macProVideo.com
Forums
  • Jauqq
    Posts: 7
    Joined: Jul 27th, 2006, 10:17
    Q for Tom Rice. Midi latency setup help? (or anyone else help?)
    I've just bought the excellent Tom Rice tutorial video 207: Logic's Latency Toolbox, and have found it to be the best video i've bought! The information is essential for anyone working with hardware synths and Logic. I've gone through it once already, but have a few questions regarding the setup procedure for midi latency. Hoping that Tom Rice will be able to share his expertise in this. I use a Motu multi input audio interface with all my hardware synths routed to its inputs, and a Unitor8 multi port midi interface. Monitoring is through Logic using software monitoring with a buffer size of 32. Up until now my method for setting external hardware synths was as follows. I'd first work out my recording delay, and i was already using the "ping" method as Tom describes. This ensured that recorded audio would be placed where it should be. However my method for each synths midi latency... or reaction time to midi data, differs from Toms. I'd set up a midi track containing four bars of same pitch, same velocity quarter notes to trigger my synth. I'd then set up an audio track, with its inputs set to the channels that my synth is plugged into on the Motu (ie 1+2) with input monitoring switched on. Then i'd recorded enable this audio track, so as to record the synth signal coming into the audio track, being triggered by the midi track. Once i've a recorded audio file, i'd then open the sample editor, and check how late off the grid the recorded audio is. I'd get an average figure, and this figure would be the synths reaction time/latency to incoming midi. As an example, say i'd find the synth averaging 3.1ms late from the grid, i'd enter a negative value of "-3.1ms" in the track delay parameter box. Running the test again, the recorded audio would now be more or less bang on the grid....where it should be. I'd repeat this for each hardware synth in turn, going along and working out each synths latency time to incoming midi. And each time entering a negative value in ms in each hardware synths instruments track delay parameter. This was how i'd do it....until I saw Tom Rice's method, which is leaving me really confused. Why is Tom not using an audio track to record to, but instead using a bus to record to? That has totally confused me!! In the video called "compensating for midi playback latency 2" there is an example of slicing using the marquee tool, and working out average's etc.... Surely it's simpler to just count the samples from grid markers to starts of transients, work out an average value and then enter this number as a negative amount in the instrument track parameter? Tom's method seems way over complex....That is unless my method is doing something very very wrong.......? Hope Tom, or someone else can explain why or if my method is way off...?
    Reply
  • Dydavee
    Posts: 16
    Joined: Sep 19th, 2007, 07:47
    Re: Q for Tom Rice. Midi latency setup help? (or anyone else hel
    Hi Jauqq, Glad you're finding the tutorial useful, and thanks for the kind words. From what you've written, the way you have been doing things is basically fine. My method is a little more complicated, as it involves a two-step process using both the track delays and the All MIDI Output Delay preference. The reason I use this method, is that ultimately it provides more flexibility than when using the track delays alone. For example, with my method, if I need to change buffer size mid-project (which I rarely do ... but you never know!), I can quickly re-sync all external MIDI instruments to the new buffer size by simply adjusting the All MIDI Output Delay (as outlined in video 45, [i]Compensating MIDI Playback Latency after Changing Buffer Size[/i]). The alternative in this situation would be having to reenter the track delay for each instrument in the project, which could potentially be quite time consuming. Also, with the method I outline, you end up entering positive track delay amounts, rather than negative ones. This leaves the option open to use a Sample Delay plug-in on the instrument's input channel, in place of the track delay parameter (as outlined in video 49, [i]Alternative Method for Compensating MIDI Playback Latency[/i]). (Incidentally, as of last week, using Sample Delay plug-ins instead of track delays is my preferred method for working around the track delay bug that I outlined in video 48, [i]Bug Shop - Part 3[/i].) So if you don't think you'll ever need to change buffer size, and the track delay bug isn't rearing its head on your system (btw, I'd be interested to hear if it's not, as it's happening on both my systems), then there's no reason to change your method - except for possibly one thing ... ... and that is, how you record the MIDI instrument's audio signal. Take another look at video 40, [i]How to Record External Instruments to Audio - The Right Way[/i], as it explains why, when recording external MIDI to audio, I set the audio track to receive its signal from a bus, instead of directly from the audio interface's input. In essence, though, for the MIDI playback latency test to be accurate, the MIDI instrument must be recorded from the point at which it is heard. Recording directly from the instrument's input into Logic, records it from before this point. This means that if you don't use the bus (or loopback) method that I talk about in this video, and instead record directly from an input, the waveform on which you will be basing your measurements will be placed early against the grid, by the amount of the system's roundtrip latency. This will lead to you under-compensating your playback latency - so it will look fine, but sound late. At a buffer size of 32 samples, this will only amount to a few ms, so its impact won't be too apparent, but when testing and working at larger buffer sizes, it will be more so. Hope that helps clear things up for you. I know I can only get my head around this stuff after a good strong coffee or two! Tom
    Reply
  • Jauqq
    Posts: 7
    Joined: Jul 27th, 2006, 10:17
    Re: Q for Tom Rice. Midi latency setup help? (or anyone else hel
    Many thanks for the reply Tom. At the moment i'm unable to confirm the track delay bug on my system, as i'm re-ordering my setup. However I can confirm that even though i'd be able to achieve very tight midi latency/jitter for soloed midi synths (less than 2ms of jitter), latency would be introduced as more midi tracks were added going along the individual midi out ports on my midi interface. I took this as being down to the serial nature of both USB and midi, and also that the whole OSX midi time stamping protocol was not working as it should in Logic 9. But i can see now that in fact this latency may well have been down to the track delay bug, as higher numbered track delays were placed at the top of my arrange page. I'll be able to confirm this in the next few days. The method i've been using for offsetting midi latency for individual synths has been the same since early Atari/ Notator days, where track delays were used to "advance the tracks of slow reacting synths to match the quicker ones". ie negative delays advance the midi track, while pluses will delay the track. This all worked fine when i was using an external mixer, but now things have changed since i'm working with audio interfaces. At the moment i no longer use an external mixer, but have all my synths plugged directly into my Motu audio interface's inputs, and these channels are routed to Logics mixer. I also only monitor through Logic using the "software monitoring" option. It seems this monitoring method totally changes things, and makes my track delay method inaccurate. I'm still confused by the second part of your reply Tom. Having a difficult time trying to get my head around what you've written. So using the "software monitoring" option, where are my hardware synths heard? And therefore from where must they be recorded? This is confusing me still..... Video 40 and much more coffee is needed I think.... Once again many thanks for taking the time to reply Tom. Appreciate your help.
    Reply
  • Dydavee
    Posts: 16
    Joined: Sep 19th, 2007, 07:47
    Re: Q for Tom Rice. Midi latency setup help? (or anyone else hel
    Hi Jauqq, What I mean when I say that MIDI instruments should be recorded from the point at which they are heard, is that you should record them from the end of their output/monitoring path. When monitoring through an external mixer, this means recording from the mixer's output. If the mixer is analog, which introduces no latency to the monitoring path, then you could record directly from the synth's output instead. So your current method works perfectly in this scenario. When monitoring through Logic's mixer, though, the end of the instrument's output/monitoring path is at the audio interface's output, so you should record from there (using the loopback method). Alternatively, you can fake a loopback recording by using the bus method outlined in the videos. Let me know if this, the coffee and video 40 haven't cleared things up for you. Tom
    Reply
  • Jauqq
    Posts: 7
    Joined: Jul 27th, 2006, 10:17
    Re: Q for Tom Rice. Midi latency setup help? (or anyone else hel
    Hi Tom, I understand it now, and no caffeine overdose was needed. Im just running through the "compensating for midi playback latency" test, but having a few technical problems trying to get things to work (mentioned in another thread BTW) Once everything has been set up correctly, will that then mean that for recording hardware synths i'd have to stick with recording through a buss? (or is this just a one off used for the calibration?)
    Reply
  • Dydavee
    Posts: 16
    Joined: Sep 19th, 2007, 07:47
    Re: Q for Tom Rice. Midi latency setup help? (or anyone else hel
    Hi Jauqq, Routing the signal through a bus won't impact on the monitoring path at all, so leave it in place for times when you want to print an external MIDI instrument to an audio track. If you don't use this, or the loopback method, your MIDI to audio recordings will be positioned out of sync. Tom
    Reply
  • Jauqq
    Posts: 7
    Joined: Jul 27th, 2006, 10:17
    Re: Q for Tom Rice. Midi latency setup help? (or anyone else hel
    Hi Tom, So leave recording through a buss in place for printing hardware synths to audio. Yes that's understood. Any ideas why i'm having problems with the marquee tool edits you describe in the video? I can see you select a single pixel marquee line across your audio regions, and then by pressing the right arrow key, the marquee line moves to the first transient across the regions. When i try that to press the right arrow key, i end up with a selection being made. Not a single marquee line. Im also having problem as you use the delete key to cut across the audio regions using the marquee tool. However when i hit delete, i end up erasing the audio file. Not use to audio editing.....
    Reply
  • Rounik Admin
    Posts: 8713
    Joined: Dec 16th, 2006, 08:13
    Re: Q for Tom Rice. Midi latency setup help? (or anyone else hel
    Hey Jauqq... for the audio edits with Marquee region/points I've answered you in this thread: [url]http://www.macprovideo.com/forum/logic/logic-pro-express&id=32698[/url] :)
    Reply
  • Jauqq
    Posts: 7
    Joined: Jul 27th, 2006, 10:17
    Re: Q for Tom Rice. Midi latency setup help? (or anyone else hel
    Hi Tom, One other quick question if i may please...? Should i be calibrating ALL external sound sources which are triggered from Logic, and not just midi synths? For example, I use an AU plug-in on a software instrument track called Silent Way which sends CV pulses from an audio interface's outputs to trigger external hardware mono synths. Recording this to Logic via the (incorrect) audio interface to Logic channel strip input method gives sample accurate, bang on the grid audio recordings. So surely that is exempt from the calibration set-up? Thanks, John
    Reply
  • Dydavee
    Posts: 16
    Joined: Sep 19th, 2007, 07:47
    Re: Q for Tom Rice. Midi latency setup help? (or anyone else hel
    Hi John, For a quick question, you sure chose a curly one! I don't have the Silent Way plug-in, or an analog synth to test this on, but having given some thought to the signal paths you are dealing with, I think this is what's going on ... Let's think about what would happen if you had a MIDI note on beat 1 of an External MIDI track feeding your synth, and a control message on beat 1 of the Silent Way's Software Instrument track also feeding your synth. Assuming you've calibrated your external MIDI track and output delays as outlined in the videos, the MIDI note will be sent before the playhead reaches beat 1, so that after passing through the synth and audio interface's input, it enters Logic's mixer exactly when the playhead reaches beat 1. At this point, the control message from the Software Instrument track will trigger Silent Way to produce an audio signal. This audio signal, and the already-triggered note from the external synth, then both make their way through the audio output path, resulting in them both arriving at the audio interface's output at exactly the same time. The note from the synth is then heard from your speakers, while the signal from Silent Way travels to a VC input on the synth. Assuming that the reaction time to the VC signal is zero, the effect of the of this signal won't be heard until the resulting output from the synth travels back in and out of Logic's mixer, to the speakers. In other words, the VC signal from Silent Way will be heard late, by the system's roundtrip latency amount, while the notes triggered from the External MIDI track will be heard spot-on in sync with the grid. So the answer to your question is to still record your synth to audio via the bus or loopback methods, so that what is recorded matches exactly what was heard. Sure, your VC will not be aligned to the grid, but it will be no different to how you were hearing it before printing to audio. (Incidentally, one way to check sync between the recording and the original is to listen to both together. If sync is tight you will hear mainly phasing, provided there aren't too many random LFOs, etc in the sound. If sync is off, you will hear flamming - or at least very little phasing.) If you do want to sync your VC signal more tightly to the grid, then you could try converting any track-based automation to region-based automation and setting the region delay to negatively compensate for the system's roundtrip latency (as displayed in the Audio Preferences). Like I said though, I can't test this, so I'm just making an educated guess. So try it out and see. As a side note, I don't think this sync problem will exist when monitoring external synths through an external mixer, as any MIDI track and output delays will be set to match the output latency of the audio path, meaning that the CV and external MIDI signal should arrive to trigger the synth at the same time.) Hope this is clear enough for, as it's definitely a brain twister! Tom
    Reply
You must be logged in to post a reply
Recent Threads
Recent Posts
Feedback
Course Advisor
Don't Know Where To Start?
Ask A Course Advisor
Ask Us!
Copy the link below and paste it into an email, forum, or Facebook to share this with your friends.
Make money when you share our links
Become a macProVideo.com Affiliate!
The current affiliate rate is: 50%
Classes Start Next Week!
Live 8-week Online Certification Classes for: