if it comes down to how much "CPU bandwidth" is being allotted to a render at a given time. Maybe if there is a large demand on compute resources when you choose to do a render, the system just doesn't "spend as much time" on your render and you get poor results?

Interesting, I guess i should have been more careful about the distinction...yes in fact i am inquiring about "Adobe Firefly 3 preview" (not beta).
What you say makes sense, but i would also like to hear it from "Adobe"....Mat?

Where does image retouching fit into this? Should only completely unretouched images be allowed as "natural creations"?

Great...anything to make it easier is welcome....now if Adobe would only start accepting editorial videos 

Thanks!

Does anyone know if you shoot a video (or for that matter stills) on an Android phone, is any of your user info/data embedded in the file created on the Android phone?

Do we need to be concerned from a security standpoint about uploading an unmodified photo or video file to a stock agency directly from your phone?

Thanks!

I am still in the early stages of learning about midjourney. While you cannot get a series with a specific model you can get a theme of files with similar vibe, colors, lighting.

Midjourney people files still need a lot processing, I get a lot of great looking files, then on close up discover all kinds of issues with broken teeth, weird fingers growing out of a neck, strange details on clothes. The experssions are better than on stable, but far from perfect.

But the biggest challenge is to get faces that not everybody has. It feels like Mj just has 10 faces to choose from for a given age and ethnic type.

I am also trying to do more with Adobe and next week I have my first real shooting with a camera again.

But like I said before, the main focus for this year will be ai illustrations and to learn what I can offer that people will then like to buy. Every illustration sale, is a sale I would not have had otherwise.

They are selling very, very slowly, even the holiday files sell slow. But I believe it will pay off eventually, once the initial files have laid their roots across the algos, which will probably take 3 years.

Regarding "number of faces", this is one thing I don't get. I have also seen on Firefly where faces get almost repeated. With all the vast quantity of images these tools are trained on, how is it that they get stuck with a limited number of "faces"? Is it some kind of "flaw in the logic" of how they work? Is it a bug? If you take human artist, say painter, they don't repeat faces in their work, yet AI seems to do that. Maybe it's some kind of short cut in the algorithms which leeds to this? Just curious.

is it me or does it seem like at times Firefly produces decent results and at other times it's near impossible to get something usable?