sbell.sd wrote: Fri Nov 15, 2024 2:41 am
Sorry about the formatting, forum doesn't seem to allow to add a .txt file.
I've been having more tuning issues now that the season is cooling.
Perhaps Moog could provide some guidance on how to tune the muse or
expectations tuning accuracy, or a better "Start Tuning" and "Quick Tune" functions.
Rarely has this been in tune to my satisfaction, then is does sound great.
The oscillators sound so that they are beating against each other,
rather they are phasing against each other. Say within .25 HZ.
Almost identical issues as someone on reddit posted "Hard questions - Moog Muse tuning".
At A440 a 1 HZ delta is about 6 cents. These results were for playing the A440 note.
My voice 7 oscillator1 was almost always at least .5 Hz and sometimes 1.5 HZ off of all the other osc1s.
The spread of the other osc1s are usually within .4 HZ of each other. Yes I know equal temperament is off. But
playing one note in a layer and the with osc1 voice1 at 440 HZ and osc1 voice7 playing 441.5 HZ, the beating is pronounced.
With this much slob in the tuning, the detune knob seems purposeless, and no it doesn't tighten up my tuning issue.
So I held off, until today, running "Start Tuning",
as the manual states "...should not be run unless there are significant tuning issues."
Start "Start Tuning" at 4:45, room temp 72F. Muse had been off hours.
I saw this in original factory autocal-verbose.log file "Start Temperature: 42.1", which is ~108F.
This indicates to me the factory temperature was warmer than 72F, the manual states "at ~70 degrees F".
At ~6:00 "Start Tuning" completed, last temp I saw was 39C.
As I noticed later the muse continued to warm for at least the next hour.
Test measurements with INITIAL PATCH modified by setting osc2 volume up.
First results
6:12 40.6C
voice osc1 osc2
1 440.58 439.45
2 440.59 439.43
3 440.50 439.82
4 440.99 439.86
5 440.66 439.47
6 440.57 439.55
7 440.65 439.63
8 440.76 439.39
Largest diff between osc1s is .45 HZ, beat frequency ~2.2 second.
Largest diff between osc2s is .47 HZ, beat frequency ~2.1 second.
But the following are not acceptable.
Largest diff between osc1s and osc2s is 1.60 HZ.
All osc1s are over A440 by between .50 and .99 HZ.
All osc2s are under 440 by between .18 and .61 HZ.
All osc1s are around 1 HZ over osc2s.
What went wrong? These are all speculative.
1. OS has timing issues, preventing measuring frequency accuracy.
2. Unit temperature has not stabilized (enough), tuning is extrapolated as synth warms.
3. Hardware is farther out of tolerance than the software can correct for.
4. Software needs to provide handling greater hardware intolerance.
5. The software was hastily shoved out the door, without testing on enough units.
Could osc1 be manually tuned down .5 HZ and osc2 up .5 HZ, yes.
But as time continued the muse continued to warmup.
6:46 40.6C 6:55 40.6C
voice osc1 osc2 osc1 osc2
1 440.56 439.45 440.47 439.46
2 440.56 439.56 440.64 439.55
3 440.54 439.95 440.61 440.01
4 441.15 439.92 441.11 439.94
5 440.72 439.48 440.78 439.53
6 440.65 439.60 440.73 439.60
7 440.66 439.68 440.64 439.70
8 440.82 439.38 440.79 439.39
Notice the .4C temperature jump.
7:01 41.0C 7:21 41.0C
voice osc1 osc2 osc1 osc2
1 440.51 439.50 440.44 439.56
2 440.63 439.60 440.60 439.58
3 440.63 440.01 440.52 440.01
4 441.10 439.94 441.03 439.87
5 440.83 439.53 440.80 439.53
6 440.75 439.65 440.73 439.65
7 440.60 439.72 440.63 439.69
8 440.89 439.37 440.82 439.41
Notice the .1C temperature drop.
Here I did a "Quick Tune"
7:32 40.9C 7:42 40.9C 7:52 40.9C "another Quick Tune"
voice osc1 osc2 osc1 osc2 osc1 osc2
1 439.52 439.36 439.49 439.38 439.39 439.37
2 439.51 439.49 439.48 439.53 439.37 439.44
3 439.58 439.26 439.48 439.25 439.42 439.22
4 439.78 439.59 439.77 439.57 439.63 439.62
5 439.37 439.38 439.45 439.43 439.33 439.41
6 439.47 439.28 439.56 439.28 439.38 439.33
7 439.85 439.48 439.87 439.43 439.85 439.43
8 439.56 439.64 439.61 439.65 439.50 439.65
Notice here is the first time after "Start Tuning" the osc1s and osc2s are closer,
however, all oscs are off A440 by about .5 HZ.
Notice the .1C temperature increase.
7:58 41.0C 8:10 41.0C 7:52 41.1C
Quick Tune
voice osc1 osc2 osc1 osc2 osc1 osc2
1 439.49 439.36 439.44 439.40 439.47 439.45
2 439.52 439.49 439.46 439.56 439.49 439.57
3 439.48 439.23 439.40 439.25 439.39 439.21
4 439.68 439.60 439.68 439.68 439.71 439.68
5 439.44 439.40 439.41 439.39 439.40 439.40
6 439.41 439.29 439.37 439.35 439.42 439.36
7 439.83 439.42 439.87 439.47 439.82 439.47
8 439.58 439.65 439.51 439.61 439.48 439.72
What can be concluded?
"Start Tuning" must be followed by a "Quick Tune"?
After "Quick Tune" voice7, osc1 is always the highest outlier.
After "Quick Tune" voice3, osc3 is always the lowest outlier.
Oscillators 1 and 2 must be manually tuned to A440.
I am not satisfied with the muse's tuning.