<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-8859-1">
<META content="MSHTML 6.00.2800.1276" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT face=Arial size=2>Rolando wrote:</FONT></DIV>
<DIV><FONT face=Arial size=2>>Thanks to everybody that replied. I would like
to use the Pinnacle and take <BR>>advantage of the hardware encoding of the
card but since it generates too <BR>>much data (between 60-120K per frame)
and my hard drives seem to be <BR>>throttling the system (profiling revealed
communication between backend and <BR>>frontend seems to top out at
2.5MB/s*), I'll have to use another capture <BR>>card. Hopefully the PVR-250
I ordered recently helps in the data flow area <BR>>(6000Kbit/s for
DVD-quality video should be less than 1MB/s between backend <BR>>and
frontend). <BR>><BR>>Again, thanks for all the replies.
<BR>><BR>>Rolando. <BR>><BR>>* With the frontend "RemoteFile" class
receiving 64k blocks of data from the <BR>>back end in about 25ms. Most of
the delay was caused by the backend reading <BR>>the data off disk. Maybe I
should hack Myth a bit and amortize the read time <BR>>per 64k block using
bigger block reads? hmmm... <BR></FONT></DIV>
<DIV><FONT face=Arial size=2> This has already been done in CVS. Either
building from CVS or waiting for .14 will help this. Block sizes now are either
128k, or 256k for higher bit rates. Before the changes I had prebuffering pauses
at RTJpeg, 480x480, quality 170. Post changes I have no pauses at RTJpeg,
640x480, quality 210.</FONT></DIV>
<DIV><FONT face=Arial size=2></FONT> </DIV>
<DIV><FONT face=Arial size=2>-Mark</DIV></FONT></BODY></HTML>