Panda, an open source project, will let any site owner willing to do a little coding and integration work to allow user video uploads and playback. Think YouTube in a box.
The software itself is free and will run on Amazon Web Services EC2, S3 and SimpleDB. You’ll have to pay for the Amazon services, but this is a nice step forward from a variety of existing paid services out there like Zencoder, SesameVault and Hey!Watch. Panda handles all aspects of uploading, transcoding and streaming, handing things off to a Flash player like JW FLV Player by default.
New Bamboo, a UK Rails consulting startup, is behind the project.








See all



That’s really cool. Talk about online video growing even more … with these types of services, there are many holes in niches that can be filled.
i gree with you the services is cool, but just one problem, how about video quality?
It uses ffmpeg.
And I’d like to state — could it possibly be worse quality than YouTube???
Does anyone know what they are using for video conversion? FFMPG?
What a beautiful gift of technology and effort by the Panda team. I suspect the open source community will love this product. Do you think the pricing at Amazon will become competitive on a bandwidth. I did look into using the EC2 service for a site consuming 10 mbps and its much cheaper to run your own network. Maybe the EC2 is a great place to prototype your video idea and if it gets traction you migrate it to a more affordable network. I could be 100% wrong about the pricing, I did my cost calculations back in march. Any feedback welcome.
Hi Wayne,
Can you expand on what you mean by “10 mbps”? Amazon charges for bandwidth used, not bandwidth per second, so I’m sort of confused.
In any case, what you’re probably not factoring into your calculations are the scalability at AWS, and the “pay only for what you use.” With your own network, you’re paying for what your line is capable of handling each month, or a set tier. At Amazon, you literally only pay for what you use. Last month my bill was 6 cents, because we just established a new account and did almost nothing with the account yet. Next month it might be $70 if we run a basic EC2 instance 24×7.
Imagine scaling your company up to 50 web boxes for peak traffic, and down to 1 web box at midnight on a sunday. That’s the real advantage to this type of setup.
The best solution is to use AWS but have one or more (squid) proxy cache machines running on your own servers (with presumably much cheaper bandwidth). So all your most frequently accessed files will be served by the proxy cache (saving $$) while less used files are served by AWS.
Best of all worlds (price/scalability/reliability).
Indeed. You can easily start making use of a proxy or CDN as your demands grow.
What I meant by 10mbps is that I’m consuming that much bandwidth, which is about 3000 GB of transfer. Based on the Amazon pricing it would cost three times more for the same bandwidth.
A lot of people use S3 for storing high volume content, but cache and serve popular content off their own servers for just the reason you mention.
Seems like a small market since my guess is that most website owners would do just as well or better to host their videos and use YouTube’s API.
If you use YouTube’s API, you’re not hosting the videos, YouTube is. You’ve then just lost all control over that content. You also have limited flexibility, and you’re at the mercy of what kind of video encoding YouTube wants to do to your video.
I see a huge market for this product among the many social networks that don’t want to be tied to YouTube or similar.
…and you could already do this on drupal as well, not sure that this is even the 10th let alone the 1st…
Is there a company doing something similar for video conferencing services?. Thanks. Max.
Great. They should offer this platform generically for all media types: photos, documents, audio, etc.
the tough part about video encoding is dealing with encoding error handling and scaling job management, I prefer to interface with an API that handles this like encoding.com rather than installing and debugging on my own EC2 account.
The article seems to imply that that’s what this software does, although I haven’t looked into it.
Another video platform social networking site…wow
great thing about this one is that it’s open source and free to use
The video player license does not allow commercial use. So, right off the bat one would have to redo all the work to integrate with another video player.
It’s dirt cheap to license it commercially though. Well worth it.
So, you work for Panda? Why don’t you just say it? What is the yearly licensing cost?
This is the new “open source”?
Complete bullshit, “open source washing”.
Who the hell is going to take on hosting costs (amazon or other) and not want to commercialize their site? Please define “dirt cheap” in unambiguous terms.
You can use any Flash video player you like with Panda. The JW FLV Media Player just happens to be by far the best one out there!
Oh, so the commercial license costs are @#$% per year?
Thanks for the info! Really helpful!
@Frank see http://www.jeroenwijering.com/....._questions
A single site license is €20.
@Frank,
Settle down there bub. It’s 20 Pound sterling (or Euro?) for a site license which seems very reasonable. The only reason I didn’t put the price in my post is because I have no idea how to create a pound symbol on my keyboard.
By the way, it took me about 18 seconds to find this info with Google. Maybe you should learn how to use it, then next time you don’t have to write douchebag remarks because someone didn’t go fetch you the information you wanted.
Jerk.
Any site using this where we can se quality of video? Link?
http://www.protectthehuman.com.....ugust-2008
to see it in use.
Looks like a bunch of Panda employees are running around commenting without identifying themselves. Just lost my trust.
Definitely company shills in here. Integrity, people! Come on now.
I’ve been posting on TC for about three years, and this is the first time I’ve ever been accused (as you accused me in a previous comment) of being an employee of anything.
It’s pretty funny, but you might want to adjust that tinfoil hat of yours Frank.
My team spent a week building this type of integration into out RawVoice Generator Platform and our users love it. Will be interesting to look at the code and see how they did their implementation.
This looks cool if it handles the EC2 instance management automatically, but is it really open source? I ask, because I couldn’t find any license. No license != open source. Visible source code != open source.
One might guess that it is MIT licensed, since that seems popular in the Rails/Merb community, but one can’t safely make assumptions about such things.
Apologies for any confusion. It certainly is MIT. There is a more explicit license in the main repository now.
Panda looks like a nice product that meets a need. In discussions with customers and potential customers at Zencoder (where I’m a partner), people who are looking for a video transcoding platform generally fall into three camps:
1- People who want something simple, cheap, and low-volume (the encoding.com crowd, though there may be another entrant into this market soon).
2- Companies who tried to build a transcoder in-house, but realized after a few months that it’s a much harder problem than they thought.
3- Companies who don’t want to worry about transcoding at all, and want to completely outsource this infrastructure to the safest option.
Panda seems a great fit for option 2, and sometimes a fit for 1 and 3 as well, though 1 may not want to run a transcoding infrastructure, and 3 may want a tech support staff they can rely on.
what about encoding.com is low volume? I guess I am part of the encoding.com crowd as we process over 2000 H.264 videos a day with them.
Right - it isn’t that encoding.com needs to be low volume, it’s that people looking for low volume often want something like encoding.com.
What would it take for this service to use MOSSO as a cloud instead of Amazon mix of ec2, s3 and simpleDB?
Why should this not be hosted on Mosso? Their media accelerator (http://help.mosso.com/article.php?id=250) promises better performance, and pricing+setup of MOSSO is a lot easier and attractive - increment CPU cycle purchase & free incoming bandwidth
The problem with server side encoding is that you always have to deal with a queue and how to manage it. The main issues are cost, bandwidth, scalability and error handling.
My company Framecaster (www.framecaster.com) is launching a distributed encoding and aggregation model for UGC that eliminates the need for a transcoding server or cluster. All the encoding happens at the submitting computer through a web plugin (IE or Firefox) to your specs (the aggregator)
It is an equally attractive model for start ups with unknown levels of video submissions or for companies that have extremely high volumes and are looking to fix their cost.
Check it out http://www.framecaster.com/howitworks.html