It's crucial that you understand the concept of latency.
Latency it the time delay between the trigger and the operation.
In terms of computer sound recording and generation:
1) The time between the midi trigger and the sound generated by a software
synth to come out of the speakers.
2) The time between the analog audio signal input to be converted to digital and stored on the hard drive.
3) The time between the digital audio track to be retrieved from the hard drive and converted to an analog signal that comes out the speakers or headphones.
4) All of the above at the same time.
The factors that contribute to this latency form a pipeline that include the A/D (analog to digital) device
(UA-25), the software drivers for that device, the operating system,
the recording application being used, the operating system again, your
internal computer speed and your hard drive speed. Remember that the sound needs to go both in and out, often simultaneously (thats called full duplex).
A latency of up to 50ms is probably tolerable, but for optimum
recording, it should be closer to 5ms. If you were to go over 50ms, you
would need some sort of time compensation, whether automatic or manual, because nothing would synchronize properly, be it the metronome or previously recorded tracks.
"What's this mean, and how does it relate to my situation?" you ask.
Lemme tell you. The recording program and drivers will incorporate software buffers to adjust for slower components in your sound recording pipeline. Without buffers, or with too few buffers, the pipeline can drop information (sound) if it is not be able to process it quickly enough. This will often result in stuttering and skips. Sound familiar?
So why not just put LOADs and LOADs of buffers in the pipeline? You most certainly can, but for every buffer you add, you gain a bit of latency. And remember that too much latency is a bad thing.
"Ouch! I still don't get it. What am I supposed to do?" you ask.
I reply, "Good question, Grasshopper." You want to stop your stuttering. You need to increase the efficiency of your sound pipeline. For regular people this means the following, in order:
- Make sure you purchase an A/D-D/A device that has low-latency drivers utilizing one or more of the following standards: ASIO, WDM, MME or CoreAudio. Selecting the one that your RECORDING APPLICATION utilizes. ASIO is widely supported and good. WDM is the worst because it is really just regular Windows drivers, but they can sometimes be optimized for low latency.
- Make sure you install your drivers properly, selecting the one that your RECORDING PROGRAM uses.
- Make sure that your recording application has the hardware selected for your A/D-D/A device and the drivers that you previously installed.
- Make sure that your recording application is using ONLY the hardware specifically designed for low latency, which is your A/D device that you previously purchased and installed.
- Make sure you include as many buffers as you can while maintaining a low latency. Often, this is dependent on your recording application.
- Set the UA-25 to 44.1KHz or 48KHz, not 96KHz. 96Khz does not allow for full duplex.
Do these things and attempt to record something. If it stutters right away or you get "underruns" you need to increase your buffers. If everything is peachy for a while, but soon you start getting skips, try increasing the buffers, or get a faster hard drive or turn off some digital effects.
In summation: A Celeron 1.6GHz should have sufficient power to record, play and add a few VST plug-ins, maybe even a dozen or more. Chances are that your hard drive is fast enough to retrieve and record 16-32 tracks at a time. Turn off all the stupid, little programs that make icons near your clock, like the DELL support and Weatherbug and AIM and AOL and HP Monitor and anything that takes processing power and hogs memory.
×