Slashdot Mirror
HDTV On Your PC And Hard Drive
Jack Kolesar writes: "So, you want to watch HDTV but you don't want to shell out thousands of dollars for a new television. Well, AMDPower.com has a review of the AccessDTV HDTV tuner card. Not only does it let you watch HDTV, but you can also record it on your harddrive. Yes, the full 19.4Mbps stream of 8VSB is stored in raw format. Now, if somebody out there could just make some linux drivers for it ..."
According to the article, the recorded video is in fact encrypted - only the specific card that did the recording will be able to play back that particular stream.
As this is a 19.4Mbps raw format file. I persume that bits.
So.
19.4 x 1024 = 19865.6 Kbps
19865.6 x 1024 = 20342374.4 Bits Per Second
Now lets divide by 8
20342374.4 / 8 = 2542796.8 Bytes Per Second
2542796.8 / (1024 x 1024) = 2.425 Mega Bytes Per Second
Now, I would like to record a move of 2 hours
2.425 x 60 x 60 x 2 = 17460 MB
or 17460 / 1024 = 17.05 Gb
Thats alot of space , evan for a 80mb hard disk.
Just a question someone might be able to answer, how well will this compress ?
If its a good level of compression, will it allow a new way for the napster type people to break into a new medium.
Cruise TT
This PC Card, like most PC HDTV solutions uses the Teralogic Janus chipset. It's pretty slick, and has actually been aorund for several years. Many major companies use Teralogic, including Tivo.
I've been trying to get information on the chipset from Teralogic for several months. On Dec 29, 1999, David Auld of Teralogic posted to the video-4-linux mailing list. "We at TeraLogic are interested in encouraging the development of Linux
drivers for the Janus DTV card." The company went so far as to offer reference cards and driver sets, and was in favor of having a total GPL driver set. You can do a google search to find the archive.
A couple months ago I e-mailed David on this subject and got a fairly kurt thanks but no thanks response.
The obvious reasons for pulling out support for the Linux driver are all MPAA based. The content controls comming down the pipe won't be in the Janus Chipset. It would have to be software based. With a linux driver could could patch an HD-Tivo, or your Windows based solution to ignore the content control flags. Most interesting would be trying to wield the DMCA against people on this. It's doubtful a linux driver would ever ack the content flags in the first place.
This is exactly what Total Recorder does (in Windows) for audio. It adds itself as the primary audio driver for the system. All audio directs to it, then it forwards it to the driver for your sound card.
The upshot to this is that you can get Total Recorder to record this stream in just about any format at whatever rate you want.
I use this for the audio books I get from Audible. The books come down in some encrypted format that requires a special plugin for Windows media player or RealPlayer, or you can push them to an Audible enabled device (like the Rio).
Before going to bed, I start the book playing in media player with total recorder saving it out as an mp3 as it goes. The next morning I convert the mp3 to wav and burn it to a standard audio cd.
This type of circumvention is very easy as long as the stream has to be decrypted somewhere on the motherboard. Having the stream sent encoded to the card and having it decrypt it is another matter. I'm sure that someone will come up with a way to decrypt it.
load "linux",8,1
That is really just a minor problem. True, the actual framebuffer portion that displays your desktop is not going to have anything but a black/blue/purple/whatevercolor box but still all modern videocards have over 16megs of ram and some of that is going to be used for video overlay buffer.
Before I continue I should say that I programmed most of the v4l-drivers of voodoo 3500 tv so I do know what I'm talking about.
Registers in that videocard are going to point exactly to where that buffer is located and accessing it is no different than just mapping that portion into your address space and copying the data from there. There's going to be a lot of data but technically it is possible. And, this way you'll actually get the clean data instead of something that has already been stretched/filtered/de-interlaced/etc by your graphics card..
To summarize.. If you can see it, it resides somewhere in memory. If you can hear it, it resides somewhere in memory. It might not stay in one place for very long but definetly long enough for someone with intermediate hw-programming skills to capture.