Today I decided to try and find the exact amount of latency for my 002 rack converters, and ran into a couple of questions. I sent a stereo track out analog outs 3/4 and returned that to analog ins 3/4 and recorded.
Will the HW Buffer Size matter if no plugin 's are being used? I know that total latency is calculated by adding the converter latency (DA & AD) plus the latency of the HW Buffer. Since I was using no plugins can I assume that 95 samples of latency is purely the converters?
My real question is why did I receive a variation between 93 and 95 samples of delay when repeating an identical procedure? Is this the instability of the 002's internal clock. Grant it, this is only a difference of roughly 2.1 and 2.2 ms, but a difference that I had not expected to see.
Comments
I tried the same thing initialy as you and was surprised to find
I tried the same thing initialy as you and was surprised to find that that 95ish sample latency did not change no matter what the buffer was set at. I then realized that I was measuring the latency in the wrong place. The setup as you describe measures the time it takes from output to input, which is the combined converter speed. However, a singer hears the latency from input to output. To measure this you need to go one step further. Route tone out a channel, in another aux channel or audio track, out another and in another, so it is now going analog twice. You should notice that the 2nd track inline will show latency according to the buffer setting , plus about 280 samples which is the three conversions at about 95 samples each.
Why it can change I do not quite get,mine does the same. Another odd thing is that my Roseetta 800 and my Alesis AI3 both run about 7 samples faster in their conversions than the 002R, while running through the 002R optically.