[mythtv-users] Slightly OT? MPEG-2 Encoding Quality

[Rod Smith]

On Friday 09 March 2007 11:10, Greg Kettmann wrote:
> I'm still pondering the many choices for building a MythTV system.  My
> first choice for capture card was a PVR-150 but they're no longer in any
> local stores and seem to be being phased out.

I don't know if they're being phased out, but I'm sure you can still find 
them. I saw a bunch in a local (Massachusetts) CompUSA a week or two ago. 
OTOH, around Christmas there was a problem being reported: Hauppauge was 
putting another card in PVR-150 boxes because of supply problems. The 
replacement card had no Linux drivers, causing problems for those buying the 
cards for use under Linux and making a PVR-150 purchase a bit of a gamble.

> The PVR-350 seems like 
> another good choice and it's got the added advantage of having a Video
> (S-Video and Composite) out capability.

Word is that the PVR-350's output is on the endangered species list; 
apparently MythTV support for it will be dropped sooner or later. Thus, I 
wouldn't buy one of these now if that output is an important feature for you.

> Is there any difference in the quality of the MPEG-2 encoders?

I'm afraid I don't know the answer to this question.

> One of the side functions of the MythTV box will be to take my old tapes
> (family / camcorder stuff) and convert them to DVD.  There was a thread,
> a week or so ago, discussing software codecs for better quality.  As a
> general rule, are there differences in the MPEG-2 encoders in the cards
> (IE a more premium card gives better output = the 350 over the 150 or
> 250)?  Are there noticeable differences?  How about using a software
> codec versus hardware?

If you encode with a frame grabber, MythTV only supports saving to MPEG-4 or 
RTJpeg. Thus, if you use MythTV to do the capturing, you'll have to transcode 
from one of these formats to MPEG-2 (at least, assuming you want a standard 
DVD). If you capture with a hardware MPEG-2 card, you can save straight into 
MPEG-2 format, which will save you some time, if nothing else.

OTOH, if you transcode (either from MPEG-4 or a high-bitrate MPEG-2), you can 
use a multi-pass transcoding step, which in theory should produce a higher 
quality final product than you'd get with a single-pass encoding step. I 
can't say I've done any careful comparisons, though, so I don't know how 
significant an improvement you'd see, or even if there'd be an improvement at 
all. If there were to be any improvement, you'd need to do your initial 
encoding at a significantly higher bitrate (and/or in a better encoding 
format, such as MPEG-4) than you'll use on the final DVD.

> I don't know if it matters but input will be either composite, through a
> VCR (for a few) or S-Video from a Sony DV camcorder (both Hi-8 and DV).

I'm no expert on camcorders, but if you're using a digital camcorder, I'd look 
for options to get the video onto the computer directly in digital format 
rather than converting digital to analog and then encoding the analog back to 
digital.

-- 
Rod Smith
http://www.rodsbooks.com