I am sure that most here are very talented, and have a gift for sensing the way musical phrases relate to each other to create the tension and release that makes good music exciting. I am also sure that vitality flows out from most of us when we record ourselves playing our best instrument. But what do we do when keyboard is not our best instrument, and our MIDI tracks stand out of the mix for all the wrong reasons?
I have had years of practice at editing hand sequenced MIDI tracks to make them sound alive. Over those years, I developed some "rules of thumb." Although tedious, these step do make a reasonable difference:
Duration
Select all notes and change every duration to 85%.
Go to every note that precedes a rest (i.e., is immediately followed by silence) and quantise their ends to the exact point each rest should begin.
Change every note that should be staccato to 47% of its already-shortened-to-85% duration (which results in a total of 40%).
If you need mezzo-staccato, reduce those notes by 71% (for a total of 60%).
If you need, staccatissimo, reduce them by 24% (for a total of 20%) and also add 10 to their current velocity.
Slurs need legato playing to overlap the notes to form coherent phrases. Achieve this by selecting the all but the final note under each slur, then increase their duration by 129% (for a total of 110%).
Velocity
Slurs should give the effect of a series of notes being sounded by a single breath, as opposed to each note being started and stopped in succession. So, in addition to the overlapping already done by increasing duration of all but the ending notes, we should also decrease the velocity of all but the starting notes. This gives the effect of an envelope that attacks on the first note then decays through all the notes that follow on in the phrase. For example, if the notes all have velocity of 80, I would keep that velocity on the first note under the slur and reduce the remainder depending on how many notes follow on: if the slur covers two notes I would reduce the second velocity to 70; if the slur covers three notes I would reduce the second velocity to 73 and the third to 68; in the case of four or more notes I would select the range covering the note-ons of the entire slur then change velocity gradually from 80 to 65. I also did a lot of velocity tweaking to provide accenting to emphasize the down-beats and the back-beats.
Humanizing
This really should be called randomizing (as indeed it is called on a hand-full of MIDI editing applications). This is a cruel device that can taunt you with a positive result then quickly bite you with a result that is what I once heard called "Robo-Sloppy." Whenever I've used humanizing, I have always cautiously applied just a tiny amount, and found the result to be a tiny improvement. What used to happen next was "I will apply a greater amount for a greater improvement." However, every time the result was messy. Don't expect much from humanizing, because it can only make a tiny improvement.
How do you enliven your MIDI tracks and get them to breathe?
I have had years of practice at editing hand sequenced MIDI tracks to make them sound alive. Over those years, I developed some "rules of thumb." Although tedious, these step do make a reasonable difference:
Duration
Select all notes and change every duration to 85%.
Go to every note that precedes a rest (i.e., is immediately followed by silence) and quantise their ends to the exact point each rest should begin.
Change every note that should be staccato to 47% of its already-shortened-to-85% duration (which results in a total of 40%).
If you need mezzo-staccato, reduce those notes by 71% (for a total of 60%).
If you need, staccatissimo, reduce them by 24% (for a total of 20%) and also add 10 to their current velocity.
Slurs need legato playing to overlap the notes to form coherent phrases. Achieve this by selecting the all but the final note under each slur, then increase their duration by 129% (for a total of 110%).
Velocity
Slurs should give the effect of a series of notes being sounded by a single breath, as opposed to each note being started and stopped in succession. So, in addition to the overlapping already done by increasing duration of all but the ending notes, we should also decrease the velocity of all but the starting notes. This gives the effect of an envelope that attacks on the first note then decays through all the notes that follow on in the phrase. For example, if the notes all have velocity of 80, I would keep that velocity on the first note under the slur and reduce the remainder depending on how many notes follow on: if the slur covers two notes I would reduce the second velocity to 70; if the slur covers three notes I would reduce the second velocity to 73 and the third to 68; in the case of four or more notes I would select the range covering the note-ons of the entire slur then change velocity gradually from 80 to 65. I also did a lot of velocity tweaking to provide accenting to emphasize the down-beats and the back-beats.
Humanizing
This really should be called randomizing (as indeed it is called on a hand-full of MIDI editing applications). This is a cruel device that can taunt you with a positive result then quickly bite you with a result that is what I once heard called "Robo-Sloppy." Whenever I've used humanizing, I have always cautiously applied just a tiny amount, and found the result to be a tiny improvement. What used to happen next was "I will apply a greater amount for a greater improvement." However, every time the result was messy. Don't expect much from humanizing, because it can only make a tiny improvement.
How do you enliven your MIDI tracks and get them to breathe?
How to make your MIDI tracks breath.
Aucun commentaire:
Enregistrer un commentaire