DIY Web "Television" Station?

randomkind asks: "Media production can be quite a bit of fun. I think it would be extremely cool to run a television station, or at least a series of television-like shows. Luckily, the web offers an alternative vehicle for publication than the standard media-mogul run cable television providers. That is why I'm looking in to creating a website that would, in essence, be a completely free television station online, with pre-recording (and sometimes live) streaming video feeds and a regular programming schedule. Think of it either as the visual analog of an internet radio station, or a very glorified webcam. My question is: Where to start? What kind of digital recording equipment is suggested? What kind of editing and effects software is needed? What kind of content delivery method is suggested (ie Real, or some other streaming video plugin)? Any special hardware requirements? Is there anything else that I ought to know about creating and running something of this sort?"

"I am looking for decent quality, but cheap solutions. I have a few thousand dollars to invest, and will have more over time, but my friends and I are solely backing this project with cash from our pockets. I've got a decent amount of working space for a production studio reserved, and enough manpower to provide, at least initially, a few hours of braodcasting a day. I've got an overflowing abundance of ideas, and lots of enthusiasm. So, where do I go from here?"

Comments

Democratization of production/distribution

[ealar dlanvuli]

Hmm, now we'll have the TV industry wake up to the same threat the music industry is facing now. Not that the concept is not obvious (after all, at a low level everything is just a stream of 0s and 1s), but this must be a red flag in their face.

Much as we needed the recording industry over the past 50 years to press those damn CDs/Records and distribute them, we're currently relying on the TV studios and networks to make/distribute their products. Also witness the current TV climate: much as the recording industry creates their own hypes and ignores non-conventional artists, the TV (and movie) industry is falling victim of their own success. Their desire to standardize everything and make it 'safe' for (their) ideal targe audience (families with kids, etc) results in a product which excells in conformity and blandness.

Given this, advances in technology which make it possible to distribute (and eventually produce) decent quality TV programs at low costs, will lead to the proliferation of 'independant' studios. With their monopoly on creation/distribution of movies vanishing in internet time, the TV studios will eventually face the same tide the music industry is facing now: We don't really like them, we don't really need them anymore; let's move to a medium we can control and just ignore the studios. Looking at the sad state of the (currently +- 30) TV stations I get via cable, this may just be good ... if nothing else, this means that as an independant producer you will be able to distribute your films to anyone who's on the internet. Wether anyone will care to watch your stuff is of course an entirely different matter.

apprentice first, for a year

[ghostlibrary]

First, go to your local cable access and do a show on there. This will:

* get you familiar with producing a TV show
* give you useful contacts in terms of staff and crew that you'll need for your venture
* give you 'street cred' in launching your web venture-- or at least teach you the lingo so you can work with folks who produce things.

Only after that would I suggest you launch your web thing. That way, you make all your mistakes while learning.

Yeah, I know a station owner/executive technically doesn't have to know anything about actually producing a show... but you will need contacts with that experience to produce your actual content, and this is a good way to hook into it.

Good luck!

For starters

[dacarr]

If one has the material already available on tape, you can always get it to the computer with a video capture card of some flavor and encode it into the stream flavor of your choice, be it Windoze, Real(bad?)Media, or QT. If there's a way to stream MPEG or DivX encoding, I'd suggest that - but keep in mind that I have no experience with media streaming. Remember, if this is to be geared more or less like public TV, you don't really have to start off with professional-looking material - to use educational shows as examples, Sesame Street, one of the longest running public education shows, has a sort of low-budget feel to it, and Fred Rogers even went with low end budgets for his shows. (It's granted you don't have to do high-end stuff for kids, but notable nonetheless, IMHO.)

The problem you'll face here is bandwidth usage. An audio feed site only has to worry about the audio data being fed out - over a 56K modem, you can get fairly reasonable sound quality, if you do it right. Now consider that a DivX movie takes up, what, one CD for two hours, give or take? That's a pretty significant chunk of data to be slinging around. On the other hand, I think Mononoke Hime weighed in on RM at about 240 MB last I looked - but again, that's a lot of data to be pushing through the pipes, and will immediately alienate your dialup people. I recall seeing video pipes that could be used for as little as 33.6 somewhere, but even that will tie up the modem.

In short, you will probably need at minimum a cable modem or DSL to use this service.

Nonetheless, good luck.

Re:For starters

[ealar dlanvuli]

You don't have to do high end stuff for adults, I'd much rather fake props and mid-level sound mastering and interesting/exciting content over the MTV music awards.

Re:bandwidth costs and formats

[Locke!Erasmus]

I think this is actually an awesome and practical use for P2P technology that won't have people (read the TV industry) squealing about copyright infringement. If randomkind releases each video segment into the public domain, or uses the GPL for his video segments,

I think you are correct in suggesting that randomkind and his associates put together a website that shows their services. Their website could include information like show synopses, upcoming features, etc.

Perhaps they could include small teasers of each show or segment, and then release the full-length files on P2P networks like kaaza and napster if possible. The website could include tips for where to locate the files.

I think the cost of the bandwidth necessary to provide direct streaming, as well as licensing to record his segments in DivX, real, mpg formats would be prohibitive. Perhaps randomkind could allow people to get the segments off P2P networks for free and make donations if they are so inclined, and as well provide a subscription service for people who don't want to go hunt for the files and want to stream it directly from his site. This might help combat the costs.

Some comments, and recommendations

[megabyte405]

This is a great idea, but must follow a few guidelines, otherwise the TV industry will strike down the entire concept. Everything MUST be original, because if any of it at all is copyrighted by someone else, the TV industry will label it as the "TV Napster", and that'll be another entirely legal and intriguing concept, down the drain. That would be my greatest concern.

For hardware: I don't know if you want to use Linux or xxxxBSD, but I know that the ATi All-In-Wonder cards work great for digitizing video under Windows. Could be a cheaper solution than buying a digital video camera. Perhaps even a Hauppage WinTV card with a VCR to turn your SVideo into Coax. Anyone with more experience on these cards care to comment on their *nix compatability?

I'd say something that could be useful is embedding the stream in a web page, where in the bottom you could have a scrollable schedule, somewhat like the TV Guide channel on cable does. Don't forget a drop down box for time zone. And, as this will be new, consumers will likely want something altogether better than TV aka few/less intrusive ads. Don't sell out.

Great idea, and I encourage you to pursue it!

Re:Some comments, and recommendations

[Zarquon]

The ATI All-in-Wonder devices are nice, but there are some caveats:
1) They don't use the same colorspace as the BTTV cards.. the programs have to be able to handle YUV input.
2) Last I checked, the drivers didn't decimate / resize frames / fields; you got 59.97 fields of 640x240.
3) No mmap support.

Given that, the drivers were remarkably efficient and stable. I was working on VBI and closed captioning, but didn't get a stable release out.

This is fairly old information (since my dev box died), about 4 months out of date. Check out http://gatos.sf.net

For linux video, the best supported cards are bt8x8-based cards, including most of the WinTV line. Many of which do have have svideo and composite inputs, as well as tuners.

R C
(Former, and hopefully future, GATOS core team member)

Re:bandwidth costs and formats

[handsomepete]

That's a really good idea. I wonder if it would be possible to create a sort of website mirror/community that would be primary hosts for the p2p downloads (like TV stations in a way... perhaps one in every state would eventually be achieved?). Maybe only use the p2p for downloading of old/archived shows. Then have all of the servers simultaneously stream new content (using a local mirror might make modem users better able to enjoy all of this).

That would probably negate at least some of the bandwidth savings that p2p would create, but it would just be plain neat to have a full fledged internet TV network presence.

Use video when appropriate

[kabanossen]

I think you should combine video footage with some data formats, flash for example, to present diagrams, maps, tables and anything else that is related to the story you're showing. Use each format for what it's good at. Video is good for showing people and nature and stuff, not diagrams. It's soo irritating to watch news over the web where the publisher has just converted their TV show to Real format without converting graphics and text into native web formats.
You could also link your shows to each other just like links on a web page, that's fairly rare and unexplored. You'll probably need to experiment a bit but I'm sure you'll learn a lot in the process.

You should absolutely not do this

[gnovos]

This is a completely new kind of a venture that you are attempting, so it is imperative that you don't enter into it with any preconcieved notions of how it is supposed to work. I strongly believe that one of the reasons why we don't have any real webcasted TV now is that the major TV companies try to just do the same ole same ole and end up with a feed that's too bulky, or too uncreative to really work in the new medium. Forget going and seeing how it's "really" done until you have figured out a system for yourself, THEN go and see if there are any pieces you can improve with conventional techniques.

Re:bandwidth costs and formats

[man_ls]

If your routers support it, you can multicast Windows Media -- it routes the packets to multiple destinations at the backbone level, so your bandwidth useage doesn't scale directly proportional to the number of users.

Look into that...

Something to avoid (possibly)

[Joe Tennies]

MPEG-4 -- I know it's the latest and greatest and should work in many streaming media players, but please read the license. You will have to pay a lot to stream MPEG-4 video. If you can eat the costs, then it's great. The thing is that you were talking about free and that'll be tough. Something you may want to keep your eyes on is Theora (www.theora.org). It will combine Vorbis Audio and VP3 Video in OGG packets. Though it may be a while before it's up and running.

Ideas for programs...

[Odinson]

I have thought about this alot, and come up with a few ideas for cheap shows to start.

• Dead car lot wars
• A comedy news talk/show
• Garth's world
• Compuhelp netTV
• A church show (many churches would be interested I think.)
• Rebuildahouse.
• Kids book reading show
• How to repair your car
• Single indepth 3 minute commercials for comercial breaks...

Mainly HOWTO shows seem to me to have the lowest overhead.

We will win if we fight them with open content, just like fighting them with open code. I've been getting back into music lately because there is opportunity once again. I think things are actually going pretty well for new art. :)

DV over IP

[Nishi-no-wan]

There's a lot of information on Digital Video over IP here. Use a FreeBSD box with a FireWire (IEEE1394) connection to a DV camera and broadcast openly or to a specific address.

I've been playing around with this at home, and it works! Sound, too.

Since there aren't many people (yet) who have xdvshow to watch the stream, a method needs to be devised to pipe this to other formats. (I was investigating this, but ran out of time last weekend. If anyone knows of a DV to QT/Real/etc. converter, please let me know!)

fyi on real

[StandardDeviant]

This is just an FYI, and I do recall you *really* have to dig on their site to find them, but you can get the basic version of the content creation program (real studio?) and an up-to-8-clients version of realserver for free. I looked into this at one point about nine months ago for my soon-to-be-ex employer (we wanted to test drive their stuff becuase we wanted to migrate away from windows media and win2k in general). I've heard several times that real server is an UNHOLY pain to deal with administratively on unix and/or behind a firewall... Good luck!