Tempo Map When There's No Tempo

[bayswater]

This comes under the "etc" part of this forum.

I have a piece I'll be adding tracks to. I usually start this sort of project by setting up a tempo map so I can get a click track, see things in an organized way in the MIDI editor, etc.

This piece has several passages where there is no discernible tempo. So the various tool in DP for setting up tempo maps do nothing.
I've dealt with this by just making up a tempo that spans the passages of interest. I also thought about splitting up the sequence so each of the new sequences starts where a tempo is apparent but this gets complicated to get the parts to follow each other properly.

It's all manageable and I can get this done with arbitrary tempos, but I was wondering how this situation would normally be handled in DP by those who do a lot of arranging and composition in pieces with passages without definable tempos.

[ed belknap]

I'm curious to see if anyone chimes in with suggestions, because I wind up using DP for recording a lot of music that doesn't easily conform to grid-based conventions. I'd especially be interested in ways to hide the bars/beats markers in the audio editing ("Sequence"?) view, or replace them with absolute time markers.

[mhschmieder]

As it turns out, I am dealing with this same challenge right now, but chose to approach it in an entirely different way, as an experiment.

I was originally supposed to be in-person for the sessions with the band I had during my brief five-month residency in Raleigh NC, but due to a family emergency I couldn't drive down there on the one and only weekend where everyone was available, so I am now overdubbing bass during production work. As I had worked on over half the material with them earlier this year anyway, there's an advantage in that I already know how they play.

Anyway, I decided to skip the tempo map and switch to a strictly time-based view of the tracks and place markers where there are major transitions and funny timing. In most cases, the inconsistent tempo isn't sloppiness but is what you might expect in an orchestral project, with sudden transitions and sometimes a buildup. It's very organic, but only a couple of actual bloopers that I decided to patch up a bit.

I don't know if I'll take this approach on other projects that are delivered to me as two-channel source recordings that were done without any tempo map or click (whether constant or variable), and it's a lot of work either way, but as I made a decision to only play live parts (vs. MIDI and/or keyboard triggered sample libraries and VI's) for the extra track or two per song, using my ears and carefully watching the cue markers has worked quite well.

I was torn on this decision as most of my stuff is still packed in PODS and so I was tempted to do a tempo map and at least add placeholder tracks with sample libraries for instruments I don't have at hand at the moment. If it was an orchestral project where I'd be adding a LOT of parts, and especially if I wasn't going to play them live (whether direct from a keyboard or as MIDI to render later), I may feel that a tempo map is unavoidable.

As to how to do it, usually I hear people saying it's best to tap the tempo in real time during playback, multiple tries if necessary. I've generally been more successful with this approach than trying to hand-adjust tempo changes in the conductor track. But again, it does depend on the material, genre, instrumentation, and whether the tempo was deliberately crafted as-is or the material should be tighter but maybe has to serve as the metronome.

[stubbsonic]

I've done a fair amount of tempo map creation using Tap Tempo-- which was made a little more challenging when DP's tap-tempo behavior was quite sloppy. I'd have to cue it up several times as the tracks might not even play for several bars!

If having the tempo bars is just for a vague structure to make editing easier-- then tapping tempo is a pretty painless way to go.

I remember one time, I created a MIDI track for my taps-- which I was able to edit pretty easily by dragging MIDI notes of my tapped beat forward or backward in relation to audio transients, etc. I nudged things to include a count-off, and I exported the chunk as an SMF. Then I used a 2nd MIDI sequencer to play those taps into DP as tap tempo. IIRC, it worked, with only a couple adjustments to get it all aligned. It was a long time ago. I think the 2nd sequencer I used was a K2000.

[bayswater]

I've done a fair amount of tempo map creation using Tap Tempo-- which was made a little more challenging when DP's tap-tempo behavior was quite sloppy. I'd have to cue it up several times as the tracks might not even play for several bars!

If having the tempo bars is just for a vague structure to make editing easier-- then tapping tempo is a pretty painless way to go.

I remember one time, I created a MIDI track for my taps-- which I was able to edit pretty easily by dragging MIDI notes of my tapped beat forward or backward in relation to audio transients, etc. I nudged things to include a count-off, and I exported the chunk as an SMF. Then I used a 2nd MIDI sequencer to play those taps into DP as tap tempo. IIRC, it worked, with only a couple adjustments to get it all aligned. It was a long time ago. I think the 2nd sequencer I used was a K2000.

Yes, this is pretty much what I’ve done. For the phrases where no apparent tempo exists (someone noodling with random notes and pauses of random length, I made up a tempo that fills the space and lines up with at least some of the notes being played at an upbeat or downbeat.

I also recorded a “click track” in audio then analysed it for tempo and applied the tempo to the sequence. That approach breaks down when the music gets into the difficult passages, and then DP seems unable to come up with a sensible map as the music speeds up or slows down, or makes an abrupt transition from an undefined tempo to a regular beat.

I could also make audio click tracks with emphasized downbeats, etc, and use only that and ignore tempo. But in cases where you might want to use MIDI quantizing, or some of the other region commands, again you’re lost.

I was also thinking there might be a more systematic way to do this that translates better into a score. The way I’ve done it, it’s difficult to avoid extreme highs and lows in the tempo, and avoid partial bars if you want everything pick up properly on a downbeat when the music become rhythmically organized again. There are a lot a ways around this, but they’re all time consuming and don’t work in every situation. Ideally you could select a range and tell DP the length of the range (e.g. in msecs) and that tell there is no tempo for that passage. But AFAIK, there is always a tempo in effect that will effect MIDI timing.

Another approach would be to always print MIDI tracks and lock them so nothing is dependent on tempo. But that’s time consuming and clunky too. You’d end up quantizing one region against the internal tempo of another. Yet another rathole.

I also considered turning this in to a series of sequences, so sections with a well define tempo map are standalone sequences, and the passages with no discernible tempo are also standalone sequences, and then use the functions to link sequences. But my mind still works like an 8 track tape deck.

For now, I’ve solved the problem for the piece at hand, but any scores I print will look goofy and will probably confuse people. No doubt this will come up again, and I’m always interested in any thoughts people have on approaches.

[stubbsonic]

Two other points, obliquely (or obtusely) related:

If playback is the ultimate goal, you can ignore tempo and just choose some specific tempo (like the oddly universal 120 BPM), and perform your tracks with care & skill.

If quantizing is required for nice playback, then you'll have to painstakingly get a good tempo map that gets all the beats/subdivisions lined up as needed. Depending on the structure, you may or many not need to bother with time-signatures-- depending on the nature of the changes.

If it would be helpful to have a visual structure with meter, time-signature, bars, etc. Then a rough tempo map is useful.

If you are going to create a printed score, it may be useful to create two versions: a playback chunk which does what you need sonically, and a for-notation version, that does what you need visually.

[bayswater]

If you are going to create a printed score, it may be useful to create two versions: a playback chunk which does what you need sonically, and a for-notation version, that does what you need visually.

Good point.

[svobodomo]

How about that?
https://www.youtube.com/watch?v=KlWZ4YjO-H0

[bayswater]

How about that?
https://www.youtube.com/watch?v=KlWZ4YjO-H0

That’s fine, but the question was about methods for establishing a tempo map when sections of the piece have no recognizable tempo that you can tap to.

[pencilina]

Hi. I work with rubato material all the time and use the technique described in this post which I updated recently:

viewtopic.php?p=505931#p505931


Hope it helps

[bayswater]

Yes, that’s pretty much what I do with things not recorded to a specific beat. But I was talking about a piece where there are sections that have nothing going on that you can use to tap out a tempo. The only alternative I’ve used is to measure the length of the’ gap and make up and arbitrary tempo that spans it, and takes things to the next identifiable part with some sort of rhythm.

The other thing I tried is to split up the piece into multiple sequences with the “musical” sections separated out. Then you can chain the sequences and it doesn’t matter what the tempo is for the sections with no musical structure.

[pencilina]

Right.... at 120 bpm a 1/2 note = 1 sec. In the approach I linked to I'd put locked 1/4s at 120 in those sections of the tapped track before tempo mapping. That said, I can usually feel/interpret some sort of pulse in most music that feels natural (which I would tap out) through those sections.