Using a TV tuner as a high speed ADC

tv tuner

The Bt878 chipset is fairly common on TV tuner cards. The chip has a built in analog to digital converter with a sample rate of 119kHz to 448kHz, well above the standard audio rate of 44kHz. The hardware has to be hacked a little to inject your signal since the chip is usually receiving audio from the turner. With some driver hacking this chip can be pushed to 896000 samples per second. Recent developments make things even easier with ALSA support.

[thanks rockarolla]

33 thoughts on “Using a TV tuner as a high speed ADC

  1. This is somewhat useful for simple gnuradio stuff. There’s actually an effort on the gnuradio list to make the main video adc useable for gnuradio stuff. They’re running into the problem of the video adc dropping data for the duration of the vblank I believe. There are undocumented instructions for the controller that people are looking into though. It’ll be sweet when/if they figure it out or someone releases a cheap software card that doesn’t have the same limitation. Especially since the tuner can on the card could be used just to get the signal you want down to IF, then decode the signal in software. Hello poor man’s usrp!

  2. would it be possible to do this hack and use some type of ‘scope software in windows without too much effort?

    It would be decently cool to have a .5khz scope for $10… But even sweeter if we hacked into the video ADC for a scope good to a few MHz… for $10.

    finally a real hack, and one for a device I already have!

  3. guest / #7,#12,#13… a digital laptop keyboard? did you mean a keyboard as in piano-style musical keyboard, or keyboard as in QWERTY.. either way, no this will not convert that way, it is an ADC, that is, analog->digital converter, what you want is a DAC if you want to turn digital into analog :) hope that helps.

  4. If it’s a delta-encoded ADC, it’ll have a DAC built in, but it might be difficult to access. Here’s to the oscilloscope project, btw. i’m thinking the anti-caps filter is to stop folks from yelling and make for a less formal atmosphere, but who knows?

  5. #9, #10 :

    The technical reason why is the presence of “text-transform: lowercase;” for BODY in hackaday’s CSS.

    View the source and you’ll see that the comment text still has the original capitalisation, the CSS just makes your browser lower-case everything.

    If it really bothers you then make a copy of the CSS with that taken out and get your browser to use it :)

  6. The highest end audio interfaces go up to sample rates as high as 192 khz. However, there’s really not much use for this. It’s just a gimmick, TRUST ME. 96 khz is the best any human can perceive. The highest frequency a person can hear with the best hearing is about 22 khz. The reason CDs are 44.1 is because it needs to fold down frequencies of two occurances of harmonics. The reason that we have even higher sampling rates is to solve the issues of even more, higher frequencies occurring at once time. For an example, 4 French horns. They all have really high frequencies in this respect it would be useful to use a sampling rate of 48, 88.1, or 96, but 192 is not even really preceivable. There’s nothing useful for going any higher… unless you are going to down-sample something a ton.

  7. Actually, erik, I have read different about he reasoning for 44.1k sampling on cds, says that the reason is because it works with both ntsc and pal tv. also the sampling rate is like resolution in a digital image. where did you get the information the 96k is the highest anyone can percive? at high frequencies 22k, 96 is only 4.3 samples per cycle. i dont really know but that seems kinda low.

  8. To the person who said this would be better adapted to an O-Scope…

    True, but you have to remember this could be better utilized in a recording scenario… If your mixer is capable of whatever bitrate, your masters will be at the same level. This allows for future-proof recording. 192kHz is about the best you can do now, but in the future, DIYers will have this rate at their will.

  9. erik jackson:
    True, the absolute highest frequency any human can hear is near 30kHz, but who’s to say that mid level frequencies can’t be helped by a higher resolution? I understand the basics of sampling rate and the human ear, but I fail to understand why people write the extended resolution off. If you’ve never heard 192kHz, then how can you say it’s no better? Even though you’ll probably pretend you’re a producer and say you’ve heard it, have you ever given it a real test? Straight from your newfangled soundcard? I didn’t think so. The future is always ahead of us, hence, new technology is as well.

  10. According to Nyquist theorem the sample frequency used must be a minimum of twice that which the human ear can hear hence the reason that Cd’s are sampled at 44.1 kHz just over twice that of human hearing to allow for implementation of a practical anti-aliasing filter. sampling any higher than that makes the sound ‘smother’ and allows for more accurate reproduction of mid range frequencies, harmonics and the likes.

  11. I am investigating about ways to hack and use TV PC cards or even satellite PC cards as ADC to use applications as win-oscilloscope though breaking the sound card frequency limits as many actual signals to test work at high ranges that you cant work properlly with the sound card range.
    Links or any information would be great.

    BTW: there out are some USB oscilloscope devices at 200MS/s approx though thats the fast and costly way to get it solved.

  12. I use this for recording off of low-frequency radio. I stayed up late a few nights writing a command line program to shift a band of signals off the high frequency recording down to a standard audio sample rate, and now can hear radio stations. It’s good that there’s still LORAN-C in my area. I can also hear the sound of the radio station WWVB which is used to set radio controlled clocks.

  13. Hi!

    I have a pctv pro with bt848 chipset.
    That one (and i think the bt878 also) have a 40Ms/s A/D for video sampling.
    I downloaded a program and source code for it. called vbi view or similar.
    Tomorrow i go and test first time with video imput, after that a 14MHz quartz oscillator.
    Anyone have another program for that?


  14. you could key the squelch on a vid cap by feeding the vertical and horizontal sync pulses threw some 50k photometers and the video threw a 1k poptentiometer and adjust it toill it givews you a steady frame to a vid cap device and make a 30fps loss-less video of the siginal and convert the video to fle or zero compression type video file and then analyse the data and convert it to graphic representation.. or hack into a scan converter and patch it to a udma 66 interface and read it into a file with debug .. maybe or something

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s