Does latency compound?

[travelin travis]

I can now see how some of the arguments around here about how latency affects performance were wrong on both sides. When I set my sound card to 48k, 256 buffer samples, there is 18 ms between when I pick the note on my guitar and when the sound actually reaches my ears thru the headphones, assuming any plugins on the track aren't adding in any significant additional latency. I guess I can measure that now too.

 

[travelin travis]

I guess I can measure that now too.

The amp modeler plugins didn't add any significant latency. 18 ms is still pretty damn high though. That's like standing 20 feet away from an amp while playing.

I measured my usb keyboard latency too. 15 ms between the time of sharply striking a note on the keyboard (recorded with a mic) and the midi note actually being recorded. Deduct 5 ms for the audio input latency. 5 ms between the time of the recorded midi note and the snare drum of bfd being triggered in reaper. Total of 15 ms. That doesn't include the trip of the snare drum sample sounding thru my headphones. Add another 5 ms. This was at 44.1k, 128 dma buffer. That seems like a pretty significant amount of latency too.

As I understand it, latency compensation is for like say having all your plugins play in time with your music. It helps correct different latency times between software during playback, it's not overall system correction.

My reason for looking into this stuff is because I think it affects performance (timing) in a negative way. The software may compensate after the fact but that doesn't change the actual performance. Btw, a loop from output to input in Reaper ended up at about 8 samples misaligned.

 

[jarocho]

You could always shift your tracks over to the left accordingly (a few ms). Latency will always be around and will always be an issue, just learn to deal with it.

 

[Robert D]

I can now see how some of the arguments around here about how latency affects performance were wrong on both sides. When I set my sound card to 48k, 256 buffer samples, there is 18 ms between when I pick the note on my guitar and when the sound actually reaches my ears thru the headphones, assuming any plugins on the track aren't adding in any significant additional latency. I guess I can measure that now too.

Use hardware monitoring - zero latency.

As I understand it, latency compensation is for like say having all your plugins play in time with your music. It helps correct different latency times between software during playback, it's not overall system correction.

That's my understanding also. You can't eliminate system latency, the idea is to make everything equally late.

The real timing problem is the way you are recording drums. You are doing the hats separately from the kick and snare. When you do that, you run the risk of your feel being different for the hats than it is for the kick and snare. It won't sound like one performance.

For sure! Back before I had a mesh head kit (and now BFD), and did drums on a keyboard, I'd use my right hand second finger (the long one) for closed high hat, and the pointer for open hat, and then the left hand thumb for kick and 2nd finger for snare. Try it on your desk, it will feel very natural to rock your left hand for the kick and snare, and the pad of your thumb vs your fingernail will even sound right. Assign a second key wherever it works best for snare for the right hand so you can do snare rolls. You can play a whole song this way if so inclined, or at least whole verses, choruses, etc. Fills can be added later, just delete the third arm high hat wherever you put the fills. You can get a pretty good groove this way.

 

[travelin travis]

For sure! Back before I had a mesh head kit (and now BFD), and did drums on a keyboard, I'd use my right hand second finger (the long one) for closed high hat, and the pointer for open hat, and then the left hand thumb for kick and 2nd finger for snare. Try it on your desk, it will feel very natural to rock your left hand for the kick and snare, and the pad of your thumb vs your fingernail will even sound right. Assign a second key wherever it works best for snare for the right hand so you can do snare rolls. You can play a whole song this way if so inclined, or at least whole verses, choruses, etc. Fills can be added later, just delete the third arm high hat wherever you put the fills. You can get a pretty good groove this way.

That's the way I've done it since I got my first midi controller keyboard. It's difficult to control hit dynamics with a keyboard for me. Add in the latency and it all feels like a mess, although it seemed subtle in the beginning. I tried the drumtrigger plugin in reaper a few nights ago, which seems pretty interesting. I just layed a mic on the floor and tapped my foot up and down to trigger the kick. It's a really cool feature of reaper but the latency still seems weird. I put the mic under a book, grabbed some sticks, and tried playing snare too. It feels a little better than using a keyboard but the dynamics are still kind of funky. I've played on a roland mesh head kit a few times too and it seems about as weird as the mic/book thing. MIDI just seems to suck for drums to me. I wish I had a good sounding space to mic real drums.

 

[travelin travis]

Woo hoo! I can get my sound card buffer down to 64 samples when using Reaper while tracking without any pops/clicks/drop outs. I can't do that in any other software that I've tried. Drums don't feel so weird now and being able to use a mic as a kick trigger helps a lot. Now to remap the kit since I don't have to play the kick via finger any more. The dynamics don't seem so hard to control at lower latency either. Maybe I was having to concentrate on the timing too much when running at higher latency. This is pretty fkn cool.

I'm betting that the weird feel I experienced on that Roland mesh head kit was caused by monitoring over the Roland speaker system instead of headphones. I bet I could put some triggers and mesh heads on a real kit and use my sound card's inputs as trigger inputs now.

Hey Robert D, does the hi hat respond well on your electronic kit with BFD?

 

[Robert D]

Hey Robert D, does the hi hat respond well on your electronic kit with BFD?

It's OK, but I use real hi-hats anyway. The base version of BFD has only closed, half open, and open hats. The XFL upgrade gets you hats with closed, 1/4 open, half open, 3/4 open, and open. My hi hat controller is fully variable, so it's strictly up to the sampler how well it works.

Good to hear you've found something that's working better for you. Reaper seems to be gaining a lot of fans.