|
MicrostockGroup Sponsors
This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.
Messages - Andrej.S.
Pages: 1 2 3 4 5 6 [7] 8 9 10 11 12 13
151
« on: April 04, 2024, 04:49 »
Top 100 weekly, top 100 lifetime, less than 10k.
Congratulations on your hard work!!!
No need to share details, I think it is already very helpful to know that such a position can indeed be reached by a single artist.
I always thought the top 100 was for large studio/team producers only. The only person I know in that level started out as a single artist but had progressed to working more as an art director and has people working for him, some full time, some on work for hire basis.
I do believe in sharing helpful information with peers and it can be done without reveailing best selling content.
Plus the algos make copying sometimes harder than people think. I know many producers who did a ton of seasonal content with ai, but get few sales, while my christmas season was quite good. But I have really, really old files that have probably accumulated in many galeries over the 10 years I didnt upload and these are the ones that started bringing in the money again. The ai stuff also sold, but less than old files revived.
For me I see it with my ai people, while they are slowly starting to sell, if I just do carrots on white background I will have better sales with that.
And if I do that with my camera, the carrots can go everywhere.
However, if I am consistent with my ai people uploads, then I am sure in 3-4 years they will also get good sales, like anything isolated or my greeting cards.
Same for illustrations. It is a new medium for me, there is ZERO algo track record. I have uploaded a lot of seasonal illustrations but sales are very, very slow. And sometimes very unexpected, I am actually having artsy files sell. Done in watercolor or oil painting style.
So my first 3000 illustrations will be test subjects, trying many different niches and many, many different art styles.
Then gradually the customers will tell me what they like and I will make more of that.
But if I ever wanted to got for the top 100, I would certainly need 30k plus files.
I do a lot of research, but wouldnt know how to do that with less than 10k even if I could identify 10k top selling concepts. Gallery uptake by clients is a very slow process. Having instant best sellers is very rare for me.
There is also the time factor, I want to grow my income over the next 2-3 years, not over 20.
Roughly 60% of what i upload is still experimental, with ai even higher, because I am testing a lot of new things.
Thank you to anyone in top levels sharing ranks and port sizes.
For newbies it should help them understand that stock needs a ton of research and is not a slot machine where you throw random files in and money comes out.
The money is related to usefulness for clients not overall random content port size.
I don't think it would be a good idea with your current portfolio. I rather have the feeling that Adobe penalizes spammer portfolios heavily and lowers the ranking. I would always test experiments with a new account (multiple are allowed) to avoid your main portfolio crashing. I would even go so far as to create a separate account for each type of image (interior, fashion, seasonal content, etc.).
152
« on: April 04, 2024, 03:23 »
On my way back down
pos 2240
you go down and I go up  but I'm always far from you rank,even though I have more content than you,I've been on Adobe Stock for much less time than you.
I don't believe it is the time factor, which is relevant to be pushed up by Adobe. I think it is rather a well balanced mix of real photos and AI images that is promoted by Adobe. A good example is the following portfolio (although he steals a lot of ideas from other contributors) https://stock.adobe.com/de/contributor/201195540/igor-link
153
« on: April 02, 2024, 10:18 »
Broke top 500 lifetime rank 491 and holding steady around 200 weekly (189 at the moment).
Nice, what's your weekly earning with this position? 1k USD or more? Would be interesting to know how the earnings are distributed among the top sellers. How big is your portfolio? Would expect min. 30k images and videos.
154
« on: March 29, 2024, 14:40 »
after a strong Monday the rest of the week was pretty bad,not even an Easter sale 
I don't know if you noticed but there are 1,076,612 results for "Easter" AI only,already a million AI contents for Easter.
this is why I am not particularly convinced that investing a lot of time in AI content is a good idea in the long term,for now yes,but I think in a year at most I don't think I will produce any more AI content,then clearly I will see,it depends on the results,but with this mass production so fast,I believe that the real difference can only be made with real contents at the end of the day.
Well, it's not as if the competition is getting smaller. Yes, the majority of AI images are currently still garbage because it takes a lot of time to find good generators and diffusion models. But the bar is definitely getting higher. Many former classic photo bestsellers from 5 to 10 years ago have been overtaken by exceptionally good AI images. My strategy is looking for niches right from the start. Popular subjects are pure luck or one have to create extremely creative concepts, which would have little competition. Easter, for example, is brutal. You won't stand a chance with standard motifs. You have to combine Easter with other themes. Vacations, online shopping, promotional offers for workshops, fashion, lifestyle, etc. In other words, what a creative director would have photoshopped manually from standard stock material in the past. And you have to build up a very balanced portfolio with many themes. Otherwise one have no chance nowdays.
155
« on: March 27, 2024, 13:03 »
In my opinion Adobe is even not selective enough.
The point is that the majority of AI images are not sold anyway if you have not won the lottery image ranking or if they are not top notch unique without flaws.
I mean yeah I can generate a port of dozen thousands images which will be accepted but it will be just a waste of time.
156
« on: March 27, 2024, 06:19 »
@stocker2014
If you apply your logic to the pre subscription era, then we should all have become millionaires with subs because of more sales.
But the opposite happened. The earnings have almost completely disappeared. Some contributors had earnings between 20k and 50k a month (!) and fell to just 1k or 2k USD.
The same will happen with videos. You just have no long experience in this industry.
157
« on: March 26, 2024, 14:24 »
The question I have for the people that got heaps of refunds, were your earnings for November oddly high (by about the same amount as the refunds)?
Pretty shoddy on the part of Getty all around.
That's really good question. Went and checked: Nov '23 was my 3rd best month last year on iStock, but nothing out of proportion.
Accounting certainly appears shady, but I don't think it's deliberate - whole system is set up in such a bad way it is easy for all kinds of mistakes to creep in. Biggest thorn to me is 15% royalty which is downright disrespectful.
Agreed
Pretty funny or odd, or I ask, how do people reason out the glaring contradiction, when SS pays us more and a higher percentage, with a higher minimum, but they can't stop shouting how bad SS is or what crooks they are? "I'd never upload anything to SS, paying 10 cents is an insult."
I won't dwell on Connect which is, fractions of cents, because no one really knows what or how those images are used. Maybe for mock-ups, or maybe not. Maybe on a website, or chat, or somewhere in Asia for some mystery use, and maybe not. We don't get anything detailed or specific. Oh but we do get hundredths of a cent royalty.
BUT... for iStock when we do get the details. How do we get downloads like this? 4 cents? 3 cents?
Type Regular
Month 2024-02
Sales date 2/5/2024
Uploaded 5/30/2013
Gross royalty $0.04762
License Fee $0.31745
Rate 15%
File Type Editorial Image
Type Regular
Month 2024-02
Sales date 2/6/2024
Uploaded 10/6/2011
Gross royalty $0.03150
License Fee $0.21000
Rate 15%
Please tell me how SS is nothing but despicable crooks, and then someone explain, how iStock isn't worse?  Oh IS is much better, they pay fractions of cents, and then 3 cents or 4 cents. Really? That's better than the terrible, insulting, demoralizing... 10 
Brutal, how is this even possible? 3 to 4 cents comissions? What the heck? I wouldn't be surprised if all agencies are ripping off their contributors since they all have become to greedy. Didn't we also have the thread about Shutterstock, where even sales were not passed on to the contributors? IStock states that they pay out almost 2 million a week. Presumably that's a similar number of sales. So mistakes can happen ... You probably don't even notice smaller ones and they are simply covered up. Nobody can check the sales system.
158
« on: March 26, 2024, 05:17 »
It is absolutely incomprehensible to me - especially in times of ever-improving AI - why newly uploaded images are not compared with images already in the database during the review process. This could be an automated process.
New images (also mirrored) could be checked as easily as with the google reverse image search.
Perhaps there is a reason why not much emphasis is placed on this.
If it is discovered that someone has an account with stolen images, the account will be closed and the money that has not yet been paid out will certainly be withheld. The return of money that has already been paid out may also be demanded. So it's a win-win situation for the agencies because they get the money for selling the images but don't have to pay anything to the contributor.
And the contributor who actually took the image does not receive any compensation anyway.
Conclusion: Accounts with stolen images and videos that are discovered are a lucrative source of income. Therefore, the effort to prevent this is non-existent.
Yeah, sounds like a disgusting strategy of a further increase of earnings. Brutal, what a mess this industry had become. I would never ever invest time of producing real photos anymore.
159
« on: March 23, 2024, 08:58 »
Brutal, does it mean that someone with a Super Duper Business Premium account has downloaded the entire database?
Who can actually guarantee that it wasn't misuse and that the images are now being used illegally for training AI models or offered elsewhere?
I mean, you can see from the account downloads that something is wrong. Why not prevent this in advance and block the account in advance?
160
« on: March 22, 2024, 20:12 »
In future I will now submit portraits only with the minimum resolution. The upscalers are currently still too poor for portraits. I will keep the originals in case better upscalers will come onto the market.
Other content such as still lifes (spa, flowers, animals, etc.) I will continue to upscale, which works quite well.
So, thank you for finetuning mine settings!
161
« on: March 22, 2024, 16:44 »
Hmm, I'm thinking about switching to the minimum resolution of 4mpx. You might be right. This resolution seems to be really better (hair is still a big upscaling problem). Two more examples: https://ibb.co/VJZBD6nhttps://ibb.co/m4VC1qHMan I don't know. Don't wanna rescale my port again ... already to much time wasted. Gigapixel sucks alot. Don't understand the hype about it. Will use probably waifu2x or neuralwriter instead. It's not comfortable with batch processes but is much faster and seems to be slight better.
162
« on: March 22, 2024, 13:29 »
Just out of interest, could you upload a sample image to compare the upscale quality? I have attached mine (not retouched yet with color postprocessing): https://ibb.co/pWJrPpP
163
« on: March 22, 2024, 06:29 »
Just had a huge day today. On it's way to BWE for photos on Adobe Stock. Thank you for your support!!
Was wondering how many photos you have on your port. For me if I sell a file a week, I'm lucky.
I have about 10,400 photos right now. About 4,200 are AI generated. How many do you have? Less than 10?
No man, around 2600 images and 500 videos. Probably because I have started recently.
Not gonna lie, that's pretty bad. I have currently a port of approx 700 AI images and weekly up to 10 sales, which is also pretty bad. I have added in the beginning much crappy stuff, which I now sort out and upload new better content.
164
« on: March 21, 2024, 10:05 »
There is an upload guide for AI images on adobe's website.
It depends wether the images are hand drawn illustrations, 3d rendered stuff, etc. or if they contain photo realistic portraits, etc.
First should be submitted as illustration and second as photo.
165
« on: March 21, 2024, 08:44 »
So, since I had massive rejection on GigaPixel4x upscaled photos of Midjourney generated AI images, I compared different options and found Midjourney's own 2x upscaling to be the most natural looking. I've been uploading the 2x Midjourney upscaled photos and are getting good acceptance rates on Adobe Stock these days, but my concern is 2x upscale may not be big enough for some buyers. So, my current option is to have Midjourney 2x upscaled photos to 2x upscale on Gigapixel AI on standard setting. It seems like the standard setting mostly just double pixel count without overly make the photos look artificial. What's your best option in upscaling AI images? Do you think 2x upscale is enough for long run? I'm worried I may need to re-upload 2x upscaled photos in 4x upscale to increase sales for buyers who want more pixels. I was thinking maybe the buyers will do own upscaling if they want to. But 2x upscaled photos look the most natural without looking artificial.
Since you're one of the contributors with most sales here on the board I would suggest to track your sales (via google lens, etc.) and look up in which resolution the images are used. I currently upscale also 2x and with lower resolutions about 4096 x 2288 or even lower. Can't imagine higher resolutions are used that much, so I would instead use the setting, which leads to the most high acceptance ratio.
166
« on: March 20, 2024, 07:41 »
One of my pasttimes is to go to book shops and flick through the first page after the cover to see where the image was taken.
I do see plenty of SS image used on covers (as well as Getty, less so AS), but it's usually as some part of composite with another image or more from the larger and more artistic agencies, Arcangel/Trevillion.
I don't think any serious publisher would use a SS or microstock image on its own knowing full well a competitor or even random business could start using it on let's say a toothpaste ad (or much worse) thus diminishing its uniqueness.
So, I think for a simple image it's OK to be paid little even for a book cover as it may not be strong enough to be used on its own. Plus it's RF subs and probably already sold 100s of times anyway (and buyers know this).
I find it difficult to justify it that way. With an extended license, e.g. for print, the degree of commercial use is usually higher. You have to distinguisch between an use for some random news / blog article as a gap filler or an use for print like a book cover. A good book cover contributes significantly to a higher revenue amount, the commercial use aspect is much higher. The same applies to print on demand stuff like t-shirts, etc. So 10 cents are just extremely ridicilous low because the buyer will earn for sure thousand times more. Ideally the extendend license would guarantee that the image is not used hundred of times but only the one buyer owns all rights. The main problem is that no one is tracking the copyrights or the use restrictions (just like a half million prints, lol!), so such agencies just sell everything for some cents.
167
« on: March 19, 2024, 08:11 »
NVIDIA Launches Blackwell-Powered DGX SuperPOD for Generative AI Supercomputing at Trillion-Parameter Scale https://nvidianews.nvidia.com/news/nvidia-blackwell-dgx-generative-ai-supercomputingNot that I understand much about the technical specifications, but what sounds promising is the following sentence GB200 Superchips deliver up to a 30x performance increase compared to the NVIDIA H100 Tensor Core GPU for large language model inference workloads. So Nvidia released the old H100 just last year and now they release a new for AI optimized chip, which delivers up to a 30x speed up. Imagine how advanced AI will become the next years if Nvidia will hold on the release pace.
168
« on: March 19, 2024, 07:36 »
Hell no. It is a lot easier and much faster to take a good series of pictures and videos with a camera than to do stuff with ai.
At least for me.
It depends on what you do.
I only take photographs of real landscapes in the field, but I don't know what future I can have.
In my case (landscape and travel), creating images with AI is incredibly cheaper and faster than buying quality photography equipment, traveling to the most profitable locations, and hoping that the weather conditions are good. And then, often, you have to do double exposures and a lot of post-production work for sunrises and sunsets, if you want your images to be sold (On average I spend 30 minutes to an hour of post-production to create a photograph that sells well).
Snapshots don't sell.
A professional buyer still buys real landscapes because he knows how to recognize them, but what AI does for many is good enough and will lead to a future of mediocrity and falsity for commercial landscape and travel photography.
The biggest problem is that, if agencies do not give privileged space to real images, these will be lost in the noise of those made with, AI because the latter will be produced at much higher rates. Sitting at a desk, in one afternoon, I can create 50 AI images of the most beautiful locations in the world without leaving home.
Landscapes are an interesting motif since you can create with AI even much better looking images. I mean one have already photoshopped much back then before AI but it was still somehow a real landscape. Another important factor with landscapes is the aesthetic factor. Many photos are used for prints like calendars or posters. So you really can't make just a few snapshots, they won't sell. The competition with landscapes is like hell. Currently one is limited with AI to the low resolution, which is needed for the landscape photos. But I would guess when the resolution is not limiting anymore AI images will take over.
169
« on: March 19, 2024, 07:30 »
I uploaded a PNG of Christmas baubles in November and it sold a couple of hundred times.
Thank you! Nice to hear as it sounds better as I would expect. Not that I will start now creating hundred of Christmas baubles but this is encouraging to try some other objects. @SuperPhoto Yeah everything in microstock is like playing lotto.
170
« on: March 19, 2024, 05:00 »
Yes.
AI can't make anything that's AUTHENTIC. Can't do news. Can't do illustrative editorial. Can't do anything real.
AI can make pictures, but most of them are crap.
Then you have poor prompting skills or use the wrong poor image generators. With AI you can already now create even better looking generic images then real photos in small resolutions. Only in full resolution the AI images fell clearly visible behind. I would adjust my opinion. Unique AI images that cannot be imitated in reality should be even more expensive.
171
« on: March 19, 2024, 04:01 »
Has anyone tried isolated PNG objects, e.g. vegetables, fruit, cars, people, etc.?
Probably these are already also extremely oversaturated so I would have to find some niches.
I would have expected designers to need these resources for their projects like print, advertising, etc.
Is it worth the upload or are regular images better selling?
172
« on: March 19, 2024, 03:58 »
Hi everybody.
I have an established AS account which performs pretty well (no AI images). I opened a separate AS account for AI images 3 months ago. It performs relatively ok, but the wait times are incredibly long. It takes 4-6 weeks on average for them to accept 200 images, I'm capped at 500 in the queue and I have started thinking if it was a wise strategy. (The whole reason for having a separate account was "if they ban AI account then my real one will remain intact)
A couple of questions.
1. Is it true that established accounts that were there for about 10 years have much faster AI acceptance rates?
2. Would you advise mixing AI and regular content in the same portfolio?
Thank you!
Although it may violate Adobe's terms and conditions. So I would not recommend it either.
It doesn't. T&C explicitly state that you can have several accounts.
Interesting, then it would even make sense to open different accounts for different content. Then the portfolio would be clearly sorted for the buyers. @Elijah AI content take an extremely long time. I don't submit real photos since in my opinion it is not worth the time. Between 2012 and 2015 I submitted vector graphics, which sold very good back then.
173
« on: March 19, 2024, 02:30 »
@blvdone
I prefer not to say the precise number but more than 7000
I'm sure your strategy pays off in the short term,we need to see if it pays off in the long term.
for the moment we are still at the beginning,then we will have to see when the competition increases.
However,you may be right,maybe I waste too much time,but I like to put content on sale that I would buy myself.
I also tried a month as a pro plan with Midjourney,because given the speed with which many of you upload,I thought that the outputs of Midjourney were more correct but this is not the case,90% of the contents have generative errors,until now all AI image generators have errors,and many things to improve.
Honestly, there isn't much more to edit AI generated photos other than erasing brand logos in very rare occasions. I don't waste time tweaking extra/missing fingers or disfigured faces. I just don't submit those bad ones and generate till I get good ones. Time is money. I just try to maximize my output per hour/day. And based on the number you gave me, my portfolio is making 2x more $$ per photo. So, no offense, but I know what I'm doing and it's working. If your photo isn't selling short term, why can you expect them to do well in long term? It just doesn't make sense unless you are producing seasonal materials way in advance.
And you're right about generating AI images on Midjourney. Many of the generated images aren't usable. It takes time to generate images that you want. You need to work on the prompts and hope AI will give you usable images without extra/missing fingers and limbs. That's where my time is spent regarding AI image creation.
Yeah, same here. In the beginning I was fixing quite long generative errors until I realized I loose to much time by doing this. So instead I now first try to formulate a better prompt or search for a better Stable Diffusion model, which generates less errors. I currently spend max. 2 minutes for retouche after upscaling. With AI upscaler I even can automatically retouche minor errors or I'm fixing them very rough and quickly and the AI makes the rest. So it's as you said time is money and you have to maximize it.
174
« on: March 17, 2024, 18:43 »
@blvdone
I prefer not to say the precise number but more than 7000
I'm sure your strategy pays off in the short term,we need to see if it pays off in the long term.
for the moment we are still at the beginning,then we will have to see when the competition increases.
However,you may be right,maybe I waste too much time,but I like to put content on sale that I would buy myself.
I also tried a month as a pro plan with Midjourney,because given the speed with which many of you upload,I thought that the outputs of Midjourney were more correct but this is not the case,90% of the contents have generative errors,until now all AI image generators have errors,and many things to improve.
Honestly, there isn't much more to edit AI generated photos other than erasing brand logos in very rare occasions. I don't waste time tweaking extra/missing fingers or disfigured faces. Time is money. I just try to maximize my output per hour/day. And based on the number you gave me, my portfolio is making 2x more $$ per photo. So, no offense, but I know what I'm doing and it's working. If your photo isn't selling short term, why can you expect them to do well in long term? It just doesn't make sense unless you are producing seasonal materials way in advance.
And you're right about generating AI images on Midjourney. Many of the generated images aren't usable. It takes time to generate images that you want. You need to work on the prompts and hope AI will give you usable images without extra/missing fingers and limbs. That's where my time is spent regarding AI image creation.
Since you once wrote that your real photos sell better, how high is approx. the share of your weekly sold AI images? >30%? I currently enhance my port, it's really pain in the ass, even if one have already generated images. Hope it pays off, even if I don't believe it.
175
« on: March 16, 2024, 17:48 »
Hi everybody.
I have an established AS account which performs pretty well (no AI images). I opened a separate AS account for AI images 3 months ago. It performs relatively ok, but the wait times are incredibly long. It takes 4-6 weeks on average for them to accept 200 images, I'm capped at 500 in the queue and I have started thinking if it was a wise strategy. (The whole reason for having a separate account was "if they ban AI account then my real one will remain intact)
A couple of questions.
1. Is it true that established accounts that were there for about 10 years have much faster AI acceptance rates?
2. Would you advise mixing AI and regular content in the same portfolio?
Thank you!
1. No. My account is from 2012 and it takes the same (long) review time. But my guess is that there are pro contributors, who are prefered. They somehow manage to upload and get through the review process with over 30k AI images / month. 2. It depends. If you take real portraits shots and generate some 3D content I would split it. Although it may violate Adobe's terms and conditions. So I would not recommend it either.
Pages: 1 2 3 4 5 6 [7] 8 9 10 11 12 13
|
Sponsors
Microstock Poll Results
Sponsors
|
Show Posts
