AMD Processors
Decrease font size
Increase font size
Topic Title: Video Capture problems with AMD-768 and 2800+ MP A
Topic Summary:
Created On: 10/08/2003 02:25 PM
Status: Read Only
Linear : Threading : Single : Branch
Search Topic Search Topic
Topic Tools Topic Tools
View similar topics View similar topics
View topic in raw text format. Print this topic.
 10/08/2003 02:25 PM
User is offline View Users Profile Print this message

Author Icon
DavidRedwine
Junior Member

Posts: 5
Joined: 10/08/2003

I built a dual cpu Athlon 2800+ MP machine using Gigabyte GA-7DPXDW-P motherboard.

I spent $3,000.00 for this machine including the DVD recorders and flat screen monitor and pro level ViewCast OSPREY 220 video capture card.

. NOW I FIND THAT THE AMD-768 SouthBridge Chipset has a flaw and gets jitter and dropout on the capture because of the AMD-768 Chipset.

and that the AMD-766 chipset motherboards DON'T have the flaw!@!@#$@#

So here I sit stuck with a $3000.00 Video editing machine that I can't use !!!

Here is the WARNING (FAQ) on ViewCast's (makers of the OSPREY 220 card) website.

Question--I am trying to capture video with my Osprey card and I am seeing glitches/artifacts in the captured video files. What could be causing this ?

Answer--We have found that certain motherboards that utilize the AMD-768 peripheral bus controller do not provide adequate bandwidth requirements to allow the Osprey cards to capture successfully (one motherboard reported to use the AMD-768 controller is the TYAN Tiger MPX (S2466N-4M). We do not have a resolution for this problem and strongly recommend using a motherboard with a different Peripheral bus controller. If you would like to use an AMD based board we would recommend utilizing the AMD760MP chipset platform, which utilizes the AMD-766 peripheral bus controller.


Does anyone have any suggestions??????
David Redwine
david.redwine@greatacts.com / david@jamesonmanagement.com
 10/09/2003 10:07 AM
User is offline View Users Profile Print this message

Author Icon
TNZ
Junior Member

Posts: 7
Joined: 10/09/2003

Did you try with a firewire acquisition card ?

On my Tyan s2469 with a noame firewire card, it works fine ... DV capture needs a bandwith of 200 Mo/minute
 10/10/2003 10:39 AM
User is offline View Users Profile Print this message

Author Icon
DavidRedwine
Junior Member

Posts: 5
Joined: 10/08/2003

hmmm, Were you actually performing "encoded capture" to windows media or to Real .rm files with the firewire card and there are no static glithces in the picture quality?

I can do that for my newer content, but the older content is VHS and I can only use the composite input on the Osprey.......I don't want to junk the Osprey for a Pinnacle or other card because I am afraid that won't get me past the buggy AMD768 problems. Does anyone know of a bios or firmware fix for the AMD768 problems with video capture?


My next move is to try to do straight capture with no encoding to see if I can make that a work-around. Maybe this sorry AMD768 chipset will not hamper the conversion style encoding since the PCI bus wont be involved (using the promise 20276 raid with striping on 2 120 gig drives.

dr
 10/10/2003 02:01 PM
User is offline View Users Profile Print this message

Author Icon
smithers
Junior Member

Posts: 4
Joined: 10/10/2003

Dang David! That is bad luck on the Osprey stuff!

I know that AMD has had some issues with the MPX chipset (versus the MP chipset which rocks for Video and Audio solutions.) They even issued a white paper on the bandwidth limitations of the PCI slots.

As far as I know, there are no driver updates to help you out with this. That being the bad news, other video companies like Canopus got around some of the limitations by moving up their cards right next to the AGP slot. This ensured that the Video Compression card got the maximum bandwidth out of the PCI bus and did not cross from the "First" PCI slots to the "Second" set of slots.

Some Video HW manufacturers have noticed this bandwidth limitations and some have not. You have a really high-end card that is going to maximize the full use of the PCI bus.

Perhaps trying this might be helpful to you for the maximum results. The hardware limitations in the MPX chipset are not helpful. The positive news for others is that all the video companies really GROK the Opteron for Video. Boxx Technologies is making a fortune off these machines!

Hope that this helps,

-------------------------
(Insert witty signature/quote here...)
 10/10/2003 02:16 PM
User is offline View Users Profile Print this message

Author Icon
DavidRedwine
Junior Member

Posts: 5
Joined: 10/08/2003

Sounds like more of the rumor mill to me. When I disable the audio codec on the Osprey, the glitches in the captured video go away. How do you explain the tiny variation in bandwidth (< 1% bandwidth of the PCI required for the sound) used by audio capture affecting the video file. my CPU utilization is 2-5%, and no matter how crude and slow I set the frame rate, or the resolution, ANY use or enabling of the audio codec causes glitches in the video.

Seems to me like there is a lot of "heads buried in the sand" claiming these glitches are somebody elses problem. How can that be when you lower the capture requirements down to the point that even an old Intel 486dx2 could handle the capture, YET a smoking hot 2 cpu machine 20 times as powerful gets glitches in the file AFTER it is written, using 7200 spin drives and striped raid and AGP video without ANY OTHER PCI CARDS IN THE MACHINE?

Thanks for the response, but I still say BALONEY. It is the Osprey 220 that has a problem with it's interrupt handling or DMA or some interaction. It's as plain as the nose on my face.....so to speak....... Disable the audio, voila no glitches. I guess I'lll be shopping for a different manufacturers product.

dr
 10/10/2003 02:45 PM
User is offline View Users Profile Print this message

Author Icon
smithers
Junior Member

Posts: 4
Joined: 10/10/2003

Hmmm, I guess that I can't explain those tiny variations. The captured video/audio should be set on your system and playback correctly. I will assume that you have the most current drivers for the Osprey, so that can't be it.

Are you using the onboard MB sound or the audio from the Osprey? You said that there were no other PCI cards in the system. Could something be conflicting in that area? All other Audio formats play correctly (WAV, MP3, CD-Audio)?

Also for another idea out of left field, how about manually setting the IRQs for the Osprey card? Any hints from ViewCast on this?

-------------------------
(Insert witty signature/quote here...)
 10/10/2003 03:57 PM
User is offline View Users Profile Print this message

Author Icon
DavidRedwine
Junior Member

Posts: 5
Joined: 10/08/2003

I disabled the on board AC97 sound and the game port.......still, If I try to capture audio (osprey) and video (osprey) I get corruption.......cpu utilization hovers at about 1 percent. The interesting thing is that I CAN use the "AMCAP" capture application successfully with Osprey sound and video input. the resulting avi files do not have corruption. If I use Osprey's vidcap32 or real producer, or Helix or anything that pertains to encoding I get corruption in the screen as I view the video and further corruption in the captured video (in the file).

Being able to capture the video and audio natively from the Osprey card indicates that sufficient bandwidth is available, and something else is wrong........any hints on Bios tweaks for the motherboard that uses AMD 760MPX and 768 Southbridge chipsets?
 10/13/2003 04:53 PM
User is offline View Users Profile Print this message

Author Icon
smithers
Junior Member

Posts: 4
Joined: 10/10/2003

Sorry David, I was gone to do Real Life for the weekend...

As for specific BIOS and MB settings, I have only heard what others have done to help certain MPX situations out. I know that some of the video folks have disabled Auto IRQ settings and set up each board manually. Certain memory allocation settings have been tweaked and so forth. I can't comment on this completely, because each motherboard is different.

Pinnacle Systems has made a cookbook on what they set everything on the Tyan Thunder boards for their Targa 3000 board. This cookbook was pretty comprehensive on what each board they tested should be set to. Perhaps this can help:

http://www.pinnaclesys.com/sup...play...ubDocTypesID=74' ">http://www.pinnaclesys.com/sup...D=404&SubDocTypesID=74

I just don't have any specifics on each board (set AGP apature size to ..., set Memory CAZ and RAZ to ...) Sorry.

Again, I gotta ask, does ViewCast have any reccommendations?

-------------------------
(Insert witty signature/quote here...)
 11/09/2003 11:41 PM
User is offline View Users Profile Print this message

Author Icon
Richard H
Junior Member

Posts: 10
Joined: 10/07/2003

Rumor?? Come on AMD released a white paper on the problem and you call it a rumor
 12/16/2003 03:07 PM
User is offline View Users Profile Print this message

Author Icon
pcy
Senior Member

Posts: 2029
Joined: 10/18/2003

Hi,

This is NOT a PCI bus Bandwidth problem.

Its a timing problem on the PCI BUS - latency not total bandwidth.

See my description of the basic problem (which affects ALL video capture/editing cards in principal) in this thread:
http://forums.amd.com/index.php?showtopic=3687' ">http://forums.amd.com/index.php?showtopic=3687


Or go to:
http://www.dc1000.org' ">http://www.dc1000.org

and read the articles by TAD about the dc1000 in the fft hand Nav bar.



Peter
Statistics
112018 users are registered to the AMD Processors forum.
There are currently 0 users logged in.

FuseTalk Hosting Executive Plan v3.2 - © 1999-2014 FuseTalk Inc. All rights reserved.



Contact AMD Terms and Conditions ©2007 Advanced Micro Devices, Inc. Privacy Trademark information