Device Details
Overview
| Name | Version: | Audio Timer 1.0 |
| Author: | JGJP |
| Device Type: | Audio Effect |
| Description: | This is a set of 4 devices that allow you to check latency across your live set. The devices communicate with each other to determine how long it takes for a MIDI or Audio pulse to travel from one to the next, there are 2 send devices (1 for MIDI, 1 for audio) and 2 receive devices (1 for MIDI, 1 for audio). The send device sends a signal and the receive device reports how much time it took for the signal to arrive, they will both display the result and with several pulses they’ll show an average of all the results so far. You can use several pairs of devices across your set by assigning each pair a unique group number, also instead of using the in-built metronomes you can send your own MIDI or audio streams through. Any audio needs to start and return to complete silence for the audio timers to register properly. You can have send and receive devices on different tracks and you can time the latency between a MIDI send and an audio receive if you have an instrument in between. |
Details
| Live Version Used: | 9.1.6 |
| Max Version Used: | 6.1.9 |
| Date Added: | Nov 27 2014 09:28:22 |
| Date Last Updated: | No Updates |
| Downloads: | 3 |
| Website: | http://jgarcia.hungry.jp/project/timer/ |
| ⓘ License: | Attribution |
Average Rating
Log in to rate this device |
-n/a- |
Comments
Login to comment on this device.
Browse the full library
The only reasonable results I get is when the audio buffer is at 512 samples (44.1 kHz). Here sending out to audio interface and right back in again in Live 11.3.2 (M1 native):
Buffer | Live overall latency | Timer latency
512 smpls | 26.4 ms | 24.58 ms
256 smpls | 14.8 ms | 24.58 ms
128 smpls | 8.96 ms | 24.58 ms
64 smpls | 6.05 ms | 24.58 ms
32 smpls | 5.33 ms | 24.58 ms
1024 smpls| 49.36 ms | 74.58 ms
2048 smpls| 96 ms | 124.58 to 171 ms (average ≈ 143 ms)
Sending directly from one track to another also gives the 24.58 ms reading, strangely only when using 32 to 1024 samples buffer. Which shouldn't affect internal timing.
Given these results it would seem there's something with calculations that haven't survived Live 10 and 11. I'll make sure to test in the latest Live 9, when I get the chance.