Am I being IRQd off.

roy wood · Post by **roy wood** » Thu Aug 10, 2006 11:46 pm

Hi all I can capture full frame DV-AVI with ease using 1394 but when I try to capture full frame analogue-Avi with my Pinnacle DC10 PCI card it drops frames badly. If I use compression Mpg2 full frame it's happy if I lower the frame size the same.
This card is rated for 720x576 capture and I've got a 2.80Ghz P4 800FSB 1Gb DDR Ram 80Gb 7200 3-Partition Hdd so it should be plenty powerful enough.
Could it be the bloody IRQ/s again the capture card is on the same one as the Usb Host Controller and the Ultra-Ata Storage Controller.
I thought I'd read somewhere that IRQ/s weren't the same problem with XP as they were in W9xxx or is that just wishfull thinking.

Terry Stetler · Post by **Terry Stetler** » Fri Aug 11, 2006 12:52 am

Most modern computers don't used the old PIC IRQ system.

Back then you had hard/fast IRQ assignments under PIC, which only supported interrupt lines from 0 to 15. Today most all systems use APIC, and it supports 255 interrupt lines per APIC. Most times only 24 interrupt lines are actually used, but still those IRQ numbers you see in Device Manager don't mean as much as in the bad old days.

Example: one of my rigs has the following devices all on IRQ22;

MB embedded IEEE-1394 controller
Matrox DV/MPEG Codec Driver (RT.X100 realtime card)
Matrox RT.X100 Flex3D Accelerator (RT.X100)
Matrox Video Bus (RT.X100)
Texas Instruments OHCI Compliant IEEE-1394 Host Controller (RT.X100)
Standard Dual Channel PCI IDE Controller
Promise SX4000 RAID5

and nobody steps on each others feet.

What could be is that the DC-10's an old card that was originally designed for use under Video for Windows system, which is supported but not the WDM/DirectShow system WinNT/2K/XP are geared to use. My experience is that while WDM drivers may come out for these outdated boards they are so hard to program for that they work as pale shadows of themselves compared to the old days.

Another problem could be that you're using embedded audio. If the DC-10 doesn't have its own audio I/O and uses the mainboards that could well cause problems. Most embedded audio isn't fully hardware, a lot of it runs in software which can tie up the CPU during captures. Been there, done that. Full hardware is better.

roy wood · Post by **roy wood** » Fri Aug 11, 2006 1:53 pm

Hi Terry thanks for that at least I can put my IRQ Paranoia back in the closet. You've given me a couple of areas I can explore. The Audio I can unplug and disable in 'options' to test that and I'll post the VFW / DSWplug-in question on the Pinnacle site to see if anyone has a work around.
Regards Roy

tyamada · Post by **tyamada** » Fri Aug 11, 2006 4:37 pm

I'm not sure what drivers you are using for you DC10 card but Piinnacle has a PCI performance enhancement utility for their cards. This link is for the DV500 card but the patch was posted in the DC-10 section years ago.
You coulld try it to see if it fixes your problem
http://www.pinnaclesys.com/PublicSite/u ... =documents

roy wood · Post by **roy wood** » Sat Aug 12, 2006 8:05 pm

Thanks tyamada mine is a late version which has the performance enhancer included. I'm coming to the conclusion I have a conflict or resource problem somewhere that only effects my ANALOGUE CAPTURE. Having shut down as much as I can and de-fragged I am NOW able to capture Full frame Avi without dropping frames for about 20 seconds before they start dropping out.

I've taken these readings from 'System Information'
Total physical Memory 1,024.00 MB
Available Physical Memory 698.46 MB
Total Virtual Memory 2.00 GB
Available Virtual Memory 1.96 GB
Page File Space 2.41 GB

Perhaps someone who understands these readings will look them over to see if they seem ok. Thanks Roy.

Terry Stetler · Post by **Terry Stetler** » Sat Aug 12, 2006 8:10 pm

They look fine, and I'm presuming that by "Having shut down as much as I can" you mean you've followed Steve's tutorial about setting up an editing profile?

http://phpbb.ulead.com.tw/EN/viewtopic.php?t=13950

Also: are you capturing to a drive that's not the same physical device as what Windows is installed to? Ideally this should be on the secondary IDE master channel, SATA or its own PCI/PCIe/RAID controller. The primary slave IDE and USB are no-no's.

roy wood · Post by **roy wood** » Sat Aug 12, 2006 9:28 pm

Hi Terry yes I've used the Tutorial. At the moment I only have the one Hdd 80Gb 3 partitions ( C:20Gb - E: 27Gb Video Files - F:27Gb Video Projects - 6GB Spare ) I will be adding extra Hdds shortly but what I can't understand is why I can capture DV-AVI Full frame even with other progs running and yet I'm seemingly running out of resources within seconds with AVI.

Post by **Ken Berry** » Sat Aug 12, 2006 10:31 pm

If you are trying to capture full, uncompressed AVI to a 27GB partition, you are going to run into problems sooner rather than later, simply because that flavour of AVI is just so HUGE. It requires 65GB of HDD for just one hour. So if you have anything else on that partition, then the program probably just can't find HDD space, especially if your working folder is also on that same partition...

If you intend capturing full uncompressed AVI, then getting a much bigger extra HDD is not just a luxury but a critical necessity...

roy wood · Post by **roy wood** » Sat Aug 12, 2006 11:29 pm

Hi Ken thanks for your input I only want to capture a few minutes of Avi at a time to edit, clean up and use filters such as 'Anti-Shake' which don't seem to work with compressed files.
I then render the Edited version to Mpeg2 and file in another Partition for later inclusion as a chapter in a full project. The original AVI is then deleted.
Is there any difference in file size between Analogue AVI and DV-AVI please because I'd previously believed Avi as 13Gb per hour.

Post by **Ken Berry** » Sat Aug 12, 2006 11:39 pm

DV is a special version of AVI which has a compression factor of about 5, but is to all intents and purposes lossless (i.e. it can be re-rendered without loss in quality). It is DV which takes up about 13GB per hour. But full, uncompressed (what you call analogue) AVI takes around 65GB. (Hence my mention of the compression factor of DV: 5 x 13GB = 65 GB.)

roy wood · Post by **roy wood** » Sun Aug 13, 2006 12:18 am

Ok I have an Analogue Camcorder and a Mini DV Camcorder if I capture 1 Hour Full Frame 720x576 Pal from each one are you saying that the file from the first will be 65GB and the second will be 13GB and that is because there is some compression in Type 1 DV. Please.

roy wood · Post by **roy wood** » Sun Aug 13, 2006 1:03 am

Come back Ken the Lights are coming on when I use Ulead to capture analogue there are 3 options YUY2. RGB24. RGB16 I had assumed that this was uncompressed AVI but it's not is it. So what I need now is to find an Analogue Codec equivalent to type 1 DV that will allow me to capture without dropped frames and use the filters. Any Suggestions please.

Post by **Ken Berry** » Sun Aug 13, 2006 1:36 am

One thing you might look at is whether your mini DV digital camera will allow pass-through i.e. where you can connect the analogue camera to the digital camera's AV inut, then have the digital camera connected to your computer via firewire. If it works, your analogue video is transformed to digital DV-AVI format (at 13 GB an hour) in the digital camera. Unfortunately, though, not all digital video cameras allow pass through.

Terry Stetler will probably jump in on the codec question. I confess I don't know much about it, but I suspect he would recommend you use the MJPEG codec. It captures files a bit larger than DV files, but is of similar (high) quality and largely lossless. (In fact, from the little I know, it was the predecessor of DV format.)

But I don't think you can get a free one. He recommends one from a company called Morgan. Go to http://morgan-multimedia.com/ Terry recommends the V3 one, and I think it costs US$15. He says it is super fast and quite configurable.

Terry Stetler · Post by **Terry Stetler** » Sun Aug 13, 2006 3:42 am

Morgan V2 is $15 and V3 is just $20, so I'd go for V3 if at all possible. That said both are very, very good. I've used them to compress YUV captures since the Celeron 850 days, so the CPU requirements are definitely not excessive. The MainConcept MJPeg is no slouch either and it's similarly priced.

Capturing to the same physical device as the operating system, no matter if it's partitioned or not, is a no-no because any time the OS decides to access the HDD you'll drop frames all over the place. Same goes for a drive on the Primary IDE/slave connector.

The best places for a capture drive are the Secondary IDE (master is best, but slave is OK if the CD/DVD goes unused), SATA or a PCI/PCIe controller. IEEE-1394 external is OK, but stay away from capturing to USB drives.

As for the relative bitrates:

Uncompressed video talkes up a TON of space and requires very fast HDD's or RAID arryays to capture. Bitrates can run 10-12 megs/second for YUV and up to ~30 megs/second for RGB24 in large frame sizes. If this is what you've been capturing it's no suprise you're dropping frames left and right. For this kind of capture I almost always use a RAID array (several drives 'striped' together to become one much faster drive).

DV25 (consumer DV) is 3.6 megs/second. Most any drive can achieve this, but its limited color resolution can affect how well composites (adding video overlays) come out. All too often you'll see bleeding on the right side of the overlay due to DV's color samples being spread across 4 horizontal pixels.

MJPeg is adjustable, ranging from 1 meg/second to 10-12 megs/second. 5-8 is considered the 'butter zone' where lossiness is minimized and drive performance does not have to be extreme. It has twice the color resolution of DV, making overlays 'neater'. Due to fast software codecs like Morgan/MainConcept this can be used even by small systems to compress YUV or RGB24 on the fly (YUV is better).

Yes, it was the editing format of choice before DV came out. For many with quailty on their mind it still is. I'm not above capturing DV shot footage from the cams S-Video output to high bitrate MJPeg if it's to have a lot of effects added.

MPEG is a poor editing format because of its use of calculated frames instead of real bitmaps for the vast majority of its frames, making it a very lossy format in edits. Editing MPEG is best done by converting it to another more edit-able format first. This can be MJPeg or a YUV format like HuffYUV (freebie; do a google for ver. 2.20) or the Quicktime/Component codec. The latter two are essentially uncompressed at 10-12 megs/second.

roy wood · Post by **roy wood** » Sun Aug 13, 2006 11:36 am

Massive thanks for your Patient and Detailed replies Terry & Ken I was completely off-track with my perception of this whole compression issue. However with your information I shall now be able to get the Max from my equiptment. Regards Roy.