delay problems with recording via interface

[Marco Braam]

I am a Logic Express user and to record voice, drums and guitars I use an M-audio interface.
There is a most annoying delay in the incoming signal which I haven't been able to get rid of.
I have already limited the in and output signals on both interface as Logic track but so far with little succes.

This problem started when I switched M-audio interfaces. I have recently bought a version which enables me to use a phantom powered Mic, but the delay is also current when the phantom power switch is turned off.

Does anyone have a clue?

Thanks,

Marco Braam

 

[EioN]

I would try freezing the tracks that are already recorded / midi to give your cpu some breathing room, or just record with a metronome and mute everything : /

 

[Peter Ostry]

What means "delay" in this case - do you hear the input signal twice or only once but too late?

 

[Marco Braam]

I hear it too late, which makes it almost impossible to play

 

[Peter Ostry]

An audible delayed monitoring signal can be produced by software monitoring. With this type of monitoring the sound goes through your interface, enters the computer and Logic, gets processed and finally sent out through the whole chain again. This takes a while depending on the plugins you use. There are several ways to minimize or completely avoid the delay:

1. Use the Low Latency mode of Logic (English manual p. 1198 "Working in Low Latency Mode"). This switches all plugins off that would produce more than an adjustable amount of latency (= delay). The sound may change dramatically of course.

2. Switch everything you don't need manually off. This way you may be able to keep an acceptable sound with less latency.

3. Don't use software monitoring but rather "false" direct monitoring. For that you get your recording signal directly from the interface and only the playback comes from Logic. This way you get just the latency of the interface which my be be around 10 ms.

4. Use "real" direct monitoring by splitting the recording signal before the interface. This is the only method to monitor without latency.

Method 3 (monitoring through the interface) is good enough for most people. If you need some reverb on the mic signal you can get the wet part of the reverb from Logic because a reverb needs some delay anyway.

A hint: You don't need to have software monitoring on just for a reverb on the recording signal. Send the recording signal to an Aux channelstrip, insert a reverb, set it to 100% wet and adjust the volume. Auxes play always, they don't need software monitoring to be on.

... the delay is also current when the phantom power switch is turned off.

You should not hear anything (or very little) without phantom power. Condenser mics need phantom power, they don't work without.

 

[Marco Braam]

Peter,

Thanks for your comment, but first of all: my quote on phantom power was misinterpreted. I meant that when I record something other than vocals and therefore I won't be needing Phantom Power the switch on the interface is turned off. That makes no difference.

I have a question on your comment: You write:
3. Don't use software monitoring but rather "false" direct monitoring. For that you get your recording signal directly from the interface and only the playback comes from Logic. This way you get just the latency of the interface which my be be around 10 ms.

4. Use "real" direct monitoring by splitting the recording signal before the interface. This is the only method to monitor without latency.

I don't know how to switch to "false" direct monitoring or "real" direct monitoring for that matter.
What's the difference and how do I get there?

 

[Peter Ostry]

I don't know how to switch to "false" direct monitoring or "real" direct monitoring for that matter.
What's the difference and how do I get there?

"False" and "real" (or "true") direct monitoring are not official terms. Unfortunately everybody calls it "direct monitoring" or "hardware monitoring" but there are two ways to do it and they are quite different. And no, you cannot switch between them. The two techniques have a different signal flow and major differences in the hardware setup.

There may also be differences between audio interfaces depending on the internal signal flow and wether their "direct" monitoring goes through the analog/digital (A/D) or digital/analog (D/A) conversion or both or none of them. However, below you see examples for two common setups without going too far into details.


Direct monitoring through the audio interface:

You want to hear the recording signal with the shortest possible latency, therefore you do not send the signal through the computer and the DAW but rather through the interface only.

The latency depends on the internal signal flow of the audio interface. Where are the A/D and D/A converters, does the recording signal pass them or not? Do you know your interface good enough? I don't. But the latency is very short anyway, many people are happy with this kind of monitoring and don't care about technical circumstances.

The signal flow is pretty clear, you have to use the internal mixer of the interface. Consult your interface manual to learn how to send the incoming recording signal to the DAW and to the headphones or monitor speaker(s). Do not send the recording signal back out from the DAW because you would hear it twice: first the direct signal and a little later the signal after it ran through the computer. From the DAW only the playback gets sent to the audio interface.


Direct hardware monitoring:

Hardliners and latency-sensitive people do not insert an interface into their recording signal for monitoring. For this method you split the signal before the interface, either directly after the microphone or after an external preamp. When we agree that analog components (practically) do not produce delay, the latency of this monitoring method is zero.

This is the very best method but splitting weak recording signals is not trivial and requires more equipment. Therefore this "real" direct monitoring is rarely used in home studios.