I’m wondering how accurate the latency measure is in Cubase. I’m using an M-Audio Fasttrack Ultra 8r on a project at 128 samples. The Device tab in LE4 shows that as about 5.3 ms in and 5.3 out. Is that an accurate read of the latency or do I need to do other tests to establish latency? I’ve read that most interfaces have hidden buffers that add to the actual latency, not sure about this device. A thread on the M-Audio forum says there are no hidden buffers on this device. How accurate is the Cubase measure with interfaces?
If you calculate the latency caused by the pure sample buffer, you´ll se, that with 5 .3 ms there already is some latency added somewhere. Apart from that, cubase can´t measure your interface´s total latency, since it can´t measure the A/D / D/A conversion.
Thanks for the info guys. When I record at 48,000, 128 samples, 24 bit, Cubase is showing 4.9 ms in 4.9 out, and the pure sample latency should be 2.67 ms, so I’m a little high there too.
As thinkingcap points out there is more latency in the conversions, so if I assume another 1 or 2 ms there, I’m at 11 or 12 ms total latency, which is good I guess – not great, but good. When I run the DPC latency checker with Cubase streaming all is little green bars, no yellow or red. I’ve tweaked and disabled a fair number of things in the system but could do some more to reduce DPC latency, which I assume is the extra latency Cubase is detecting.
Converter latency may or may not be included in the Cubase measurement - it is NOT with my RME interfaces (and I’m doubting that your M-Audio interface is including it either). I can only assume there is some extra buffering going on in your interface.
The DPC latency is a different thing all together and the latency it shows has nothing to do with the latency of your audio interface (other than the fact that a system with good DPC performance should be able to run at lower audio buffer settings without crackling etc.)
There is tool out there called the Centrance Latency Test Utility that will show the true round-trip latency of your system using a loop-back. If you jumper your D/A back to your A/D you will not be able to determine the individual latency of each converter - just the total of the two plus the hardware and software buffering. If your interface has digital I/O you should be able to measure just the interface latency.
and if you want to use plugins at recording
1 disable asio direct monitoring
2 disable adjust for recording latency
3 disable auto vst compensation
4 put your interface as low as your computer can do 64 /128/256 buffersize
after you have do this
put a plugin in the first click track that you see on the video.
when you have the value of the latency of the plugin in the second track
put this value in the track delay where is insering the plugin.
the value will be different for every plugin you use.
under 256 samples of asio buffer size
you will have no problem to record your plugins in real time and the result will be safe because you have compensate the latency before with the track delay of cubase.
thanks
weesch
