Forum to discuss all matters relating to the MPC1000 and MPC2500 operating systems created by 'JJ' (all versions).
User avatar
By naugie Thu Oct 05, 2006 12:23 pm
ok im just goin gaga in the looping world how do i transpose loops? the manual way without acid without fruity without software... artist around the times of the linn 9000 figured this out but the secret seems to be lost forever.. how was it done manually? dont say by ear... i know damn well math plays in this somewhere...
User avatar

By chilom Thu Oct 05, 2006 2:27 pm
what do you mean by transposing a loop to bar ?

By tombola Thu Oct 05, 2006 2:50 pm
By ear.

This is how I used to do it in the olden days.

One way:
1) Make your loop in a wave editor (like the one in the MPC, perhaps)
2) Set it to 'one shot'
3) Trigger it in the first beat of each bar (assuming it's one bar long)
4) Adjust the tempo until it loops properly.

Another way:
As above except
4) Adjust the pitch (i.e. the playback speed) until it loops properly.

The third way (which certainly wasn't possible with a Linn 9000)
As above except:
1.5) Timestretch your loop until it's the length of the bar (or half a bar, if you're making drum'n'bass!).
User avatar

By naugie Thu Oct 05, 2006 2:53 pm
say we got a sample thats 2bars long and we wanna have it fit 6bars at tempo 160 using semitone and cents to stretch the sample out. this is what i mean by transposing loop to bar. say if the sample is 121.4 bpm and the tempo were workn with is 160bpm the samples lentgh would have to be + 4.78 in semitone and cents for it to fit 2 bars in a 160bpm track now if we transpose an octave down the sample will be slower but will still be in sync. and i mean 12 keys up or down using our new + 4.78 as our default setting for the sample loop. i cheated to get this answer but how to get this answer is a mystery. how the hell did i figure this out? i wanna know how these calculator got these answer.. what do the math formulas look like to figure this out?

===small rant dont reply to this rant==========
btw this is nothing new i stumbled across its just sad that u and i and everyone else dont know how to manually do this.. its a luxury if our mpc could do it and its an infant still that many programs are now offering this as a feature. its not time stretching its not pitchshifting. but this an old discussion of how pitch with resampling its original sound = lentgh in bytes well to me atleast u other will argue im wrong
User avatar

By chilom Thu Oct 05, 2006 4:23 pm
oh ok so the math way for calculing it is using the following formula :
TempoFrom (in your example 121.4)
TempoTO (160)
pitch=(ln(TempoTO/TempoFrom))/0.05776227)

so i have here 4,7796637340886574934681458207224

i i use mathround i will say 4.78 is a good value.

if TempoFrom is greater than tempoTO then it is a negative pitch
if not ( like the exemple above) it is a positive pitch.

there is no mystery here ;)
User avatar

By naugie Thu Oct 05, 2006 6:41 pm
i should tell u how i cheated first before i get into this formula....

i used timestretch in mpc to tell me what the tempo was to my loop because i dont know how to manualy figure out tempo through math yet but i did not apply timestretch to this sample. i then used the tempo that mpc told me was for the loop and droped it in the deepsound calculator to generate the correct semitone and cent..now i can change the pitch by 12 keys to that sound up or down even 24 keys.. the sound will be in complete sync without timestretching. the advantage to this is a cleaner sounding patch phrase and when pp is applied its also better then timestretch in my opinion .. help me out i do not wanna be slave to proprietary software.. its lisense renewals are hurting my wallet so i weill use math kill off my demons.

im checkn out ur formula what is pitch=(in? first off i didnt have a pitch wehn i started the project.. the way i got a pitch was by knowing the samples bpm and the main tracks bpm.. with those 2 bpm somehow the calculator in deepsound generated the correct semitone and cent to use.. ya i cant understand can u show me in steps how ur formula works?
User avatar

By gunmetal Thu Oct 05, 2006 6:44 pm
1,2,3,4.... tap your feet. listen to the beat nod your head enough said!
User avatar

By chilom Thu Oct 05, 2006 9:19 pm
ok

for calculating the bpm of sample.

go in trim mode, and write down the sample length. or note the start - end sample lenght.

then bpm is known as to be :

BPM = 60 /((((sample end - sample start)/44100)/TimeSig)/Number of BAr)

sample end - sample start = go in trim mode and check the sample value.
Time signature = 4/4, 8/4 etc....
Number of bars = mesures
44100= this is the definition of our audio resolution

for tempo transposition, using the pitch value :

T1= Original Tempo (ie: 121.4 bpm as your example)
T2 = final tempo (160 bpm)

so the pitch needed for the transposition is pitch =(ln(T2/T1))/0.05776227)

so in your example again :

pitch = (ln(160/121.4)/0.05776227)
pitch = (ln(1,317957166392092257001647446458)/0.05776227)
pitch = (0,27608293660842903413770703889657/0.05776227)
pitch = 4,7796413923557546152134782600476

picth ~= 4.78
the sign is given by the fact T2 is greather than T1 thats all.

this is how deepsound calculates it.

changing a sample by +12 pitch or -12 , makes the orginal BPM of the loop divide by two (when -12) or multiply by 2. it works with 24, 36 and 48, and respectivly by 4,8,16

for the timestrech factor betwenn two BPMs:

timestrech factor = (BPM2/BPM1)*100 thats all ( expressed in percent)
User avatar

By Od3rus Fri Oct 06, 2006 2:51 am
or for a different feel you could just chop it on the hits and play the attacks/initial bits. Picks up the pace and attacks are what people recognize, can sound cool, can sound like shit... depending on the source material and the desired outcome.
Everyone else already ave the other answers.
=)
User avatar

By naugie Sun Oct 08, 2006 4:23 pm
yooo alright im back. wish i replied sooner but its hard at times to find time. alright im doin this but im kinda stuck at the first point witch is how to manually detect tempo.. im workn with a new sample witch btw is 160bpm when i made it, but akai os 2.1 detects it as 162.1 but thats prolly due to my manual cutting ability. remember even though i cheated and used akai os to tell me tempo i wanna pretend that i dont know the tempo for this sample.the sample size is 132253. im still workn with 160bpm for the sequence and the time sig is 4/4.. for some reason i get a different answer for the tempo wehn i do the following steps u gave me.. i dont come remotely close to the tempo mpc gives me and what tempo i know i started with. the deepsound calculator gives me another tempo witch is how i wanna do it sense there calculators are so dead on accurate. its answer for this sample was 160.057.. i wanna upload this sample so we can work with it together closely so u can help me further understand this. does anyone have a fileserver to upload this drum loop?

this is what u wrote
BPM = 60 /((((sample end - sample start)/44100)/TimeSig)/Number of BAr)

now correct me if i do wrong here
BPM = 60 /((((sample end - sample start=132253)/44100)/TimeSig=4/4)/Number of BAr=2)

my question is when i got to the time sig part how do i divide my answer?

132253/44100=2.9989342 the rest from here im lost
User avatar

By chilom Sun Oct 08, 2006 5:18 pm
your sample lenght is 132253

so :
for time sig you have to put only the numerator (ie : 4, 8 ...)
BPM = 60 /((((sample end - sample start)/44100)/TimeSig)/Number of BAr)
BPM = 60 /((((132253)/44100)/4)/2)
BPM = 60 / (((132253)/44100)/4)/2)

BPM = 60 / ((2,9989342403628117913832199546485/4)/2)
BPM= 60 / (0,749733560090702947845804988662/2)
BPM = 60 / 0,374866780045351473922902494331
BPM = 160,05686071393465554656605143175

so 160 is correct :)