Tempo edit in MIDI recording

[Sueg]

I have created a system track. Now I need to edit the tempo for teaching and performance purposes. How do I do this?

[Jokeyman123]

What other information did you record in the system track? The system track is the only track that will record and store your original tempo setting, or settings. You cannot "overdub" another tempo setting within the same recording if that's what you want to do, and record it without deleting whatever other information you have already recorded in that system track, including the tempo, an unfortunate limitation of the PX560, and I have found no way to change that. I now always set my tempo first, record it to a blank system track (if I am not recording the auto-accompaniment tracks to that track at the same time) then I record my other parts on the individual tracks 1-16.  You can record different tempos in that system track as it plays by using the tempo controls, with or without recording anything else in the system track. Again, you would have to record the system track in one "pass" and use the other tracks for individual instrument parts, which would follow your tempo changes. I have tested this-and it will work. The metronome will follow these tempo changes, and if you record auto-accompaniment tracks at the same time, the tempo will change with those rhythms. You need 2 hands-one to play your chords for accompaniment, and your other hand and finger to change the tempo setting from the midi recorder screen while you are playing your chords for accompaniment triggering. Hope this is what you wanted to do! 

[Dan Kaufeld]

I also wanted to change the tempo after completing a MIDI recording.  I was able to do it with the following methodology.  It is a little bit of a pain, but worked fine for me.

→ Applies to System Track only!  I strongly suggest making a copy of your original song so that that you don't lose it if something goes wrong!

1. Insert a measure at beginning of song (page EN-63 in the User's Guide).

2. Use Punch-in recording (page EN-58). Select “Punch In Sync”.

3. Set tempo to desired setting

4. Start punch-in recording, then immediately stop before the end of the first measure.

5. Delete the first measure you previously added (page EN-64).

The new tempo will be retained!

 

I have only verified this methodology for changing the tempo for an entire song, from the beginning, with only a system track, but I think it should work with multiple tracks.  Obviously, the system track is what must be used to effect the tempo change, as this is where the tempo data is stored.

 

I suspect this method or some variation thereof could be used at other points in the system track for tempo changes within a song.  Insert a measure the measure before the desired tempo change is desired. Set up to "Punch in Sync".  Play song to the inserted measure, pause and change the tempo, then punch in and immediately stop.  Then the inserted measure should be able to be deleted and the changed tempo should apply thereafter.  I haven't tested this yet. When I have, I will attempt to update this post or add a new one.

[Jokeyman123]

Good idea-never thought of punch-in/out on system track-didn't think this would work. I will try the tempo punch-in/out in the middle of a song too. I was thinking that any attempt to re-record any part of the system track would also delete other data from the rest of that track. If this works in the middle of a song-could also be done for other data-like volume, pan, maybe any changes that the system track records originally. Will have to try this out. I'm thinking any changes that can be done with the mixer with punch-in/out-will these also be recorded without affecting the rest of the system track? It should work. Very nice if it does. 

[Dan Kaufeld]

l will be interested in what you find, Jokeyman123.  As far as I could tell, the only thing affected by the tempo change in the inserted first measure was the tempo change, the other system track data seemed to be unaffected.  I did a little experimenting today in the middle of the song and did not achieve success, but the particular song I used for the test was not synced up with the system beat because of ritards I included as I played the piece for the original system track recording. This was not a problem for an inserted measure at the very beginning, but further into the song the measures I had recorded were out of sync with the recorder's beats, so the inserted measure was misaligned relative to the song.  The change in tempo was picked up in the inserted measure OK, but when I deleted the inserted measure, some originally recorded notes were removed or mangled (not a big surprise) and the tempo change was not retained (disappointing!).   I'll experiment some more and try a song that is recorded on tempo, and see if that makes a difference. It would be great if the PX-560 had an "overdub" record option... I have this with a "legacy" Ensoniq KT-88 keyboard sequencer (along with step editing). Maybe with overdub you could just change the tempo at will throughout the song.  I suppose that would consume a lot of extra real-time processing power and might be beyond realization, though.  With the Ensoniq, each song was composed of segments.  Each segment could have its own tempo, and it was easy to modify.

 

I actually achieved success in the past in changing the tempo in the middle of a previously recorded song on the PX-560, but seem to have misplaced the notes I took on how I did it. 😖 So it is a worthy challenge to rediscover the methodology!  The PX-560 is a great keyboard, I really like it, and it seems like there is so much function in there that frequently there are non-obvious ways to do things if you experiment enough.  It is great to have this forum to share discoveries.

[Jokeyman123]

Yes, I have posted here a few times-an overdub or cut and paste function would be helpful with the 560. Haven't had a chance to follow up on your findings but will, I record and play quite a bit with the 560. Also have much experience with the Ensoniqs-from the SQ-1, then the KS-32, the MR-76 and 61 and TS-12-great keyboards in their own ways-I've kept my TS-12 which I felt was the best of all of those-still a formidable beast which is now next to my Casios, as I've been revisiting the older models to compare to these newest Casios.

 

I don't know if you've seen my other post about midi clock/sync with other keys and modules-but the PX560 I've just discovered-although it has no midi clock sync-keeps its tempo accurately enough to manually sync with just about anything i've thrown at it-this is a real plus since older internal tempo clocks were not always accurate enough to do this and needed midi clock sync. I synced the PX560 song  recorder tracks with my Yamaha QY's and 2 Roland digital recorders-kept in exact sync for over five minutes with all of this. As long as i record using the tempo settings for the other equipment, it works. Never tried it until now because I never thought it would work. And since the PX song recorder will not record from external midi controllers-I play Yamaha midi wind controllers and a malletkat-now gives me a lot more flexibility in recording to other devices, and syncing with PX560 tracks.  Just thought I'd pass this on.

[Dan Kaufeld]

Thanks for the info.  Amazing that the tempo clocks can stay in sync that well. 

[Dan Kaufeld]

Update on changing tempo in existing MIDI recordings - a way do it within the body of the recording, and a nicer way to reset the starting tempo.  Note - I made some corrections in here on 10/30 due to some discoveries I made while experimenting.  Then changes are denoted by blue text.  I have further discussion of what I found in my 10/30 post.

 

Within the body of the recording - use Auto Punch-in recording on an existing measure in the System Track, on a very short interval.  (Turns out that inserting and deleting measures does not behave like it does at the beginning of a MIDI recording...  in the body of the recording, if you delete the measure in which you make the tempo change, you lose the tempo change!)

 

  → Applies to System Track only!  I strongly suggest making a copy of your original song so that that you don't lose it if something goes wrong!

1. Touch "Record" mode. Use "Record Type" “Punch In A-B”.  (EN-59 in the user guide).
2. Choose the measure just before where you want the tempo change.
3. On the last beat of that measure set A to the last beat, tick 95.  Set B to the last beat, tick 96.*
4. Use fast forward to navigate to the beginning of that measure.  Set the tempo to what you want the new one to be.
5. Touch the "play" arrow.  With this, the tempo change should be implemented, without disrupting what you had played before, except for the last beat tick 95-96 range, which will be deleted. Go back and check the results.  The new tempo will continue to the end of the recording, unless you change it again later in the recording!

       * I think the user guide makes setting A and B harder than it needs to be.  If you start out by hitting "record" and selecting "Rec Type" "Punch In A-B" you can then press "set A" and set the measure, beat, and ticks by touching these and using the finger dial or "Yes" / "No" buttons to adjust them.  Then press "set B", which will already have the A values preloaded... all you have to do is press "Tick" and increment the B ticks from 95 to 96.  There is a lot more flexibility to the order in which you can set these parameters than the User Guide indicates.

 

Caveats

1. This methodology requires your MIDI recording to have close adherence to the system beats/metronome. If you're a bit of sloppy player (like I am), make sure you do not have critical notes for the succeeding measure inadvertently starting at the end of the measure you selected for the tempo change, or you may get unhappy results... like losing those notes or mangling them.  You might be able to move the tempo change autopunch A-B interval back a little on a re-do if this happens. (Depends on what you have going on in that measure.  The further up you move from the end, the more likely you are to "mis-pace" other notes!)  The previous struck out text turns out not to be the case. When you change the tempo according to the slightly modified instructions above, the tempo change will start at beat 1 in the measure you selected, even though you set it later in the measure.  For this reason, make the change in the measure where you want the changed tempo, not the measure before.  In any case,  I recommend quantizing the track (EN-64) before doing the tempo change procedure above.
2. Theoretically, this methodology causes a little bit of a gap in whatever is still playing at the end of the measure -- 1 tick wide.  I used a simplistic piano test recording to test my methodology.  I found that when I had sustained piano notes over the measure boundary, the gap was undetectable.  But this was very simplistic, one super-simple piano piece with melody and chords on only the system track.  I will be experimenting with multiple tracks and instruments in the (probably not near) future, and make updates here on what I find.  My test case was a radical tempo change from 200 to 30.  I thought the 30 might accentuate the gap, but I didn't notice it.

 

A nicer way to reset the starting tempo - still requires measure insert and delete, though! This seems to work just fine, I made several different tempo changes using this method and they always worked fine, I did not notice any problems.

 

→ Applies to System Track only!  I strongly suggest making a copy of your original song so that that you don't lose it if something goes wrong!

1. Insert a measure at beginning of your MIDI recording (page EN-63 in the User's Guide).
2. Go to "Record" mode. Use "Record Type" “Punch In A-B”.  Set A to anything in the first measure.  I used beat 1.  Set B to something later in the first measure.  I used beat 2.
3. Set tempo to desired setting 
4. Press "play" arrow.  With this, the tempo change should be implemented.
5. Delete the first measure you previously added (page EN-64).  Check the result.

The new tempo will be retained!

 

Edited October 30, 2020 by Dan Kaufeld
Corrected some statements due to later discoveries.

[Jokeyman123]

So deleting the first measure doesn't delete the tempo change? This means the system track has a "chase" function-somehow it is tagging the rest of the system track as keeping the same tempo. One would never figure that to be true. Is good to know.

 

Here's another experiment I haven't tried-and i don't think it should work based on the manual's info, but I'll try it.

 

If I copy the system track to another track-does it copy all its data-tempo, pan, effects send etc? Hmmm....and of course embedding a tempo change in the system track is pre-supposing it has no other musical content in it since you cannot overdub as i can with other modules and keyboards. So if the system track can be copied to another track...with its data intact. Wish I had more time to test all this, maybe tomorrow will be better.

[Dan Kaufeld]

Another update - CAUTION! with regard to updating tempo in the body of a MIDI recording!  10/30 updates in blue below:

 

This is with respect to the first method I posted yesterday.  I tried an experiment today to change the tempo back in one of my test cases.  It did not work.  I do not seem to be able to modify the changed tempo in the middle of a recording at all.  I've tried a few different approaches including the obvious one to just repeat the original "change" steps with a new tempo, but the modified tempo seems to be locked in.  So if you decide to use this, be aware that you can make the change, but it is a one-way street at this point!  Today I found that you can indeed change the tempo back, or to whatever new value you'd like.  It is not "locked in". The catch is that you have to start your "Punch in A-B" with beat 1, tick 1, since that is where the tempo change will start, regardless of where you made the change it in the measure.  Since this will erase the played data you have in that measure, you have to replay the notes in that measure during the punch-in if you want to retain them while resetting the tempo.  I know this works fine if you do the whole measure.  (beat 1 tick 1 to beat "last" tick 96.)  I did an experiment with a partial measure, with a note on beat 1, tick 2, and did the tempo change punch-in for beat 1, tick 1-2.  The tempo change worked OK but took out the note at beat 2.  More varied experiments might show where notes can be without getting cleared.  I may try these, but my feeling is that even if it works with the note on beat 1 tick 3, the gap is getting to the point where it may become audible, and the process becomes unmanageable.  It's easier and better to just replay that whole measure if you really need to re-change the tempo.  Another, likely easier (but more tedious) way to handle this might be to make a copy of the recording at each step so you can just go back to a version with the original tempo, and make whatever change you need based on that, rather than trying to reproduce what you originally played. 

 

Also, at this point I have not gotten to the point of testing this overall methodology with multiple tracks.  I will be doing that eventually, that's a key reason I am interested in this, I want to be able to do system track parameter edits on multiple track recordings.  I don't anticipate a problem with multiple tracks since only the system track controls several recording parameters, including tempo.  Some of these may be changed using the mixer.  But tempo, hence this effort.

 

The other method for changing the base tempo of the overall recording by inserting and deleting a new first measure, making the tempo change there, as described in the last part of my posting yesterday, does not have this problem.   I have verified that - you can successfully change it to different values, repeatedly.  But if you have modified a tempo in the middle of the piece using the first method, that value will be retained.  Your new base tempo will remain in effect until the measure where tempo was modified, and then will change to that tempo.

 

Jokeyman123 - I do not see a way to copy the system track to one of the other 16 recorder tracks, but I have copied the system track to another system track (new recording).  From what I can see (not that much!) the new system track appears to be identical with the original, the tempo changes and other data are retained.  It would be very interesting to know how the system track parameter data is stored.  Like you point out, it seems strange that you can change it in the inserted first measure and the change is retained when you erase the first measure.  And that behavior does not occur anywhere else in the system track, apparently.  I tried to determine if the mid-recording tempo "chase" data is stored at the measure, beat or tick level, I have not been able to determine this yet.  I was hoping "tick" level but my experiments so far have not isolated this. 

Edited October 30, 2020 by Dan Kaufeld
Corrected some statements due to later discoveries.

[Dan Kaufeld]

Tempo change within the body of a MIDI recording - experiment discussion

 

Long story short: I found that that you can use the Punch In A-B recording mode to change the tempo in the body of a previously recorded MIDI song.  If you do this, the beat change is retained at the measure level and takes effect at beat 1, tick 1, regardless of where in the measure you set A and B.  I found two ways to do this:

1. Re-record the entire measure with A=beat 1 tick one and B=beat "last" tick 96 (or possibly a subset of it) where you want the tempo changed.  This means you have to play whatever you had on the system track again, along with changing the tempo.  

2. Choose an A-B location where notes are not being played in the measure.  I chose the last beat, tick 95-96 for this as the likelihood of having something played there seemed to be lower.  (Might not be the case if you have a lot going on with rhythm, etc.)  With this method, you don't have to replay the measure, the original content will be preserved except for the A-B region you set to change the tempo.  I found that for the test sample I used (a very simple piano song), I could not hear the gap.

-->If you want to change the tempo again for this measure (and beyond), you can do it, but you have to use beat 1 tick 1 for A.  I found the most reliable way to do this was to do the whole measure as per #1 above.  It is conceivable that you could do a shorter A-B interval, but you have to start it at beat 1 tick 1.  As per one of my updated postings above, I tried A-B for beat 1, tick 1 and 2, and had previously inserted a note a tick 2, and the tempo change took effect, but the note at 2 was not retained. 

--> I recommend doing interim copies of your recording as you go through this process, especially one just before you do your tempo change, and use that copy if you aren't happy with the tempo change you made and want to go back to the original or change to something else.

 

Long story - this whole effort came about because tempo changes are common with most music, especially ritards and atempos.  There are various ways to handle this with the PX-560's MIDI recorder.  One is to initially put these into the system track.  If you are well-organized and coordinated, when you get to the targeted measures, you can adjust the tempo live during the initial recording and it will be recorded.  Doing it this way will change the system beat and metronome and you stay in sync.  Hopefully.  A second way is to ignore the metronome altogether and just play your piece the way you want it.  Inevitably the recording will be out of sync with the system beat, and the recording may sound just fine, but this method complicates later editing.  The third way I was working on was to record  the entire piece at one metronome setting, and then go back and edit the tempos afterwards.  This is somewhat clunky as the tempo change will be a step function by measure.  Sometimes this sounds OK depending on the musical content, other times it is too abrupt and does not sound like a nice smooth transition.  Still, I wanted to be able to do this, it is very useful to have this capability.

 

I originally thought that if I could sneak in a tempo change at the end of a measure, it would be set for the next measure.  And it seemed to work... until I listened and watched more closely and figured out what was happening.  The PX-560 adds to the confusion a bit... if you use the fast forward function and look at the tempo, measure by measure, when you set the tempo at the end of the measure like I did, and look at the first beat, it looks like the previous tempo was retained!  All well and good, it should change at the end of the measure!  But beat 2 shows the change! What's going on?  I tried to watch the tempo while the piece was playing instead of stepping through it - it went by too fast for me to catch.  This was at a tempo of 100, I guess I should try this with a very slow tempo, maybe then I could see better.  But instead I took a video of the display with my phone to I could step through the frames.  Lo and behold, the video showed the tempo change right at beat 1!  This was for a tempo change I made at tick 95-96 A-B setting at the last beat of the measure.  So it seems that the tempo data is stored at the first beat of a measure regardless of where in the measure you change it.  I guess I can guess why this would be -- the law of conservation of data and processor cycles.  Storing tempo for every tick would cause 384x the tempo data and processing vs storing it once per measure for a 4/4 time signature!

 

I did many other variations, I won't elaborate greatly on those, because I saw some things that I think I misunderstood. One I had a lot of hope for was inserting a measure and changing the timing in that, then deleting the measure.  Before the deletion, the tempo change is there, starting with the first beat of the inserted measure.  But after deletion, the tempo change is lost.  So this behavior is different than inserting for the very first measure.  It would be very interesting and helpful to know the algorithm the PX-560 uses to store the tempo data.  Also, a really nice function to have in the MIDI recorder would be an overdub recording capability for the system track.  That way you could focus on the tempo initially then add the rhythm track or other content later.  

[Jokeyman123]

So...unlike some built-in sequence recorders designed into keyboards i've worked with-that allow for more continuous gradual tempo changes-beat by beat for example-the PX560 will always need a one measure space to do this. You may be right about limited cpu cycles that cannot adjust finer calibrations for tempo changes than one measure-there is also no step recorder here and no midi clock-so you may be onto something here-the CPU may not be able to do all that in addition to all the other tasks it must process.  And yes, if you have a continuous recording on the system track-punch in/out does no good there. Worth knowing how to do this-I prefer doing my composing/recording without software if possible, so i study and experiment with these fundamental capabilities-or lack thereof! I think the MZ-X500 has some of these abilities-a shame that the 560 doesn't share those same features  Might be why there can be no added functionality through firmware updates-the CPU might not be up to any additional tasks such as you are testing out. And why there is no overdubbing capability for the system track. Typical Casio design that way..but would still be helpful.

[Jokeyman123]

Something else just flashed into my little brain cell-I noticed originally when I first started recording with the PX560-and this threw me off quite a bit until I understood about the system track and tempo-how necessary it really is to establish and record your tempo initially-when i tried to record on other tracks-without using the tempo (system) track at all-I could set the tempo before i recorded any other track-and it would not stay constant-I might be "misremembering" this-but I seemed to recall the tempo would drift-as if the internal sequencer clock was not stable enough to hold the tempo unless I recorded the tempo to the system track first. I will check this again, it's been awhile as I haven't tried recording that way since-a limitation IMO because if the clock were steady enough-I wouldn't need the system track for this at all-and I could keep the system track empty until i had all my other tracks recorded, then use the system track for recording all the other real-time changes I might want to perform-such as panning, variations in dynamics, tempo anything that can be changed with the mixer. Your work has me convinced-with punch-in/out on that track-alot of changes (I think) can be recorded on the same system track-at different measures of course-but still presents some very interesting possibilities, if the CPU is up to it. Just need more time.....!

[Dan Kaufeld]

Very interesting observations. I don't know if the following is is related to what you saw with sequencer clock stability, but I have observed that there seems to be two metronome tempo entities on the PX-560 - the one activated by a named button on the control panel, and the one that shows up in the MIDI recorder.  I originally thought that these two were the same, but I found that if I used the named button, it was generally out of sync with the recorder.  My impression was that they ran at the same rate, one just seemed to be out of phase with the other.  But I didn't scientifically investigate this by running them together for a long time interval. I ended up just relying on the MIDI recorder metronone for recordings.                                                                             

[Jokeyman123]

So do I. this is the clock that seems to stay in sync manually with other midi devices. I never used the other one much at all but I'll check that too.

[Josh W]

On 5/27/2020 at 8:00 PM, Jokeyman123 said:

the system track is the only track that will record and store your original tempo setting ... I now always set my tempo first, record it to a blank system track (if I am not recording the auto-accompaniment tracks to that track at the same time) then I record my other parts on the individual tracks 1-16

 

Just wanted to say I came here for a related issue which is that I recorded on tracks 1 & 2 at 80bpm and every time I played back it reverted to the default 120. Went back and recorded a blank system track at 80 and problem solved! Thanks for the useful info.

[Jokeyman123]

There is alot of info here we had to deduce out for ourselves for the PX560. In fact there is now so much info-many times i just google a phrase to find my own posts, since i forget alot lately! What bothers me the most-I can't figure out how to do something with my Casios-I Google for an answer-and end up with my own post from-1, 2, 3 years ago!!! Good thing there are others here-with better memories than me!! and keeping tempos in sync with other midi devices-without midi clock, becomess even more important with other new technologies (I'm thinking bluetooth) that do not connect directly with a physical midi cable. Brings another question-how accurate are the newer PX-S series when trying to sync like this-manuallt without any connection-bluetooithe or other midi? Will make a separate post now about this.