INFO : Rate/Tempo conversion

[cinemascope]

Hi to all,

first of all sorry if this question has been raised tons of times and I do not have found it . . .

I have a tremolo effect on a patch that was set with a rate of 62 . Now I want to sync it using master tempo and master clock division.

Which master tempo / clock divisions combinations are suitable to achieve the same rate?

Thanks a million to whose have stepped by !

[KyleM]

I would think that 124 Bpm would be twice as fast... You could sync the master clock using the "tap tempo" function while the tremolo patch is playing and that would line them up. Then just see what the tempo on the master clock is after you tap it.

[cinemascope]

Hi KyleM,

Unfortunately it is not , as syncing it to a master of 124 and clock division of 2 is way faster, btw your hint is precious and for sure will let me find which tempo corresponds!!!

Really appreciate it!

Wish you a lovely day ahead.

[KyleM]

You are welcome Cinemascope and I'm glad I could help!

[Mr_-G-]

Let's see... I recorded a wav file at tremolo=62 and the max amount. I could count 26 channel bounces in 5 sec. So the frequency is 5.2Hz or 312 bpm (5.2*60). However the master clock does not go that high, so I think you need to put the clock to 104 and the div to 1/4T (which should be a multiplier by 3, hence 104*3=312)

[cinemascope]

OMG , you rock guys!!

Mr_-G- , your count is correct! 104,1/4T does the trick.

Wondering if the rate manteins a fixed ratio with bpm so we can have a ratio conversion here.

[Mr_-G-]

Just checked the length of a tremolo cycle for a few values:

Code: Select all

Rate	 Cycle time (s)
0        15.46
1         6.786
3         3.194
5         2.082
10        1.122
20        0.581
30        0.393
40        0.299
62        0.191
90        0.135
110       0.108
127       0.096

Note that these are determined by hand, so they are not very accurate.
If you plot those values in a scatter graph (remove the first two rows for the moment) the plot looks like a line in log-log space.
So I computed the regression line and that gives: time = 9.5456066251*rate^-0.9456066251. The fit is very good considering those are experimentally determined values (R-squared is 0.9993).
You could use that formula to compute the approximate cycle length for a given Rate value.
Once obtained the cycle time, converting to Hz is trivial: frequency = 1/time and to bpm is simple too: bpm = frequency * 60.
Just as an example, a rate value for 62 (what you asked originally) gives a bpm of 311.349 which is quite close to what I measured before also experimentally (312 bpm).

Remember I said to ignore the first two rows? The rate 0 is a special case because the computation involves log(0) and that generates an error; for rate=1 I noted the point falls somewhat outside the plot so the fitting is based on the rest of the points.
Unfortunately the discrepancy for very small Rate values seems to be a bit larger. Perhaps the instruments have a loop up table of values which have been quantised, rather than being computed as above.
Anyway, I hope this is useful.

[pablomastodon]

That's HARD CORE, Mr. G! Thanks for Herculean effort!

Pablo

[Johannes]

Comes to mind: I f**** love science
Great effort indeed!
PS: I'd kill for that kind of R2...

[Mr_-G-]

Thanks! One more detail is needed here, although it was suggested earlier.
Since the values for Rate=48 and above produce tremolos quicker than 240bpm (the maximum Master Clock value), then one needs to set the Master clock to the formula result divided by 3 and the DIV value to 1/4T.
Enjoy!