|
Re: Cubase SX 3 - External Instruments Problem
| quote: | | Originally posted by alanzo So, my question is, has anyone been able to get the Cubase SX 3 External Instrument feature working fully? If so, how? |
Use it all time without any problems. In my current setup, I have 5 external synths hooked up.
As mentioned you have to measure the latency seperately for each synth. If you are to setup a new synth, this would be the approach.
1) Connect the synth to an input on your sound device (let's call this port 4 and 5 in this example).
2) Make sure that your sound card doesn't have hardware monitoring (or through-routing) enabled on the port 4 and 5.
3) Make sure that port 4 and 5 HAS been configured in the device manager as active ports (and showable), but HAS not been configured as input ports in studio setup (F4-screen)æ
4) In studio setup (F4) add a new external instrument, choose the correct name and audio ports.
5) Now create a new project add a midi track, add your external instrument (by clicking F11 and choose it) and route you midi output to the instrument. Important note: Route the output to the INSTRUMENT and not just the correct midi port.
6) Now record a 4-on-the-floor pattern in, say, 120 BPM and choose a sound with a fast attack on your synth. Make sure your hits are over-quantized to the exact beats, loop selection over it, enable the metronome and hit play. You will now hear that the sounds are out of sync. This is expectable.
7) Compensate for latency. Press "e" (edit) button on the external instrument and you'll see a little screen that lets you set the delay in seconds. Although there are approaches to measure this to the exact sample, the quick-and-dirty solution is to adjust the delay now, while listening to the metronome vs. the sound from your synth. Once you start hearing them in sync, that's it. Your done. Cubase remembers the settings for next time you use that synth.
However, even though you now have audio in sync, you might still experience problems and this has to due with what latency arrises from. The following information can then be very important to troubleshoot or understand why "it doesn't get any better".
Latency comes from these factors:
1) Buffering in your computer. Your computer has a buffer set to allow some time for processing the audio in the CPU. The higher the buffer settins are, the higher the latency will be. You can actually calculate this latency exact. It's very simple: You just divide the number of samples in your buffer settings with the sample rate you work in. Now you have the latency. This is important to note, as it means that the latency varies accordingly to the sample rate. In other words: The "syncing" you performed above (in step 7) has to be done for the specific sample rate you work in. And every time you change your sample rate, you have to set the delay/latency compensation again.
2) A/D conversion. Your sound device needs to convert the sound from analog to digital and this takes some time. Most audio devices have a specific amount of time they take to do this. Some, however, use a verying amount of time that depends on other things such as how many of the other ports are in use at the same time. I would avoid such devices if I were you, however, unfortunately this is not information the manufacturer will/can give you. I bought an Echo Audiofire myself and found out that the A/D (and D/A) conversion doesn't take the same time always, but depends on how many of the ports are in use at the same time.
Anyway, this information might help you explain why you, even though you have syncronised your device, later will find that it's suddenly a little bit out of sync. But please note, that we're are talking very low variance here! It's enough so that it won't let you do a phase-null test, but it's often not enough to be audible (except when used for drums and sometimes basses).
3) Midi latency. The midi protocol is old and often there's latency involved in using midi. Another (and far worse) problem is that there's also jitter to take into account. Jitter is when latency is not the same, but changes a bit all the time. The amount of latency and jitter depends on your midi device. As a general rule, USB devices are bad, serial and firewire devices are better. You should however, when working in Cubase, get yourself a Midex8. This device (as the only one available) uses Linear Time Base. This protocol makes sure that all midi signals arrive on time, so it effectively removes this item from your "latency equation".
4) Synth latency. Your synth itself has latency. This is due to different factors in the synth, such as processing power, D/A conversion etc. Often this is a very specific number of milliseconds that's always the same. This makes it possible to measure it and take it into account.
So now, if you add the latency caused by 1, 2, 3 and 4 you should have a somewhat approximation of your total latency. You can see now that depending on what type of gear you're using, the latency can vary and that in some cases you might have to deal with changing latency and even jitter (changing latency and jitter is NOT the same thing).
In my setup (see my signature), I've recently added a Midex8 device for triggering the synths. It solved most of my problems and gave me rock solid timing - well, almost. I'm still not on sample precision which is due to the somewhat strange behaviour of my AudioFire12 device (where the A/D and D/A-conversion on the specific port takes longer or lesser time depending on how many ports are in use at the same time. But it's close enough. I'm using all my hardware synths with Cubase, fully satisfied with the results.
Btw: If you, or someone else, wants to know how to measure the EXACT latency (talking sample accuracy, here), let me know. I have a great procedure for doing this.
Regards
___________________
RiValMan
- Pumping the bits using:
My beloved MacBook Pro, Logic Pro X, UAD hardware and software and tons of plug-ins...
Last edited by RivalMan on Sep-24-2006 at 13:09
|
