Cubase record-mode dropout problem

skippy

New member
This is a very vexing problem. I've searched this site, the various Cubase and PC_DAW sites, and the sites for all my hardware vendors, and haven't been able to solve it. I've followed every PC setup/optimization suggestion I've seen, and I still haven't been able to solve it. Most recently, I sacrificed a live potted plant to the Gawds, and yet still no joy.

Symptom: when recording, Cubase will drop out of record at random times. This happens about 10% of the time while tracking. It occurs at a different point each time, and is completely nonreproducible. It also occurs regardless of how many tracks I'm recording: it is equally likely to occur whether I'm recording one track as an overdub, or flying 16 tracks in at once via ADAT litepipe. CPU load (according to the perfomance meter) has nothing to do with it either: I can be running 1 track and only monitoring the click, or I can be playing around with lots of tracks and plugins, and have the meter up in the 70% range: no difference. 10% of the time, the thing just stops recording mid-take, freezes the screen while it updates the waveform display, and continues playing back as if nothing has happened, leaving me pissed as hell.

And yes, I'm sure that I've got the right marker out past the end of the song... (;-)

System: P3 800MHz, 512Mb, ASUS CUSL2-C, 20Mb SCSI system disk, 60MB SCSI audio disk (both 7200RPM), Initio In1-9XXX UW SCSI adapter, Yamaha CRW8824S SCSI CD-R, W98SE stripped down as far as humanly possible. Running Cubase VST32 5.0 R3 PB2. Sound card is an RME Hammerfall, clocked in slave mode via word clock from my Fostex D1624. The system also has an Audiophile 2496 installed, but I use it only for its MIDI interface at this time, receiving MTC from the Fostex. Also have Midiman Midisport 2x2 installed for instrument MIDI, with the latest drivers from Midiman for both. Networking is via a 3com Etherlink 100baseT Ethernet interface, connected to my network via a switch (not a hub), so there is essentially no inbound traffic congestion. Graphics are Matrox Millenium G450, and there's nothing else on the system: Cubase, Wavelab, and Sound Diver are the only apps. It's totally dedicated to the DAW task, and to allowing my sound data to be backed up over the network.

I've stripped, tuned, tweeked, optimized the buffer sizes, and finally run out of patience. Followed all of the setup advice from Slackmaster2k and others here that the search function could come up with, all the advice on the Cubase support sites, all the tuning tips from the RME site. Audio sync is solid as a rock: word clock is properly terminated, and the Hammerfall is the only load on it.

And yet I still can't get through a session without it randomly dropping out of record at least once or twice. I've seen lots of references to this problem on the Usenet archives and other sites- but none of the fixes suggested there either work, or apply (most of them I'd already done by following the tuning suggestions, for example disabling power management and systray junk, bumping up the buffer sizes, and the other more or less obvious stuff).

Any suggestions would be highly appreciated: It really does annoy me. No, actually, it drives me apeshit wildman bugfuck _nuts_. The good news is that I always track to the 1624 simultaneously, so I don't lose the take- but then I have to fly the data down, and I'd like to avoid that overhead. It oughta just work, dadgummit! Until I can get this to go away, I'll never trust the silly box.

Whaddaya think, computer recording guru persons?
 
The only things I can think of off the top of my head are not very likely. One thing you can try is updating to the latest version, either R4 or R5 PB3. The other thing is that MIDI messages are somehow taking it out of record. Make sure the remote control is disabled (hmmm...and even if it is, could there be a bug?) You might try disconnecting your MIDI totally and doing some experimental audio recording.

Jim
 
I've noticed that when this site's server is going especially slow, adding a new post does not bump the thread up to the top or show that there is another reply. Let me try again.
 
I'll try the MIDI disconnect thing first, but I don't hold out much hope- without the MTC sync, it'll be a damned sight harder to make use of the backup track for the 1624, or to fly tracks back and forth. It's worth a try, though- I hadn't gotten that desperate yet...

I'm not using any remote at the moment, so I don't think that that's likely to be it. There shouldn't be anything flying around via MIDI other than timecode once I'm into the middle of a track.

I'll also look into doing the upgrade. That is almost always worthwhile. I had been just trying to make one change at a time, and it is time for that one! Thanks very much for the response.
 
Updated to R4 this evening, and that seems to have done it, maybe, I hope. I'll need to put 10-15 hours on it before I'll believe it, but tonight has been the first night in 6 months that the damned thing hasn't screwed up mid-take at least once. Three hours, and no dropouts. One can always hope... I'll let you know.

Thanks for the suggestion. It's really been unbelieveably annoying to have done everything right (according to conventional wisdom), and still have encountered fatal problems literally every time I used the machine. If anybody else is running VST32 5.0 R3 PB2, consider upgrading to R4 ASAP- if not sooner. Looks to me like that version is broken, even though that is what you will get on your CD-ROM when you buy a new installation...

For those who say that it's more or less trivial to build a DAW and make it work first try: couldn't prove it by me. This problem isn't even acknowledged by Steinberg, and there's nothing in the release notes for the R4 patch that would indicate that it might help in any case...

It is wierd how this thread doesn't show any of the new posts, isn't it? It's gone all day without showing any of these updates... Maybe this time the counters will update.
 
I knew you weren't using the remote, I just thought Cubase might have been mistaking the MTC or something else for a remote command to switch out of record. I wouldn't be surprised if some bug fixes are not documented, especially if the programmer finds the bug and not the QA person. Also, it's possible that you fix a bug that had manifested itself in more than one way, so you inadvertently "fix" bugs you didn't know you had.

They're pumping out new beta versions like once a day. They're up to R5 PB6 as of yesterday. You can keep an eye on it by ftp-ing to ftp.steinberg.net. I guess if anyone wanted to report a bug, this would be the time to do it.

By the way, skippy, I was just enjoying your great posts on the thread 24 bit / 96 kHz. Really interesting stuff!

Jim
 
Eeek- I honestly didn't realize that the product was _that_ unstable. That's certainly not what I gathered from the folks who sold it to me... I'm not too kindly disposed towards beta code, though, so I'll try and stay with the official releases whereever possible (which is what I thought I had!).

I'll check out the FTP site; looks like there is a decent enough revision history with each release. Now I understand what "PB2" meant: it absolutely blows my mind that they'd do a CDROM distribution with public beta code on it! That's what I got with my clean install.

I'm glad you enjoyed those posts- I surely made a fool out of myself, talking about ancient history as if it was still the state of the art. Jeez, I musta slept since then.

It's been a lot of fun getting back up to speed on the current stuff. That wasn't my intention when I started building the new room- all I wanted to do was get back to recording. But there's just no way to keep the nerd in me from coming out, seems like!
 
Just a followup: I'm now running R5PB6. Jim, many, many thanks for hooking me up with the cubase.net and ftp.steinberg.net resources. This looks like a good version: it's getting good reviews over there, and everything that I need actually *works* (like recording). I'm probably 10 hours in now, with no dropouts, and no crashes. Halle-freakin'-lujah.

It has a couple of odd screen refresh bugs, like when you resize the mixer windows (the clipping algorithm is broken, so it flickers ghost channel strips at you to the right of the cursor). But other than that, it makes noises pretty dadgum well. Certain chunks of the code seem to have been rewritten for more efficient use of resources: I can now set the latency of my Hammerfall down to 11ms and run 16 tracks and boatloads of plugins with no drama, and less than 40% CPU usage showing on the meter. Happy camper- and I'll be happier when the real R5 comes out...

I'm probably the last person on the planet to find out about those resources, since I naively assumed that I'd be getting production-quality code when I wrote the big check. If anyone else uses Cubase and hasn't taken a look at these, please do. It may very well make your life more pleasant...
 
That's good to hear. I think I'm going to upgrade to R5 PB6 then. I normally stay away from the betas. I'm really surprised they put a beta version on your production CD.

I don't think I mentioned cubase.net; you musta found it yourself. I always just check the first page there to see if there's any news. I never even noticed that it had forums until now when you mentioned it. A while back I used to go to Cubase for Windows Users. But I just checked it now and the forums look dead. I wonder what happened. Anyway, there are some useful tips there in the support section about tuning Windows for best performance. But I guess if it ain't broke, don't fix it.
 
Scroll down to the beta forums- they are at the very bottom of the cubase.net forums page. There's active and ongoing discussion of all the outstanding issues with the various releases and public betas. Following your pointer to ftp.steinberg.net led me to those, and I have found that the "search" function there, and the archives of that group, are a very, *very* useful tool for researching- shall we say- "antisocial" behaviors. Just make sure that you punch "refresh" each time you go in. Perversely, my W98SE Internet Explorer on my Windoze play machine works *only* from its cached copy of that site, so I have to manually crowbar it into looking again and refreshing everything from soup to nuts to see any new posts. Why? Search me. I'm a Unix/Netscape guy for everything but DAW topics. All thanks to Microsoft for non-updating page views, I guess...

I also would normally avoid betas like the plague. However, a couple of the leading lights there have pronounced R5PB6, and I quote, "the best Cubase ever". Cool: If these individuals have been beta testing this product as long as it appears they have, that carries significant weight with me, and led me to move up fro R4. Check out a gentleman who posts by the name "Hippo" on that site: he would appear to be the a major-league power user, and spends what must be a great deal of his time reproducing user bugs...

Like I said: I would never have known the "PB2" meant "public beta 2, pray for me" without that pointer from you. I suspect that R5PB6 will transmogrify into a real release-quality R5 very shortly, as it does put some *significant bugs* in R4 to bed. Anyway, the information you provided really did lead me to that resource, and it has proved invaluable. This very evening, I just finished up 3 straight hours of tracking and editing with a sort-of-client (since I am not doing this for money anymore, dammit!) who will probably turn into a collaborator: no droputs, no crashes, no Blue Screens Of Death. Print it, stripe a rough to cassette, send 'em home. That's a first. Yup, the first time I ever had that never-to-be-sufficiently-damned DAW *fail to melt down* with talent in the room.

Anyway, check out the beta forums there. The "beta lounge" forum is where potential bugs are discussed and reproduced by multiple people. Only after they are confirmed in the "lounge" are they welcomed in the actual beta forum. It's an interesting sort mechanism (read: experiment in social engineering), and it is instructive as hell to read through.

Damn. This dinosaur has actually used the Web for something *useful*. Alert the media! Hey, Ed: you heard it here first...
 
Last edited:
Back
Top