Tempo Slider Bar

[Doug Kerr]

Hi, p,

Well I have no idea how I got screwed up before, but in fact I now confirm on E/W 5.0.0 the behavior you described.

In summary: Regardless of the measure in which the cursor is placed, the tempo slider will report the current embedded tempo at the beginning of that measure. If I move the slider, it changes the embedded tempo for that measure, and changes all other embedded tempos throughout the score by the same ratio as the change made here.

I am having great difficulty imagining how I could have believed to have gotten the other result in two separate test runs, but for now, the above is what I found.

The current manual clearly reflects this behavior.

Thanks for your report.

Best regards,

Doug

[polarbreeze]

If you save the file, any change will be saved; if you don't want this, you have the "reset" button at bottom of the tempo window.
I can't imagine something clearer or more efficient.

The confusion is that (a) the slider is calibrated in absolute values but it does not actually operate that way - it changes it by a ratio; and (b) the slider makes changes in places in the score that you can't see so you may not realize what it's doing.

BTW, I have nothing against adding a percentage slider, but we should be aware that percentage isn't perhaps the good parameter. Having sequenced many MIDI files, I know that a 10% change makes a great difference, that 5% is fairly noticeable, and that you will even notice 3%. Thus I'm not so fond of a slider always staying at the very bottom of its range...

I think that the ratio slider would have "zero" in the centre so you can go either faster or slower; and it would have a non-linear scale for the reason you mention.

[polarbreeze]

The confusion is that (a) the slider is calibrated in absolute values but it does not actually operate that way - it changes it by a ratio; and (b) the slider makes changes in places in the score that you can't see so you may not realize what it's doing.

But that's exactly the way you play when working a piece.

It depends - but that's not the point. All I'm saying is that if it means a ratio it should be labelled as a ratio but if it is labelled in absolute values it should make the changes in absolute values.

You decide to slow down at some place, and you play the piece. Then you decide to play the whole piece faster -> for the few measures suggested above, will you slow down at the same absolute tempo? I don't think so! You will adapt the slown to the new general tempo, therefore you will "make changes at places you can't see" when starting to play, and that's the way the tempo window works.

It depends. If I have an Allegro passage followed by an Andante passage I make the tempo decisions on those independently and it would be really annoying if when I re-think the speed of the Andante part the app decides to apply that same proportional change to the Allegro part that came before it.

Anyway, as I said above it depends on the task at hand and the volume of discussion in this thread confirms that the current implementation is confusing.

[polarbreeze]

It depends. If I have an Allegro passage followed by an Andante passage I make the tempo decisions on those independently

Personnally I would not do like that. Why? Because an audience hears a new tempo by contrast with the preceding one.
If you don't change the tempo of the second movement, it means that you're changing your mind about the tempo relationships between movements, which affects the balance of the whole piece, and that kind of artistic decision should certainly not be taken by a software.

My point exactly. I want to make those artistic decisions myself and I don't want Encore changing a tempo that I've already decided is the tempo I want.

[Doug Kerr]

Hi, p,

All I'm saying is that if it means a ratio it should be labelled as a ratio but if it is labelled in absolute values it should make the changes in absolute values.

And indeed it works in terms of absolute value for what it does directly, and then tempos in other places are changed proportionately. We can think of the result in terms of percentage of the pre-existing values if we wish.

But a "real time" slider isn't suited to being "calibrated" in terms of ratio (or percentage of the current value).

Here's why. If we wanted the slider to work in terms of ratio, imagine this scenario:

The tempo at bar 1 is 120. Put the cursor there. Would the slider be at 100%? It would have to be for this to work.

Then we move the slider to 80%. Now the tempo there is 96. What would the slider now read? Well, 80%. (The other possibility would be than as soon as we let go of the slider, it changed the tempo to 96 and then the slider popped back to 100%. Not cool.)

So if, when we moved the slider to 80% it stayed there, when would the slider go back to 100% so we could, if we want, make another change in terms of percentage? When we moved the cursor?

That's why a "real-time" slider isn't good for work in terms of percentage. Percentage changes have to be done with a "one shot" enter-execute action: "Change to xx% of what we have now - execute."

In any case, keep in mind that if we want to change all the tempos in the entire score, or in some contiguous portion of it, to a certain percentage of their current values, we [E/W 5.0.0/] have a perfectly handy way to do that: with the Change Tempo dialog (using the second radio button). (That is in fact a set-execute operation, suited for work in terms of change to a percentage of the current values.)

Best regards,

Doug

[commnthings]

I (finally) reread the manual. The Tempo Slider does a GLOBAL change to all tempi based on the ratio of the New / Current tempi read from the slider bar for the measure in question.

I did some further testing with the current Tempo Slider bar (seems unchanged for many years). It actually behaves (or at least attempts to behave) as I've indicated I'm looking for. That is, it does a GLOBAL change to all tempi based on a ratio of the change you're making for the measure you're working on. The trouble is, the ratios it calculates are not very accurate. To demonstrate:

I opened a new score, 8 systems, 1 staff per system, 3 measures per system. For each system I manually set the tempi as follows:

System Measures Tempo
1 1 - 3 50
2 4 - 6 100
3 7 - 9 150
4 10 - 12 100
5 13 - 15 200
6 16 - 18 75
7 19 - 21 125
8 22 - 24 100

I put the cursor in meas. 2 and, using the slider bar, change toe tempo to 60 (a ratio of 120%. I then check the tempi for the rest of the piece. The results I got were

System Measures Tempo / Change %
1 1 - 3 60 / (120%)
2 4 - 6 120 / (120%)
3 7 - 9 180 / (120%)
4 10 - 12 120 / (120%)
5 13 - 15 240 / (120%)
6 16 - 18 85 / (113%)
7 19 - 21 145 / (116%)
8 22 - 24 120 / (120%)

Note that the changes for systems 6 and 7 were quite different from the rest and were significantly less than the desired 120%.

I then put the cursor in system 3 and changed the tempo to 150 (a decrease to 83%, returning the tempo to the original value). The results were

System Measures Tempo / Change %
1 1 - 3 49 / (82%)
2 4 - 6 101 / (84%)
3 7 - 9 150 / (83%)
4 10 - 12 101 / (84%)
5 13 - 15 199 / (83%)
6 16 - 18 70 / (82%)
7 19 - 21 121 / (83%)
8 22 - 24 101 / (84%)

Here, the changes were pretty accurate, ranging from 82 to 84%. Probably roundoff error. I then hit the <Reset> button.

System Measures Tempo / Change %
1 1 - 3 50 / (100%)
2 4 - 6 97 / (97%)
3 7 - 9 147 / (98%)
4 10 - 12 97 / (97%)
5 13 - 15 197 / (99%)
6 16 - 18 75 / (100%)
7 19 - 21 125 / (100%)
8 22 - 24 97 / (897%)

Not bad, but not consistent either. What then got weird was when I repeated the test (going back to the original saved score). I got different results.
Original:
System Measures Tempo / Change %
1 1 - 3 50
2 4 - 6 100
3 7 - 9 150
4 10 - 12 100
5 13 - 15 200
6 16 - 18 75
7 19 - 21 125
8 22 - 24 100

I put the cursor in meas. 2 and, using the slider bar, change toe tempo to 60 (a ratio of 120%). The results I got were

System Measures Tempo / Change %
1 1 - 3 60 / (120%)
2 4 - 6 120 / (120%)
3 7 - 9 180 / (120%)
4 10 - 12 120 / (120%)
5 13 - 15 240 / (120%)
6 16 - 18 93 / (124%)
7 19 - 21 153 / (122%)
8 22 - 24 120 / (120%)

Note that the changes for systems 6 and 7 were quite different from the original test and were significantly more than the desired 120%.

I then put the cursor in system 3 and changed the tempo to 150 (a decrease to 83%, returning the tempo to the original value). The results were

System Measures Tempo / Change %
1 1 - 3 51 / (85%)
2 4 - 6 101 / (84%)
3 7 - 9 150 / (83%)
4 10 - 12 99 / (83%)
5 13 - 15 201 / (84%)
6 16 - 18 77 / (83%)
7 19 - 21 127 / (83%)
8 22 - 24 99 / (83%)

Didn't do the <Reset>.

The point here is not the magnitude of the ratio errors (Again, roundoff error) but the fact that they are not consistent.

It also shows that the mechanism that I'm looking for is already there. Instead of calculating the ratio from the New / Current tempi you would simply read it off the (new) slider bar and apply, exactly as it does (inconsistently) now.

Another thing I noted was that if I did a manual tempo change somewhere (using the regular menu (Measures | Tempo) function and then did the Tempo Slider bar thing, the <Reset> did NOT account for the manually made change. That might actually be a difficult fix(?) to incorporate.