<br><br><div><span class="gmail_quote">On 12/9/05, <b class="gmail_sendername">John P Poet</b> <<a href="mailto:jppoet@gmail.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">jppoet@gmail.com</a>
> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
On 12/9/05, mrwester <<a href="mailto:mrwester@gmail.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">mrwester@gmail.com</a>> wrote:<br>> Hi-<br>> My mythtv system has 1xPVR-350, 1xPVR-500 using ivtv
0.4, and a newly added<br>> DViCO FusionHDTV 5 Lite using DVB drivers. 1x200GB seagate+1x300GB seagate
<br>> HD LVM with XFS for /video and 1x40GB seagate for FC4 OS. I have DMA<br>> enabled. DVI out via fx5200 at 1280x768. Chaintech nforce2 mobo, AMD2900<br>> with 1gb ram. All 4 tuners can capture on their own with no problems. I
<br>> can capture with all 3 Hauppauge tuners at the same time no problems.<br>> Trying to capture HDTV while capturing with the Hauppuge cards results in<br>> normal analog recordings but the digital recording has dropouts and
<br>> scrambled bars across it. I need to double check if capturing digital NTSC<br>> with the DViCO card also has the dropouts when one or more PVR cards are in<br>> use.<br>><br>> I'm new to the HDTV part of mythtv and looking for feedback to tell me if
<br>> what I'm trying to do is realistic or not. And if it is realistic, where<br>> should I look to troubleshoot. Thanks for any insight.<br>><br>> Mike<br><br>I have four HD-3000 cards writing to a 5 disk software RAID5 formatted
<br>as XFS. I can record on all four cards simultaneously without any<br>problem. I do *not* use LVM -- which I believe adds some overhead.<br>Software raid5 on a 3GHz Pentium 4 system results in very good disk<br>throughput.
<br><br>As a test, I tried turning OFF hyperthreading on my 3.0GHz P4, and<br>started having problems recording more than two HD programs at a time.<br> Don't know how Athlons perform, but HT is a win for Pentiums.<br><br>
Another datapoint - My raid5 had to rebuild one of the disks. During<br>this rebuild, I could still record one HD stream without a problem,<br>but two resulted in a LOT of corruption, and a lot of messages about<br>it taking a LONG time to get data from the HD tuner card.
<br><br>I noticed that Daniel add a ringbuffer into myth's DVB processing<br>today, so that should help any system which is marginal.<br><br>Don't know if any of this helps you, but thought it might...</blockquote><div><br>
<br>
Thanks for all the replies, they give me some idea of what is possible- if capture of 4 HD streams is possible, then I'd think
the 17gb/hr that the 3xhauppauge + the DViCO are writing to disk
when recording 1080i should be pretty easily attainable.
<br>
<br>
Currently, I've got the 2 tuners of the 500 as 1 and 2, the pvr-350 as
3 and the fusionHDTV 5 lite as #4. I only see issues when the
pvr-350 and the fusionHDTV 5 lite are both capturing... If I have
all 4 tuners capturing- which will give me recording corruption on the
fusionHDTV5 lite, and then systematically stop recording on any of the
other tuners, only the PVR-350 makes a difference, and stopping capture
by this tuner will clear up the fusionHDTV capture, irregardless of what
the pvr-500 is doing. <br>
<br>
Relevant bits of lspci -v indicate that the 350 and the fusionHDTV 5 lite are not sharing and IRQ, so that's not an issue... <br>
<br>
<br>
01:0a.0 Multimedia video controller: Brooktree Corporation Bt878 Video Capture (rev 11)<br>
Subsystem: DViCO Corporation DViCO FusionHDTV5 Lite<br>
Flags: bus master, medium devsel, latency 32, IRQ 11<br>
Memory at e4000000 (32-bit, prefetchable) [size=4K]<br>
Capabilities: [44] Vital Product Data<br>
Capabilities: [4c] Power Management version 2<br>
<br>
01:0a.1 Multimedia controller: Brooktree Corporation Bt878 Audio Capture (rev 11)<br>
Subsystem: DViCO Corporation DViCO FusionHDTV5 Lite<br>
Flags: bus master, medium devsel, latency 32, IRQ 11<br>
Memory at e4001000 (32-bit, prefetchable) [size=4K]<br>
Capabilities: [44] Vital Product Data<br>
Capabilities: [4c] Power Management version 2<br>
<br>
01:0b.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL-8139/8139C/8139C+ (rev 10)<br>
Subsystem: Realtek Semiconductor Co., Ltd. RT8139<br>
Flags: bus master, medium devsel, latency 32, IRQ 5<br>
I/O ports at d000 [size=256]<br>
Memory at ea010000 (32-bit, non-prefetchable) [size=256]<br>
[virtual] Expansion ROM at e4010000 [disabled] [size=64K]<br>
Capabilities: [50] Power Management version 2<br>
<br>
02:08.0 Multimedia video controller: Internext Compression Inc iTVC16 (CX23416) MPEG-2 Encoder (rev 01)<br>
Subsystem: Hauppauge computer works Inc.: Unknown device e807<br>
Flags: bus master, medium devsel, latency 64, IRQ 9<br>
Memory at dc000000 (32-bit, prefetchable) [size=64M]<br>
Capabilities: [44] Power Management version 2<br>
<br>
02:09.0 Multimedia video controller: Internext Compression Inc iTVC16 (CX23416) MPEG-2 Encoder (rev 01)<br>
Subsystem: Hauppauge computer works Inc.: Unknown device e817<br>
Flags: bus master, medium devsel, latency 64, IRQ 11<br>
Memory at e0000000 (32-bit, prefetchable) [size=64M]<br>
Capabilities: [44] Power Management version 2<br>
<br>
If I get a chance to switch some cards/slots around tomorrow, I'll see
if it makes any difference, but from lspci, it appears that the PVR-350
is sharing with tuner 1 of the PVR-500 and the fusionHDTV 5 lite is
sharing with tuner 2 of the PVR-500. I've got to re-do the
experiment and go to the logs to see what's going on, but my wife
wanted to catch up on a few days of unwatched shows, so the system was
in use all evening... If anyone has ideas, I'm open to
suggestions. Thanks,<br>
<br>
Mike<br>
<br>
</div></div><br>