Michael T. Dean
mtdean at thirdcontact.com
Wed Aug 2 01:53:44 UTC 2006
On 08/01/2006 08:33 PM, Brad Templeton wrote:
>On Tue, Aug 01, 2006 at 04:47:25PM -0700, Jafa wrote:
>>To MythTV it is a capture device (actual two as it has dual tuners).
>>Live TV loops through the backend then onto the front end.
>>The CPU requirements for the backend are very low (I see around 3% on a
>>cheap 2GHz Sempron).
>We do have to stop calling ATSC/QAM tuner capture devices. They
>are demodulators of a digital signal rather than frame capturers.
But don't they need a _tuner_ so the demodulator can find the carrier
wave from which to recover the information content for _capture_? Isn't
it more correct to call an ATSC/QAM HDTV card a tuner capture device
than to call a device (i.e. a PVR-x50 or a BTTV card) using the
S-Video/Composite inputs a tuner?
>However, what was your thinking on placing one or two NTSC capture/compressor
>chips in the box, or in a variant of the box? Then it would be able to
>be all a user needs for Myth. How much are those chips?
>You could, in theory also do just plain capture without compression, since
>HD people all have powerful CPUs, but that's not quite as good because
>you get issues when people want to watch HD and compress 1 or 2 SD at the same
Heh. 720x480 @ 30 frames/sec (60 fields/sec) takes 15MiB/sec (yes,
that's a capital B as in 15,552,000 bytes/sec) at 12-bit color (and no,
that's not 12 bits/color/pixel, that's 4 bits/color/pixel). Therefore,
uncompressed SDTV takes 124416000 bits/sec = 124Mbps. Yes, that's more
than 6 times the maximal bitrate (i.e. 19.39Mbps) of (most) HDTV
(ignoring the 38.78Mbps of high-data-rate mode, that you won't see for
terrestrial broadcast). So, really for transmission across the network,
"even" SDTV must be compressed. (Imagine that MPEG-2 can get that same
124Mbps down to 2.2Mbps with pretty good quality and 4.4Mbps with good
quality. Now that's compression.)
More information about the mythtv-users