Live's Export MIDI File resolution is 96 ppqn.

[tomjordan]

NOTE: This message has been updated to show some new developments..
namely, that when using the MIDI file export command in Live, it outputs a MIDI file at 96 ppqn.

At first I thought that the resolution in general was 96, but gladly, I was wrong..

This thread shows the original messages and how we came to find out that it was only the MIDI file export , and not the playback or record resolution.

OLD VERSION STARTS HERE...


I've just finished some tests...

It seems like Live's resolution is 96 ppqn.

Compared with Logic's 960 ppqn, Logic has 10x more resolution.

I generated a MIDI file from my live coding environment, where there is a note on every tick, at 960 ppqn.

I imported that into Logic, looked at the event list, and there was a note on every tick.. 1 1 1 1 , 1 1 1 2, 1 1 1 3 ...

The file has 25 notes, ranging from 1 1 1 1 - 1 1 1 25, with every tick represented in between.

I imported the same original file into Ableton, then exported it back out from Ableton, which I then brought up in Logic.

The file now has many repeated values, and only changes twice

The values of the ticks are now: 10 notes at tick value 1 1 1 1, 10 notes at tick value 1 1 1 11, and 10 notes at tick value 1 1 1 21, for a total of 25 notes.

To check this, I then generated the same MIDI file from my environment, but now with a ppqn of 96.

I brought it up in Logic, and its tick values matched exactly to the file that was exported from Ableton.

Since the MIDI file standard has the capability of using a wide range of ppqn,
I can only assume that Ableton's resolution is 96 ppqn, since it wouldn't make sense to export it at a lesser rate.

To me, this explains why there is no event list with tick values in Live, a feature that I've longed for, and why the ppqn is not advertised, or widely known.

I sincerely hope that they add a feature where you can choose between the standard resolution at 96,
or opt for a finer resolution at 480 or 960 (whichever they choose to use), with a possible decrease in available processing power.

I can see why they would want to have it at 96, since it makes the MIDI very lightweight, and a much less drain on the resources,
but for a professional workstation, it is not acceptable.

There are good reasons why Logic is at 960 and other professional DAWs don't shirk this responsibility.

[TRS80]

If this is true I"m surprised and disappointed.

Hell, back in the early 90's Opcode Studio Vision had a 480 resolution. We used to adjust the feel of a track by ticking up or back a tick or two. High MIDI resolution is a big deal.

[tomjordan]

I know, I used to use Studio Vision too..
I thought that Live's rate would be 480, or at the least 240, but I had to check it because my sequences were getting a different feel in Live than in Logic.. , more "square" if that makes sense..
I'm afraid that it is 96. Anyone with Logic can repeat the test by making a sequence in Logic with notes at each tick, exporting it to Albeton, then exporting it back to Logic and comparing it with the original sequence.

[kb420]

How can that be possible if I can go to the grid in the midi editor and make the resolution 1/16834 and draw notes in?

Let me say this again, 1/16834 which works out to roughly 4208.5 PPQN. You can draw notes in at this resolution. As far as being able to actually record incoming midi data at that resolution, I think the limitations of latency and midi itself would prevent that from being possible.

[Angstrom]

I've just finished some tests...
Ableton's MIDI resolution is 96 ppqn.

you seem quite certain of this.
Of course, it would be quite embarrassing to be so certain if it turned out that this 'fact' was in fact incorrect.

In cases when I make a bold assumptions which seem to go against sense I normally give myself some wriggle room, just in case the error is with me, rather than with the app.

Whenever I have boldly claimed : "live does not recognize pitchbend correctly" I usually find I have egg on my face pretty soon after. And an omelette hat to match.

I would suggest to you that you have made a mistake

[tomjordan]

If I found out that I was wrong about this I'd be far from embarrassed, I'd be ecstatic.

Please somebody prove me wrong !

You may be able to draw notes in at 4208.xxx ppqn, but it probably rounds it off, there is no way you could tell just by hearing at that resolution, anyways.

[kb420]

Did you read my earlier post?

[TRS80]

Great to find someone who remembers Opcode!! I was devastated when it was discontinued because I learned that program inside and out. It was a shame how that company got bought and mothballed by Gibson. Great group of programmers behind that software. But I digress.

I've found a way to deal with the low MIDI rez. In session mode, enable the "track delay" feature. Then you can push the track forwards and backwards in milliseconds. Or, if Live has good rez, at least this technique substitutes for the "tick" feature I'm used to.

Still, the lower MIDI rez of Live is a bummer, and a reason why I'll probably end up shelling out $$$ for Logic Pro. I don't think I'll be able to quit Live though, so I'll have to maintain upgrades to both programs.

Yes I do hope tomjordan is wrong too. But Tom I love that you know Opcode

[stringtapper]

I think you're mistaken. You have Max for Live. Make a [transport] object and monitor the PPQ out the fourth outlet and you will see a resolution of 480.


Angstrom's blue omelet hat:

[tomjordan]

(monitor

think you're mistaken. You have Max for Live. Make a [transport] object and monitor the PPQ out the fourth outlet and you will see a resolution of 480.

Thanks for the tip, I was thinking of doing something like that..
I have been thinking up ways to get around this by playing sequences directly from Max, and bypassing recording them on a Live MIDI track. From that respect, I'd bet that you can get a finer resolution.

[ethios4]

I dont' know enough to participate in this discussion, but would these results also apply to data sent to internal plugins? Or is this just having to do with external midi?

[Angstrom]

Angstrom's blue omelet hat:

oh ffs

[Angstrom]

You may be able to draw notes in at 4208.xxx ppqn, but it probably rounds it off, there is no way you could tell just by hearing at that resolution, anyways.

It's possible that there may be an issue with the midi export, if internal tests can easily give higher resolution, but exported midi is noticeably low-res.
My basis for thinking that: I made a clip with thousands of notes in, exported it as midi and dragged that exported clip back into the live - the imported version of the midi clip is low res.

I'll try turning midi note thinning off in options.txt and see what happens, it may be thinning in the wrong place at the wrong time

[stringtapper]

Angstrom's blue omelet hat:

oh ffs

Sorry man had to do it

May the age of wibble music and interesting hats never be forgot!

[tomjordan]

All I know is I imported a sequence into Ableton with a 960 ppqn, then exported it back out, and now it is at a 96 ppqn resolution.
That much can be repeatedly proven, if you read over the results that I got.

Now you could say that it may have something to do with the importing / exporting function in Ableton, but from a programmer's standpoint, that would be highly illogical since you can easily set any ppqn of a MIDI file with just an integer value.

Why would you have a higher resolution during normal operation, then export it at a lower resolution, or import one and change it's resolution to a lower rate when you supposedly can accommodate it?

There is (conveniently) no way to check the tick value in Live, so you have to rely on importing and exporting.

Another test could be done that generates a fine-resolution sequence of extremely short impulse bursts in Max and recorded both as MIDI and as audio in Live.
Do this at a very slow tempo.
Then print the MIDI track to another audio track. See if the space between the bursts are the same, and if they mathematically match the projected value(s).
It couldn't be done at single ticks, since the duration can only handle single ticks.
Place an impulse every two ticks with a duration of one tick.