marvel skipping frame

[Patrick]

That's interesting. I notice that you have a SB Live sound card. I'd like to hear the Doc's comments on this! He has come out in the past with some rather critical comments directed towards PCI sound cards, especially this one.

[CMB]

This time last year i read the posts about G200 and SBLive probs and pooped my pants, lucky i timed getting my cards as they brought out the new drivers that fixed the probs i assume

i have read of other peoples probs, but really my sblive works great too :O)
its a crap out having to by the new driver upgrade though from creative, £9????????
since when have you had too BUY driver for a peice of hardware I hope creative stops this before others join in 2
its lucky i have a pal who just bought a Platinum cos you get all those new drivers with it, The mp3 coders pretty good too :O)

[This_Idiot]

Bruno,

It doesn't matter how fast your HD is, if you are using it for OS, Application AND AVI then you are going to get into trouble sooner or later. Not only are you trying to write (or read) ann uninterrupted stream of 3Mb/s, your application is also trying to cache to the same HD, and windows is trying to use the same HD for its memory management.

There is no guarantee that there isn't another factor that will be holding your system back, but this is about the biggest one that you will face.

A lot of us here have got over the initial problems a long time back, so we tend to look for answers within our current scope of expertise rather than sit back and take the wide view.

Your first task is to get a reasonable HS setup and make sure that DMA is enabled (I know that Grigory says that PIO4 is fine, but I have fallen foul of this argument myself, especially using non-intel chipsets)

[Bruno]

Thanks,
Any recommendations on what I should buy as a second drive?

[Grigory]

Chris,

To be precise, I was always talking about PIO4 DMA vs Ultra DMA drives.
Old drive may not support Ultra DMA. However, most of them that support PIO4 also can support "multiword DMA..." mode which is 16.6 MB/sec, but still DMA. So, by speed, it is PIO4, by functioning, it is DMA.

However, many members here reported playback problems on many fast drives, even if they were avi-only, defragmented, and all system optimizations were done.

The problem is with system-to RR incompatibility. For example, Abit boards show many problems, as you can see.

My permanent position is as follows:
the cards like RR_G are the last purchase for PC. It is impossible to built a system around RR_G, keeping in mind all these compatibility issues. Many of us could accept 1.5 bigger price of RR_G itself, but with better compatibility to escape wasting money for numerous upgrades of nearly all PC components.

Remember:
Fasttrack
Several fastest drives
probably MB change
sound card
Power supply ( for those who have double editing card detection or tuner not detected)
what else?

The cost of a system which (maybe?) start working well becomes too high for many of us. I made some of these steps, and, as you remember, asked many questions to get an explanations: why we need the most expensive upgrades or any upgrades to make inexpensive card working satisfactory?

In my case, I spent more money for upgrades then for RR_G.

I could resist to skip Fastrak purchase, but did a lot of tests to see that a problem is not with drive performance.

I was switching to all possible PIO modes, DMA modes, changing connections, soundcard, CPU speed, PCI slots, LX6 to BH6 motherboard.
The result always was as follows:
RR_G drops frames on playback. Capturing drops is another story, I don't want to discuss it here.
For any non (Matrox MJPEG) digital video with up to doubled data rate, the playback was always smooth.
I have now perfect operation of Raptor card, and my old good Avermedia TV phone as analog capture device. The last works well for 10 MB/sec uncompressed half width captures from VHS tapes. I use the cheapest in Moscow Fujitsu 10.2 and 17 G drives.

My position still remains: I could admit 250$ instead of 180$ I have paid for RR_G, provided it would work on my very common PC configuration.
I try to keep apart from RR_g discussion here, but sometimes I cannot resist from saying that recommendations for hard drives may cost a lot, but not work, because they are almost outdated for current state of drive performance.

Grigory

[Guci]

Hai

I understand that using OS on the same capture video is a bad idea. But what if you make partition? since I have 27 GB I could reserve 7 gb for my OS and 20 GB for capturing.

Any comments?

Thanks

[Pertti]

Guci,

Partitionning the drive doesn't help, since you are still using the same heads.

Pertti

[This_Idiot]

Hi Grigory.

I wasn't calling you out, I was saying that although you have had success with PIO4, I haven't.

I have only been running video capture equipment for a couple of years, using Matrox equipment from the RR-S, thru RR-G and Marvel. Starting with a seriously underpowered P133 and a single drive (where it IS possible to capture and edit successfully if you are realistic enough to limit your expectations), through AMD, Celeron and PII processors and a number of different mobos (mainly Aopen and Gigabytes). The AMD was a bitch to get set up, solely due to the difficulties in getting DMA working. Until I found the answer, I was dropping frames all over the place. Afterwards I didn't have a problem. Does that sound like a Matrox problem ?

And even on the current PII 233 that I am using as a testbed, choosing the wrong busmastering drivers results in desperate performance. Choosing the right ones fixes the problems.

I may not be a scientist, but I recognise the value of a "fair test" and apply it wherever possible. So when I say that dropping frames is USUALLY a function of a substandard HD system, and that DMA enabled drives perform better than PIO4, this isn't a "scientific" explanation, but one that I have observed. Ergo, I am not talking out of my backside.

Your first post to this thread shows that you jumped to a conclusion without determining the FACTS, and that your recommendations in this case are very probably WRONG. Capturing full-chat to a single drive is probably the biggest cause of frame drops our there.

Chris

[Marijn]

Hi Guci,

I might have a hint for you. You may want to dissable the virtual memory (swap file). I am using 64MB of ram and a few months ago when I didn't have a second drive (only a second partition) I always disabled the virtual memory.

I also had some drivers of my HP scanner that were checking the scanner every now and then. So before I capture, after a boot and before starting MSP, I use the CTRL-ALT-DEL combination to disable every task that is running except for the "Explorer" task. Disabling this one will hang your system.

Since you have 128MB of ram, you probably will have enough. This may sound a bit drastic, but this way you would be able to see whether it is actually windows that is interfering with your playback.

After capturing or print out to tape, you enable the virtual memory of course...

Good luck,

Marijn

[Guci]

Hai pertti

Thanks

[sweeper]

Guci,
I also had similar problems with my Marvel G200. It always skipped a few frames now and then when I captured and when I played back. Few days ago I started a new project with amount of capture, that does not fit to my 7 GB WD disk. I went out and bought Maxtor 15 GB 7200 rpm. I put it as Primary Slave, formated with Fdisk. The same system, the same tape, where I always got dropped frames just two hours before, and now, 10 GB of full resolution capture in a raw with PC-VCR from simple VHS-VCR, 1 frame dropped. I played back the first and the last of five 2 GB files, no dropped frames. My previous HD was split with fdisk into 2GB System and 5 GB Video and it did not help.
I have ASUS P2B, PII-350@392, 64 MB, SB Live, a year old Marvel with old drivers, W98.
BTW, check your OS, I saw many complaines about Matrox from people with betas, codidas, and nationaly adopted Windows.

[Keith Wiebe]

I gave up on the Rainbow Runner after 2 mobo's and 2 hd's. When running only 1 hd I had to disable virtual memory for smooth playback but with 2 I don't see any difference but I did have it set at 128mb min and max. I replace my Maxtor 8.4 gig 5200? drive with a Maxtor 10.2gig 7200 rpm and the only thing I gained was more reliable capture at full quality but I still wasn't able to capture some footage at full quality without droppnig 15 frames but my new Canopus Raptor did so without any problems at all! I'm wondering if the people with less skips upon playback aren't the PAL people as they only have 25fps to deal with vs NTSC 30 (or should I say 29.97fps). It's easier to deal with 25 than 30!

[GinoCyber]

Grigory, what kind of advice is that? I take it you know nothing about the BE6 boards that Guci has.

Allow me to educate you this time

-PCI slot 1 shares it's resources with the AGP slot.
- PCI slot 3 shares with the Ultra 66.
- PCI 4 and 5 share the same bus mastering lines.

Abit also says that they have major issues when USB is sharing with the Ultra 66.

Now that's a brilliant design if I ever saw one

As for your BH6, Abit doesn't even recommend putting cards in slots 1,4,and 5 if these cards need an irq.

Did you ever think that Guci might have a configuration problem or both a HD and a configuration problem?

Sheeeeeeesh

Besides, having the video stuff on a separate hard drive is never bad advice.

Guci says he has a 66 drive. What do you think would happen if he has another bus mastering card on slot 3 and or USB enabled?

Guci, you say you have a 66 drive. See if you can borrow a 33 drive and use the 33 connectors or even try another motherboard if you can. The Abit boards are certainly not the best out there.

Proper way to install a 66 drive on these mobo's is by putting the hard drive on the 33 port and have the utility from the hard drive manufacturer adjust the hard drive to 66 mode. Only then it will be properly used the next time when you put it on the 66 port.

PS: I think our forum referree will be needed soon

Ciao

[Grigory]

Hi,

"The video sometimes skip the frame like loosing many frame (freeze) eventhough the software codec report it no drop frame."

Maybe, my english is not as good, but this phrase inmmy understanding says that video can be captured without dropped frames, but cannot be played back without short stops.

This is a problem which I had for many months of attempts to make any single movie on tape with RR_G.

Trying to solve or explain a reason of problem, I tried to lower the drive performance as far as I could, in attempt to get dropped frames on captures. NO dropped frames with stable video signal, in PIO4 mode without DMA, too.
Still the same number of dropped frames on playback. Lowering the drive performance by a factor of 2-3 DID not made situation worse.
Chris, don't you think that in my case, which seems to be similar to the original problem, the trouble was outside the drive performance issues?

This, of course, has no relation to any science. My attempts to make RR functioning just included this pure diagnostic check.

Making large or small swap file, reducing cache size to 8 M, or enabling default Windows cache size management, did not alter the RR operation. Still good capture ability, and still dropped frames on playback. And, on any drive from three.

The only one movie I could made on tape was converted from MJPEG to DV and exported with Raptor. For all other movies I had to abandon the idea of tape export and used mpeg on CD.

I lost a lot of time and money too, in attempts to make RR_G playing MJPEG without glitches - no success.

Gino,

Thanks for my education. I've read this info before in BE6 manual and decided not to experiment with it.
I have 3 PCI slots in use now:
BT848 card takes one,
DV raptor another
Hollywood+ the third.

AGP is occupied by G200 Mystique.

All is working. RR_G was thrown out after the next "double editing card detected ..." message. This is the only one card that is not working on Abit board.
All other are working fine, in any slot I like.

So, the best suggestion is to change motherboard.
Another one is to buy new hard drive.
... new case with new power supply,
... better CPU,
... more RAM,

=
... new PC

This was exactly what I was talking about in my previous post.
I wish all you good luck with building a new PC around RR. I am happy to understand that it is possible in principle. The problem for many users is, however, in too large number of variants.

BTW, why Matrox has no such suggestion: sell PC's with RR installed and working? This all-in-one solution sounds the best.

Gino, how about duel? I will defend mysef by RR_G card. Is it strong enough?

Grigory

[GinoCyber]

What do you mean by duel?

Ciao