I need lots of cheap, crappy stock photography- where to find?

July 12, 2004 9:06 AM

Looking for cheap, crappy stock photography. I recently revived an old project of mine, a photomosic program, and I need source material, lots of it.

I wrote this program a long long time ago for fun, and recently decided to rewrite it again now that I know some math. I'm making good progress. I'd provide a link but at the moment it's highly unstable.

Anyway, I need hundreds of thousands of images for this to work well. They don't need to be high quality, in fact, images that are around 200x200 pixels are really pretty perfect.

I have gone the route of buying crappy stock photo cds at $5-10. These generally have about 10,000 images each. So I'd need, say, 20-50 of these to really get going. Ouch.

There has got to be somewhere that I can get stuff like this. Most stock photo sites offer previews that are about the right size but they're generally resistant to spidering. For the last week I've been (very gently) downloading lots of movie covers from amazon, but even if I get ALL of their covers it won't be near enough.

To make things worse, I probably actually need MILLIONS of pictures, because someone might want to, say, make a mosaic entirely out of pictures of animals or something.

Anyone have any ideas? Anyone want to donate CDs of crappy stock photos? My company has 10 or 20 cds of stock photos that I was drooling over. Looked at them today, and they have only 100 photos per CD, since they're intended for print use. Doh! I want the exact opposite.

posted by RustyBrooks to media & arts (26 comments total)

Oh, and, another reason I'm not posting the link to my site just yet is that, well, the most readily available source of images is porn. So my current "library" is about 250,000 porn pictures.
posted by RustyBrooks at 9:16 AM on July 12, 2004

I was going to suggest

www.sxc.hu

It's a free stock photo library. However, i don't think there's millions of images in there! Last count was 93,000.

Maybe you could try some sort of automated script using the advanced Google image search? Typing in "cat" gives me over a million results...
posted by derbs at 9:26 AM on July 12, 2004

sounds a bit like debris--which uses random images found on the internet... which equals a lot of porn.
posted by lotsofno at 9:34 AM on July 12, 2004

toy could use the random personal picture finder at http://www.diddly.com/random/ and harvest results. Or just copy it's search term generation styles.
posted by Flat Feet Pete at 10:40 AM on July 12, 2004

Have you checked out openphoto.net and y6y6y6's pd photo site?
posted by mathowie at 10:41 AM on July 12, 2004

These all look like great sources, thanks. I have plenty of time so I'll take it easy on these sites.

I am trying to avoid totally random pictures if possible, because I would like to be able to organise them by general theme (without categorizing that many pics myself)

Debris has some seriously cool output, thanks that link.

I actually have sxc.hu on my list, to look at later (it's currently down)
posted by RustyBrooks at 10:53 AM on July 12, 2004

You could give istockphoto.com a try. Won't have millions of images, but it's pretty darn cheap.
posted by ssmith at 11:22 AM on July 12, 2004

Hmm. Am I missing the idea of istockphoto or is it really 20 web-res pictures for $10? (with slightly better deals for larger purchases)
posted by RustyBrooks at 11:30 AM on July 12, 2004

"y6y6y6's pd photo site"

I have around 3000 images. I could be motivated to write a script to pull out just the ones that are public domain (the vast majority) and send you a CD if you think such a small number would be worthwhile. Just ask.

I've been thinking about selling something like this anyway. Especially after finding out this site spidered everything and is basically selling a stripped down version of my website for $20.

This page also has some info on government public domain archives.

My list of public domain photos, which will lead you to a huge number of PD sites, is here.
posted by y6y6y6 at 11:33 AM on July 12, 2004

y6y6y6: I'm not sure 3000 images would be worth anyone going to the trouble of making a physical CD. I'll add it to the list and keep it in mind though. I am targeting larger numbers now in order to keep my personal involvement-time low. Like I said I have plenty of time to wait for pictures to download, I'm in no hurry, but I'd like to keep my sorting, maintenace, etc time as far down as possible.

Nice of someone to steal all your stuff. I'd sell it too if you could get $20 for it. Maybe you should rip off their ad and use it to sell your CD ;) I once saw something that I'm *pretty sure* was a photo of mine on the cover of a very low-run financial magazine.

Someone emailed me that I should rip frames from movies. Not a bad idea, really. Taking every frame would be a waste of time, but say, one sample every 5 seconds means 1000-1500 images per full length movie. There's always the copyright issues though. Maybe I should set up a half dozen web cams in my office and capture all the wackiness.
posted by RustyBrooks at 11:42 AM on July 12, 2004

"Nice of someone to steal all your stuff."

[off topic]
Well, it's not really stealing. I've explicitly put the images in the public domain. Which means people can do whatever they want with them. Even download them all and sell them.

It does bug me that they don't give me credit anywhere. But while that pushes them into the sleezebag zone, it's still 100% legal.

Also - They sell this thing on eBay.
posted by y6y6y6 at 11:56 AM on July 12, 2004

try photo.net, PBase, Yahoo news photos, or even GIS. That GIS link for "picture" results in 2.6 million hits alone...
posted by Kwantsar at 1:24 PM on July 12, 2004

try photo.net, PBase, Yahoo news photos, or even GIS. That GIS link for "picture" results in 2.6 million hits alone...
posted by Kwantsar at 1:28 PM on July 12, 2004

photo.net is actually one of the sources I used long ago. I should revisit it, definitely. There is probably a lot more pictures there now than there used to be. Back in the day the only pics there were philg's photos.

Google will be a last resort. Totally unsorted, of extremely varying quality. I guess I don't want stuff quite as crappy as a lot of stuff on google is.

Yahoo news photos is a very intriguing idea, I'll definitely look into that. Particularly since I've been wondering where I might get a lot of pictures of celebrities. imdb.com is on my potential list as well for the same reason.

One of the main criteria I have for the images I'm looking for is recognisability. The images in the output will likely be quite small and it would be nice if someone could look at the image and say "Hey, this is a picture of an academy award made out of actors and actresses" or "This is a picture of roger stalbach made out of sports photos" or some other such nonsense. Right now all I have is "This is a picture made out of porn!" which for some people, a definite plus, and for others gets me the response "Dude, I'm at WORK".

Lots of great advice here, you guys are really pulling through for me, thanks!
posted by RustyBrooks at 2:25 PM on July 12, 2004

Maybe I should set up a half dozen web cams in my office and capture all the wackiness

Look into model releases
posted by nakedcodemonkey at 2:39 PM on July 12, 2004

After doing some poking at photo.net I'm thinking you could spider stuff there and get a couple million images.

But that leaves me with the question - What if someone doesn't want you using their photo? What if photo.net objects to you grabbing images from thier site? I'm betting Yahoo's lawyers would have a few things to say about it.

And I suspect these sites already have things in place to defeat spiders. But you could probably get 100,000 before they blocked your IP.
posted by y6y6y6 at 2:40 PM on July 12, 2004

I just found out that yahoo, at the very least, does not like you grabbing a few hundred pages all at once. I'm seeing what they feel about 5 pages a minute.

It's a sticky question. I personally would not object to anyone using my pictures for such things, but I understand that some people would and I understand why. I really don't want to be an asshole but I do tend to do this stuff without thinking about other people's potential reactions too much.

I'm going to have to think about this a bit. There's quite a bit of public domain material in this thread, I'll see how far I can get with that.

Regarding fair use... I'm not lawyer but as I understand it we have limited rights to reproduce materials in magazines, newspapers, etc. This program is for my private use. If I'm going to use it commercially I shouldn't have a problem securing the images I need, I hope. Yahoo can certainly object to me downloading a lot of pictures, but do they have a legal right to prevent me from saving it, printing it out? What about printing out a bunch of pictures and making them into a collage which I hang on my wall?
posted by RustyBrooks at 2:56 PM on July 12, 2004

I was joking about the webcam thing. Actually we do have a half dozen webcams around the office that are for security purposes. You can even go to an (internal) website to view them. I had to move one that is above my cubicle, to, ahem, not see my screen. I just didn't like the idea. It can see my bald spot though.
posted by RustyBrooks at 2:57 PM on July 12, 2004

The net is full of generous people. Instead of filching (at a miniumum, quite a lot of bandwidth) from various sites, how about doing what many of the other free photo archives do: leveraging the power of the masses to do so much collecting and tagging.

For instance, set up a bulk upload/categorization interface, and solicit submissions for your project. The donors warrant that they have the necessary copyright/license and model releases to donate their photos for this purpose, and you warrant that the donated photos will not used for other purposes.
posted by nakedcodemonkey at 3:03 PM on July 12, 2004

U.S. government photos are PD. See what the NASA, FAA, JPL, USDA, EPA, Interior Dept, etc. websites have in the way of public archives.

Once you know what kind of photos they keep on file, you could do a Freedom of Information Act request for specific collections (e.g. grazing animals, 1930-2004; rocket & shuttle launches from Cape Canaveral; employee ID card photos, 1995; etc). Same principle used by the Memory Hole to obtain all those military coffin photos.
posted by nakedcodemonkey at 3:13 PM on July 12, 2004

Naked code monkey, it's not a bad idea, and all afternoon I've been contemplating how I could set this up as a service to other people without getting hosed myself. It bears thought. I actually have some software that I wrote some time ago that might apply here. If I made a service like this, I would probably make it something like what I want, instead of general purpose. That is, it takes a lot of bandwidth and processing power to browse images, select the ones you want to download, etc, and it's particularly bad if people are spidering you. But, I could make it so that you could do a search and download (or have mailed to you or something) the huge chunk of photos that the search returned, or you could download categories, etc. I don't really have the time resources to work on that right now probably though.

It raises the issue of liability also. If it is not currently the case that I could be brought to task by uses who uploading infringing images, it may soon be. As I recall there is pending legislation regarding providing a vehicle for someone to circumvent copyright. I could be wrong of course. There are steps I could take to protect myself but I fear anything I built could get turned into a big unsecured ftp site for people to drop wares, porn, what have you.

On preview: I'm in the process of slowly downloading NASA's photos. Wow, FOIA, that's heavy. What an idea. Maybe I'll get some huge boxes of CDs in the mail soon ;)
posted by RustyBrooks at 3:18 PM on July 12, 2004

Oh you could just ask for every digital photo that a particular dept's got on file, and wait for the envelopes trucks to start arriving.
posted by nakedcodemonkey at 3:22 PM on July 12, 2004

I just found out that yahoo, at the very least, does not like you grabbing a few hundred pages all at once. I'm seeing what they feel about 5 pages a minute.

Yahoo hasn't done anything to suppress the PERL Finance::Quote module, which grabs all sorts of financial data from yahoo. Since finance people are (at least) as possessive of their data as creative types are, I would guess that Yahoo is doing simple bandwidth filtering. Also, most of those photos come from press agencies like AP, Reuters, etc. No one has shut down Fark, B3ta, Worth1000 or SomethingAwful for their frequent photoshop contests. Since these frequently are based on Yahoo News photos, that might be good enough precedent for you.

Google will be a last resort. Totally unsorted, of extremely varying quality. I guess I don't want stuff quite as crappy as a lot of stuff on google is.

Google is only unsorted if you don't give it any search terms. diamond gives 828,000 different interpretations of the word.

I was joking about the webcam thing. Actually we do have a half dozen webcams around the office that are for security purposes.

I had a single webcam at a previous job. Over the course of about six months, it faced out over seattle, inside the server room, and occassionally at people. It took pictures at either 640x480, or 320x240 (don't remember which) several times per minute. I still have two DVDs full of images that may be interesting to you.

Also: you may want to try extracting frames from movies. There are command line utilities for all platforms that will turn a single MPEG file into a very, very large pile of jpegs.
posted by Kwantsar at 3:24 PM on July 12, 2004

I'd be cautious with photo.net. Some of their regulars get pretty touchy about unauthorised use. Here's their terms of use.
posted by normy at 3:44 PM on July 12, 2004

Looking at photo.net makes me sad. I am not worthy.
posted by RustyBrooks at 3:52 PM on July 12, 2004

Self link.
posted by armoured-ant at 7:25 PM on July 12, 2004