Cubase 5/7 dropout during recording with N-12

I worked two years with mac mini and Yamaha N-12 without issues. Then switched to Win7 64bit and I am experiencing dropouts during recording. It does not depend on track count (1 or 12). It is very rare, but the problem is that no warning is displayed. I can hear the dropout when listening to recording. It is a very short silence in the wave file. Is there a way to determine if it is the FireWire dropout or SSD dropout? Is the overload message box used for playback dropouts related to recording or recording is without any warnings?

Also if it is a disk (SSD) dropout, will I have silence in all 12 track from my recording as it actually happens?

One more thought - the number of track does not affect the frequency of reproduction. If it was disk/cpu issue the reproduction ratio will increase with number of tracks. But if it is cable or firewire issue probably it will not depend on tracks count.

What’s your audio latency setting? May be too low?

Aloha m,

1-Did this just start happening right after the change
or did C5/Win 7 work OK for a while and then go wonky?

2-OS wise, do C5 and Win 7 play well together or would
C5 and XP be a better fit?

3-Are your n12 drivers/tools up (or down) to date?
http://download.yamaha.com/search/product/?site=usa.yamaha.com&language=en&category_id1=16244&category_id2=16268&category_id3=&product_id=592444
{’-’}

I faced the problem after switching from mac to win7. All drivers are latest. Buffer is huge - 1024.

2.5ghz core 2 duo, 6gb ram, ssd disk. I changed the mac mini with this similar machine because playback performane was ways better even with half the mac ram. But then I faced the recording issue. I am using the motherboard firewire, which shares same virtual irq as sata controler. I am not sure if I can run the latency tester tool for some hours, so it can detect the issue. The drop is always with same short duration of silence in the file for all recorded tracks. It is as short as cubase click, but no warning is displayed while recording, so you can omit that. I ordered TI chipset PCI to test, because if it was a performance issue it should depend on tracks number and latency buffer, but it does not. The other difference after mac setup is the firewire cable.

One more thing. I did some work with this setup without issues. Then the drops started, but I found that ssd is full. I freeed some space and finished the job without drops. But after that on a new project I faced the issue again. Currently I have 20GB free. Maybe it must be more, considering the ram size of 6gb and corresponding swap file.

I was going to ask if you had the TI chipset but I see you’ve done that. You are running a separate SSD for systems and another for recording files, correct?

No, just one ssd for all, since random access is not an issue for ssds. I will try the latency checker tomorrow, but I am not sure if it keeps long log history, because drops are very rare.


I ran the checker and there are only small green bars, peaking at 140us

Ok, I just installed TI chipset firewire card and as opposite to motherboard firewire, now I can run two n-12 at 96kHz. With motherboard firewire maximum was 48kHz, so probably there is bandwidth improvement. But unfortunately the problem still exists. I run the latency checker tool in background, while Cubase 5 was recording 24 channels at 96kHz and the same short dropout (32ms) appeared after 3mins and 12sec. No any red bars on the latency checker. I just recorded the mixer preamp noise, normalized all tracks, so I can see the silence dropout. It is again for all the recorded tracks.

After further zooming I found that all channels stop in sync (to silence for both n-12), but second n-12 channels recovers 20 samples later. It looks like firewire stream break.

I would start using process of elimination. First, try a different interface to see if there’s incompatibility between the N-12s and W7 64. If that’s not the problem, I would entertain using a separate SSD or HD for transfer to humor the test.

I mentioned above that 2 x n12 can work at 96k, which was not possible with motherboard firewire, but it is very inconsistent. Sometimes after reboot it cannot start at 96k, but works at 48k. Without closing cubase, I unplugged the firewire of the 2nd n12 and plug it again. After that cubase started working at 96k with both mixers. Other strange thing I noticed is that after closing cubase, the latency checker starts hitting red bars. After opening cubase 5 or sound forge it goes back in green bars. I am not sure if it is a latency checker bug, because during the red bars if I restart the latency checker it shows green bars.

Today I moved the firewire pci to other slot and then I did a very long stable recording 24ch at 96k. Restarted cubase a couple of times and it synced flawlessly, no dropouts at all. Then I restarted to see if it is fixed. After that the same inconsistent and unstable behavior. Can such inconsistency be caused by drivers? Maybe I have to by other firewire cables.

I also tried the legacy firewire driver, but nothing better has happened.

Behavior is the same with Cubase 7.

Tried with two new firewire cables. After first boot all was perfect at 96k. After reboot only 48k, but did not record a lot to face a dropout. Strange inconsistency in sample rate connection.

New test: Start machine and no sync at 96k at the beginning. Then I found that after disconnecting the second n12 mixer from daisy chain, while running Cubase5, with only one mixer it syncs to 96k. Then if I reconnect the second one, both sync at 96. So maybe there is some race between both n12s and drivers. I am not sure if drop-outs and this sync issue at higher rates are related.

Do you have the capability to connect both N12s to the computer as opposed to daisy chaining?

Yes, the firewire pci has 3 ports, but I was not sure if it is ok.

I don’t know the protocol of your system but I’m not sure it would clearly see the separation of both N12s through a single firewire cable. You should check it out. I don’t think it could hurt your 'puter.

Aloha guys,
I recently connected two n12’ (daisy chained) for a session with C7.
Now I have done this many times in the past but only while using C6/6.5

Everything worked flawlessly so after the C7 session I asked the guys to stay for a moment
and just jam a lil to assist me with a studio question. (beer and pizza helped them to decide :slight_smile:)

While they played a couple of songs I tried the same set-up
while using C5.

Perfect! No probs.

C5/two n12’s/six piece band.

Nice.

C7 is kool but C5 still rocks.

BTW
because the fellas were just jamming and ‘off the clock’,
there were some very nice performance moments recorded.
Nothing commercial but some nice playing.
All captured on C5.
{’-’}

Thanks for that update Curteye. That’s good to know but the one component where Mixexperience differs is that the problem started when there was a switch from Mac to PC with W7 64bit. By the looks of your set-up, you are on a Mac platform.

Have you checked your CPU usage during recording? A 2.5 GHz Core 2 Duo might be enough to run fine on OS X but not on Windows. There may be some background process that starts every once in a while and peaks CPU usage which causes the dropout.

Does Cubase detect recording dropouts? It does not display a warning, but the dropout is there. Maybe if it is a firewire driver dropout it won’t be detected?

It could be anything - even an electrical issue. I once had problems with USB devices interfering with the firewire port on an iMac causing lost data on a hardware raid. Try removing everything that is not absolutely necessary from the computer’s ports, no daisy-chaining Firewire devices etc. Record to an internal drive.
I am not an electrical engineer but there seem to be all kinds of electrical interferences which supposedly can sometimes even be solved by simply rotating the power plug by 180 degrees in the socket.