jump to beginning show previous Showing Logic-users Thread 74613 of 105812 show next jump to end

Forum Index | Read LUG: Policy/Rules Messages Threads Digests | Post New Message | Search!

From: Vincent Kenis <vincent.k@...>
Date: Tue, 2 Dec 2003 at 4:46:01 PM
Subject: Re: [LAM] softsynth internal latency is OK
Message #151312
>The EXS24 in fact is not sample accurate. If you play 2 kick >drums with each other simultaeneously via midi to the EXS24 >you will hear flanging of course, because MIDI can only transmit one event at a time...
Viewed 229 times, 1 reply, 2 messages in thread. Reply to this message.
From: "John Pitcairn" <johnp@...>
Date: Tue, 2 Dec 2003 at 5:25:37 PM
Subject: Re: [LAM] softsynth internal latency is OK
Message #151316
This is a reply to #151312.
>The EXS24 in fact is not sample accurate. If you play 2 kick >drums with each other simultaeneously via midi to the EXS24 >you will hear flanging It's not that the EXS isn't sample accurate on playback - it is. It's just that there's no way any softsynth can possible be sample accurate for realtime midi input - how can it possibly compensate without knowing in advance when you're going to send it a note? Additionally, there is always a little jitter in the midi system, and audio buffer cycling will add some randomness to the response time, the result being that the softsynth response will be a little different for every incoming realtime midi note. There's an important consequence of this for softsynths in Logic: because realtime midi input is not and cannot be sample-accurate, and because playback of the midi recording _is_ always compensated to be sample accurate, the timing you hear on playback _is not_ what you heard while recording the midi. Playback will be on average a little early compared to what you heard when you played in realtime. If the musician is (naturally) compensating for the perceived realtime softsynth latency by playing a little early, so what you hear in realtime is in time, when you record that to midi and play it back, it'll play back a little ahead of time (by the audio buffer size). Midi jitter and audio buffer cycling do have a randomizing effect on the input, so it's masked to some extent (because no musician can compensate for the randomness), but it could certainly be improved by _not_ compensating the playback for realtime- recorded softsynth midi. Personally, I'd very much like to be able to turn the "sample-accurate" compensation for Audio Instruments off on a track-by-track basis. It makes sense for step-entered, scored, quantized, pencilled-in, or imported midi, but not for realtime-recorded midi. John Pitcairn
Viewed 216 times, 0 replies, 2 messages in thread. Reply to this message.

Forum Index | Read LUG: Policy/Rules Messages Threads Digests | Post New Message | Search!

Forum Index | Read LUG: Policy/Rules Messages Threads Digests | Post New Message | Search!


© 1994-2008, All Rights Reserved.