Forum index > Bug reports

transpose loses range w/ hi multipliers

AuthorMessage
noyzelab
theres a problem with the transpose when setting high multiplier values. this is the process to replicate this issue.

init voice to get a 1op sine.

go to the base tab & set transpose to -60. set op1 multiplier to 40. then descend on the keyboard from say a few notes above C - note 60, once i go below C note 60 it stops transposing down, every note is the same as that C. experiment with moving the transpose value up by increments of 1, and u'll notice that for each increment, u gain an extra note downwards past note 60, but of course only to the same frequency limit. i experimented a bit more with this, and it does the same thing for all the operator multiplier values.

if i load this voice into an initialised patch, the transpose in the patch page 2 menu will only transpose up, not down.
KODAMO
This is due to the fact that MIDI notes range from C-1 to G9.
By transposing down by 60, any note below 60 is already zero (the lowest) so it can't go any lower.
Consider lowering all your multipliers in your sound (40 becomes 20, 1 becomes 0.5 etc.) instead of using extreme transpose values
noyzelab
i hear what you're saying, but surely i should be able to transpose the oscillators down and retain the full 128 note midi range. thats really the point of having higher op ratios. the dx7 allows for up to 61.69 in order to obtain a wider range of spectral forms.

to organise c:m ratios according to their spectral identity i use barry truax method of taking farey sequence values, as discussed here =>

http://noyzelab.blogspot.com/2016/04/farey-sequence-tables-for-fm-synthesis.html

as you can see, reducing the op ratio will reduce the amount of spectral families available.

in any case, the transposition problem appears even when reducing a ratio from 40 to 20 or to 10 etc. it doesn't fix it. i still lose part of the lower range. enclosed is 40 to 1 reduced to 10 to 0.25 [which i've set to multiplier 0 & quartertones at 6 to approximate 0.25]. as you can hear, this is transposed down by 60 and every note below middle C is the same. its impossible to get it to play all the way down.
noyzelab
this work around gets the pitch down without using transposition, but requires a pitch env per op set at range = 8oct. i'll investigate and see how using a mod matrix to send one pitch env for multiple ops might keep it at only using one.

while i was doing this i noticed that pasting the pitch env doesn't include the range, it would be handy if this was also included.
KODAMO
Thank you for the example files you provided.
The loss of range is due to the fact that the Transpose parameter on Voice doesn't transpose the oscillators, it transposes the incoming MIDI note and that's not easy to change.
The workaround you found may be a good enough solution for the rare cases where you need extremely high frequency modulators.

Reply