Forum to discuss all matters relating to the MPC1000 and MPC2500 operating systems created by 'JJ' (all versions).
By ssyniu Sun Sep 10, 2017 7:12 am
tapedeck wrote:
ssyniu wrote:All right so this trick is affecting sample rate I would never thought about it.
Thanks dude.

yes. think about it this way - if you sample one second of audio at 44.1k, that 1 second of audio has 44.1k samples in it. if you then slow that down by half, those same 44.1k samples have to spread out across 2 seconds now.

44100 / 2 = 22050 samples per second .... effectively half the sample rate.

that is a pretty simplified version, and the way each machine 'interpolates' the new samples (filling in the gaps from stretching 44.1k samples from 1 second to 2 seconds) is usually pretty unique, and contributes a lot to the sound of the machine. this is why pitching things sounds different across different samplers (emu's are very noticeable for example).
this also explains why sampling at a lower frequency is not the same as sampling at a higher frequency and then pitching down - because there is no need for interpolation if you sample at low rate to begin with.

nerdalert :mrgreen:


Nerd alert continue ;-)
Last edited by ssyniu on Sun Sep 10, 2017 7:17 am, edited 2 times in total.
By ssyniu Sun Sep 10, 2017 7:15 am
Thanks for your answers.
this tells me that it was a lot of sample rate reduction happening when this trick was done on the SP1200 (SP sample rate is about 26.04 kHz) plus trick(when sampling with 45rpm speed and pitched down about -5 semitones to get the initial pitch).
Is it possible to figure out(mathematically) the final sample rate???
Probably I would need to know the initial speed of the sample (before -5 semitone trick)and than the speed of the sample after the trick and than figure out how much the sample was slower from the initial speed.
User avatar
By tapedeck Sun Sep 10, 2017 6:55 pm
ssyniu wrote:...(SP sample rate is about 26.04 kHz) .... (when sampling with 45rpm speed and pitched down about -5 semitones to get the initial pitch).
Is it possible to figure out(mathematically) the final sample rate???


ok lets try this...
original sample rate is 26.04k = 26040 samples per second.
i'm also gonna run with your -5 semitones assumption.
we know that -12 semitones would be down a full octave, so that would be half the sampling rate, 13020, so the answer is somewhere between 13020 and 26040. i'm gonna go ahead and make a guess it's ~20000.

you can look up the actual ratio of a semitone as being 2^(1/12) - that 1 is the actual semitone number.

so that gives us...
original rate * (2^(semitone/12)) (i'm using multiplication here cause its a ratio).

so....
26040 * (2^(-5/12)) = 19507.9581409

close enough to ~20k :mrgreen:

simpler way (but not what you asked, really) is to find the ratio difference from 45rpm to 33.3rpm and then just multiply by that.
45/33.3 = 26040/x
x * 45/33.3 = 26040
x = 26040 * 33.3/45
x = 19269.6 ....which is pretty close (but not exactly) -5 semitones.

i'm guessing here that pitching down -5 semitones keeps the sample sounding just a little sharp, as the actual ratio of 33.3/45 is somewhere between -5 and -6 semitones (18413.0605821). good news is, your -5 semitones does appear to be the closest you can get to original pitch.

important to remember tho, that the numbers don't really matter here, as its still gonna be the interpolation / resampling algorithm that really colours the sound differently across machines.

please correct me if my math is all wrong
By ssyniu Mon Sep 11, 2017 12:38 pm
tapedeck wrote:
please correct me if my math is all wrong


Nope you are right.
I did actually test today to see exactly how much "pitching" down need to be done going from 45RPM recording to the original sample speed and the closest you can get is -5 semitones and 20 cents roughly.
I did this test using MPC3000
I can post loops if someone wants to hear the difference in sound between sampled at normal speed and the one with 45rpm trick.
Use this address to see the example picture:

https://ibb.co/eLy88a

bigger zoom

https://ibb.co/gvg38a

Thanks for Interesting discussion tapedeck.
All the best to you.
User avatar
By tapedeck Mon Sep 11, 2017 3:43 pm
yer welcome. i plugged in -5.2 semitones into the formula and yea, it came even close. actually -5.21 is the closest i could get to 45/33.3 ratio.
By ssyniu Mon Sep 11, 2017 4:42 pm
tapedeck wrote:yer welcome. i plugged in -5.2 semitones into the formula and yea, it came even close. actually -5.21 is the closest i could get to 45/33.3 ratio.


Yes thats right but in MPC the the smallest value or unit is 10cents so can't go more precise when tunning than 10 cents.
All the best
User avatar
By tapedeck Mon Sep 11, 2017 5:02 pm
ssyniu wrote:
tapedeck wrote:yer welcome. i plugged in -5.2 semitones into the formula and yea, it came even close. actually -5.21 is the closest i could get to 45/33.3 ratio.


Yes thats right but in MPC the the smallest value or unit is 10cents so can't go more precise when tunning than 10 cents.
All the best

that's what i figured...just being thorough :mrgreen: