Sad limit in stretching samples... why, oh Yamaha, why?

Covers the Yamaha MODX6, MODX7 and MODX8

Moderators: Derek, parametric, Saul

GuyDenruyter
Member
Member
Posts: 43
Joined: Tue Nov 13, 2018 10:28 am

Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by GuyDenruyter »

I very sadly have come across a very annoying limit of the Modx/Montage, it has to do with the limits of how samples can be stretched (pitched).

When importing a single sample, it apparently cannot be stretched over a wide pitch range. You import, say, a one sample wav that is pitched around C3, well, whatever you try, you cannot get the pitch higher than e.g. C6.

It was already illustrated some years ago in https://www.yamahasynth.com/ask-a-quest ... in-montage .

I think it has to do with limited processor capabilities (playing a 48000 wav 3 octaves higher is 8 times the original frequency, and sadly the Modx is not performant enough to do this).

This is a serious shortcoming, as it means that there is no way to properly use a basic custom waveform for synth capabilities. I have never come across a synth that has this kind of limits. The Alesis Fusion - an old timer - for instance, had possibilities to pitch up without I have ever noticed such limits, and pitch down until zero speed (which gave a great "break" effect, for instance).

Why, oh Yamaha, why?!


RLMUSIC Project YouTube Channel
User avatar
anotherscott
Member
Member
Posts: 387
Joined: Sun Jan 13, 2013 6:36 pm

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by anotherscott »

AFAIK, there are always limits in stretching samples upward. The Alesis you mention had multiple kinds of synthesis in it, not just sample playback. I suspect that the sound you're describing was acheived with one of its other engines.


GuyDenruyter
Member
Member
Posts: 43
Joined: Tue Nov 13, 2018 10:28 am

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by GuyDenruyter »

Hello, at least on the Alesis there were no such limits - I am talking about the sample playing here, not about other ways of sound generating. I hope Yamaha will solve this error...


User avatar
Derek
Global Moderator
Global Moderator
Posts: 5318
Joined: Fri Dec 07, 2007 1:00 am
Location.: Wales, UK
Contact:

VIP

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by Derek »

Are your samples multi-samples? Trying to stretch a sample too far usually sounds awful in my experience.


Regards
Derek Cook

http://www.carregddu.co.uk
http://www.echoes-music.co.uk
http://www.xfactory-librarians.co.uk
http://www.ex5tech.co.uk
User avatar
Rara
Global Moderator
Global Moderator
Posts: 686
Joined: Sat Jul 07, 2012 4:34 pm
Location.: Australia

VIP

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by Rara »

GuyDenruyter wrote: Mon Sep 21, 2020 6:56 am Hello, at least on the Alesis there were no such limits - I am talking about the sample playing here, not about other ways of sound generating. I hope Yamaha will solve this error...
Some of the AWM samples in my SY35 have the same issue. if i transpose the pitch of the sample too far
it gets to a point where notes will repeat themselves.
I can play a C1 (I ignore yamahas idea of middle C :roll: ) on my SY35 via remote keyboard and it will sound C2
(I do not remember off hand which patches do this, many do ... )
Derek wrote: Mon Sep 21, 2020 4:57 pm Are your samples multi-samples? Trying to stretch a sample too far usually sounds awful in my experience.
The answer to your question is in one of the reply posts by "Bad Mister" you linked a page to.
They can only go so far before they sound weird either up or down generally more upwards than down. Just like Derek mentioned above.
Multi sample C to F as one zone, F# to B for the next zone and so on, all the way up the keyboard.
Stretching downwards is always going to be a bit more forgiving, I read somewhere this was a trick
users of fairlight cmi used back in the day, to maximise the limited ram they had back then.

Rara 0-)


User avatar
Zeitfraktur
Posts: 5
Joined: Mon Aug 31, 2020 5:30 pm
Location.: Germany

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by Zeitfraktur »

Interestingly the limit seems to depend on different factors? E.g. Sample Rate.
Usually 44.1k samples stretch up exactly 2 octaves higher, but I also had one exception.
Different rates seem to and up at different limits, but I could not see a formula (square/log blocks my thinking) :)
I tried/checked the limits of some of my single/simple cycle waveforms:
PastedGraphic-15.png
Sascha
You do not have the required permissions to view the files attached to this post.


GuyDenruyter
Member
Member
Posts: 43
Joined: Tue Nov 13, 2018 10:28 am

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by GuyDenruyter »

Just to be clear - my question is not about whether a sample sounds "awful" or not when stretched. Simple waveforms (squares, saws, pulses, combinations of those, ...) sound as they are, whatever the pitch is (and when talking about more complex waveforms, maybe I want an "awfully" sounding sample). The point is, the Modx designers have built in an (undocumented) limit, so you cannot squeeze a sample so it plays at a pitch 3 or 4 octaves higher. Any oldschool synth can do this. Why? Very annoying limit.


GuyDenruyter
Member
Member
Posts: 43
Joined: Tue Nov 13, 2018 10:28 am

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by GuyDenruyter »

Zeitfraktur wrote: Tue Sep 22, 2020 8:44 am Interestingly the limit seems to depend on different factors? E.g. Sample Rate.
Usually 44.1k samples stretch up exactly 2 octaves higher, but I also had one exception.
Different rates seem to and up at different limits, but I could not see a formula (square/log blocks my thinking) :)
I tried/checked the limits of some of my single/simple cycle waveforms:
PastedGraphic-15.png

Sascha
Very interesting analysis! It confirms my findings. Question remains.... why, oh why... Yamaha?


User avatar
Zeitfraktur
Posts: 5
Joined: Mon Aug 31, 2020 5:30 pm
Location.: Germany

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by Zeitfraktur »

GuyDenruyter wrote: Thu Sep 24, 2020 10:21 am Very interesting analysis! It confirms my findings. Question remains.... why, oh why... Yamaha?
I assume to save DSP/CPU from exponentially increasing stretch demand...


GuyDenruyter
Member
Member
Posts: 43
Joined: Tue Nov 13, 2018 10:28 am

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by GuyDenruyter »

Zeitfraktur wrote: Tue Sep 22, 2020 8:44 am Interestingly the limit seems to depend on different factors? E.g. Sample Rate.
Usually 44.1k samples stretch up exactly 2 octaves higher, but I also had one exception.
Different rates seem to and up at different limits, but I could not see a formula (square/log blocks my thinking) :)
I tried/checked the limits of some of my single/simple cycle waveforms:
PastedGraphic-15.png

Sascha
Based on your table, I have discovered that the Modx cannot play at a rate higher than 180.000 samples per second. This could of course be a technical limit (but then again, so many oldschool synths can easily do this...).

In the table below I have added the frequency at which the highest note can be played, as well as the sample playback rate (last column) of this highest note. As you can see, for all your examples, the limit is around 180.000.

Image
You do not have the required permissions to view the files attached to this post.


User avatar
Zeitfraktur
Posts: 5
Joined: Mon Aug 31, 2020 5:30 pm
Location.: Germany

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by Zeitfraktur »

GuyDenruyter wrote: Thu Sep 24, 2020 10:47 am but then again, so many oldschool synths can easily do this...
I guess the old gear used quite simple transpose algorithms without any sophisticated interpolation and anti aliasing etc?
In the table below I have added the frequency at which the highest note can be played, as well as the sample playback rate (last column) of this highest note. As you can see, for all your examples, the limit is around 180.000.
What is the formula to calculate the target sample rate, e.g. when transposing 14 half-tones up?

Sascha


GuyDenruyter
Member
Member
Posts: 43
Joined: Tue Nov 13, 2018 10:28 am

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by GuyDenruyter »

Zeitfraktur wrote: Thu Sep 24, 2020 11:03 am
GuyDenruyter wrote: Thu Sep 24, 2020 10:47 am but then again, so many oldschool synths can easily do this...
I guess the old gear used quite simple transpose algorithms without any sophisticated interpolation and anti aliasing etc?
That could be... but then it would be nice to have this "raw" option, sometimes we just want this raw, aliased sound :)
GuyDenruyter wrote: Thu Sep 24, 2020 10:47 am In the table below I have added the frequency at which the highest note can be played, as well as the sample playback rate (last column) of this highest note. As you can see, for all your examples, the limit is around 180.000.
What is the formula to calculate the target sample rate, e.g. when transposing 14 half-tones up?
First, I look up the frequencies of the notes (could do this by a formula also, but Wikipedia is a good help). Then I calculate the ratio between both notes (column 6 in my table). Then I multiply the original sample rate with this ratio (e.g. if C3 is sampled at 44100, and you play it 2 octaves higher (which means 4 times the frequency), then the samples have to be processed 4 times as fast, hence 44100 x 4).


User avatar
Zeitfraktur
Posts: 5
Joined: Mon Aug 31, 2020 5:30 pm
Location.: Germany

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by Zeitfraktur »

GuyDenruyter wrote: Thu Sep 24, 2020 1:32 pm
Zeitfraktur wrote: Thu Sep 24, 2020 11:03 am What is the formula to calculate the target sample rate, e.g. when transposing 14 half-tones up?
First, I look up the frequencies of the notes (could do this by a formula also, but Wikipedia is a good help). Then I calculate the ratio between both notes (column 6 in my table). Then I multiply the original sample rate with this ratio (e.g. if C3 is sampled at 44100, and you play it 2 octaves higher (which means 4 times the frequency), then the samples have to be processed 4 times as fast, hence 44100 x 4).
Ha... found it! :idea:
You calculate the factor between n-semitones with this: (12th root of 2) power of n.
Thus in Excel/Numbers: factor = (2^(1÷12))^n where n is the delta of semitones. E.g. 12 tones up is 2.0 x but 6 tones up is 1,4142 x

Sascha


GuyDenruyter
Member
Member
Posts: 43
Joined: Tue Nov 13, 2018 10:28 am

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by GuyDenruyter »

Indeed! Isn't math beautiful? :)


User avatar
Ivan Jochner
Member
Member
Posts: 110
Joined: Thu Jan 31, 2019 2:35 pm

Re: Sad limit in stretching samples... why, oh Yamaha, why?

Unread post by Ivan Jochner »

GuyDenruyter wrote: Fri Sep 25, 2020 8:53 am Indeed! Isn't math beautiful? :)
Guys are you actually sure that you have chosen in your life the right path to be piano players Lol?))


Post Reply