ctdonath@rodan.acs.syr.edu (Carl T. Donath) (07/20/90)
I just got an interesting spec sheet... OKI Semiconductor (tel 408-720-1900) sells the "MSM6322 Speech Signal Pitch Control LSI" chip. To quote part of the sheet: "General: The MSM6322 converts in realtime the pitch of the speech signal in a range of one octave upward or downward. Two pitch control methods can be selected. one is to change the pitch in 17 steps by two switch inputs, and the other is to select one of 16 steps by four binary input lines. Since a microphone preamplifier and a low pass filter are built in, the pitch conversion set can easily be configured by connecting a microphone, amplifier, and speaker in the peripheral circuit. "Features: - Built-in microphone preamp - Built-in low pass filters (4th order LPF on input and 3rd order on output) - Built-in 8-bit A/D converter - Built-in 9-bit D/A converter - Speech pitch alterable in 17 steps - Oscillation frequency at 4MHz - 5v single power supply - 24-pin plastic flat package - Silicon construction (Si gate CMOS IC) Note: designed for application to home electronic equipment (toys)" There's a couple of application circuits given that are very simple to build. Thought this might be useful to whoever is discussing time/frequency shifting... - Carl
davisonj@ecn.purdue.edu (John M Davison) (07/20/90)
In the interest of accurate, consistent terminology, I am compelled to point out that frequency shifting and pitch shifting are two very different things; let's keep it that way. Pitch shifting = frequency scaling freq(new) = freq(old) * factor Frequency shifting: freq(new) = freq(old) + factor -davisonj@en.ecn.purdue.edu
bw@hpcvlx.cv.hp.com (Bill Wilhelmi) (07/24/90)
/ hpcvlx:comp.music / davisonj@ecn.purdue.edu (John M Davison) / 7:53 am Jul 20, 1990 / > In the interest of accurate, consistent terminology, I am >compelled to point out that frequency shifting and pitch shifting are >two very different things; let's keep it that way. > >Pitch shifting = frequency scaling >freq(new) = freq(old) * factor > >Frequency shifting: >freq(new) = freq(old) + factor > >-davisonj@en.ecn.purdue.edu >---------- Being on the fringe of this sort of stuff, I can't help but notice an oddity. If factor = (2 * freq(old)), then freq(new) = freq(old) + factor [ Freq Shift ] = freq(old) + 2 * freq(old) = freq(old) * 3 [ Pitch Shift ] For Pitch Shifting and Frequency Shifting to have any differentiation of meaning, severe constraints must be placed "factor". The point is that according to the definition above, the value of "factor" doesn't seem to give enough information about whether the shift is a Pitch or Frequency shift. The only difference seems to be that pitch shifts are large-scale frequency shifts. But they still both modify the frequency. Am I missing some big point? Bill Wilhelmi USMail: Hewlett-Packard - Interface Technology Operation 1000 NE Circle Blvd. Corvallis, OR 97330 ARPA: bw%hp-pcd@hplabs.HP.COM UUCP: {ihnp4, cmcl2, decvax, sun, rice, tektronix}!hplabs!hp-pcd!bw
jthornto@fs1.ee.ubc.ca (THORNTON JOHAN A) (07/25/90)
In article <110260001@hpcvlx.cv.hp.com> bw@hpcvlx.cv.hp.com (Bill Wilhelmi) writes: >/ hpcvlx:comp.music / davisonj@ecn.purdue.edu (John M Davison) / 7:53 am Jul 20, 1990 / > >Pitch shifting = frequency scaling > >freq(new) = freq(old) * factor > > > >Frequency shifting: > >freq(new) = freq(old) + factor > >Being on the fringe of this sort of stuff, I can't help but notice an >oddity. If factor = (2 * freq(old)), then > > freq(new) = freq(old) + factor [ Freq Shift ] > = freq(old) + 2 * freq(old) > = freq(old) * 3 [ Pitch Shift ] > > >For Pitch Shifting and Frequency Shifting to have any differentiation of >meaning, severe constraints must be placed "factor". The point is that >according to the definition above, the value of "factor" doesn't seem >to give enough information about whether the shift is a Pitch or Frequency >shift. The only difference seems to be that pitch shifts are large-scale >frequency shifts. But they still both modify the frequency. Am I missing >some big point? > >Bill Wilhelmi Well, yes. Let's look at a signal that contains a 100Hz, a 200Hz and a 300Hz sine. The spectrum of this signal is: A| | | | | | | | | --------------------- 0 1 2 3 4 5 (x 100Hz) If we do a pitch shift on this, say a factor of 1.5, we do f = f(old) * 1.5 and get frequencies of 150, 300 and 450 Hz. This new signal will sound like the old one but a perfect fifth higher. The new spectrum is: A| | | | | | | | | --------------------- 0 1 2 3 4 5 If we do a frequency shift on the original signal, say by 50Hz, we do f = f(old) + 50 and get 150, 250 and 350 Hz. Note that these no longer have the same harmonic relationship. The spectrum now looks like: A| | | | | | | | | --------------------- 0 1 2 3 4 5 A pitch change stretches the spectrum while a frequency shift slides it. A frequency shift will generally change a harmonic sound into an inharmonic sound. ------- _/__/ ----------------------------------------------------- _| ___| E l e c t r i c a l | Johan Thornton, Esq. | | |_/ E n g i n E E r i n g |------------------------- |/| __| U n i v e r s i t y | jthornto@fs1.ee.ubc.ca |-| |/__ o f B r i t i s h |------------------------- | |_____| C o l u m b i a | This space for rent ---- |__|/_| ------------------------------------------------------
davisonj@ecn.purdue.edu (John M Davison) (07/25/90)
In article <110260001@hpcvlx.cv.hp.com> bw@hpcvlx.cv.hp.com (Bill Wilhelmi) writes: >For Pitch Shifting and Frequency Shifting to have any differentiation of >meaning, severe constraints must be placed "factor". The point is that >according to the definition above, the value of "factor" doesn't seem >to give enough information about whether the shift is a Pitch or Frequency >shift. The only difference seems to be that pitch shifts are large-scale >frequency shifts. But they still both modify the frequency. Am I missing >some big point? The aforementioned factor is not, as you suggest, some function of the incoming frequency. It is a real, positive constant. (Of course, if the models were modified so that the "factor" is another input, the factor could be real, positive function, including functions of the audio input. However, this is not central to the operation of either frequency-modifying signal processor.) -davisonj@en.ecn.purdue.edu
bw@hpcvlx.cv.hp.com (Bill Wilhelmi) (07/27/90)
Could anyone tell that I am not an EE? Thanks for the responses--this is an interesting topic. - Bill Wilhelmi
Alvin@cup.portal.com (Alvin Henry White) (07/29/90)
A ne to say again why I need time change without pitch change. I have a tape recording of a book. I have the book. I am trying to use it to teach someone to read. The recording goes so fast that the reader, who is just trying to learn the language, is too slow to be able to find, or follow, the place in the book. I have a shareware program called Music Transcription System that has something similar to the old bouncing ball sing-a-long that will point to each syllable in time to the music. The operator can change the metronome speed and the music changes speed. I just bought a new thing called a Sound Blaster Card that says it has sampling. If I record the words in a set of sound files linked to a dictionary I would like to make it read in time to the music. In essence, as fast or as slow as the child would feel comfortable reading as they gain profficiency. Following the bouncing ball. Alvin H. White, Gen. Sect. G.O.D.S.B.R.A.I.N. 1505 De Rose Way #66 San Jose, CA 95126 USA [ alvin@cup.portal.com (OR) ..!sun!portal.cup.portal.com!alvin ] Government Online Database Systems Bureau for Resource Allocations to Information Networks
dbell@cup.portal.com (David J Bell) (07/29/90)
Alvin, I don't know where you would find one of these today, but my wife bought a cassette recorder several years ago that would be just the ticket: I believe it was by Panasonic, I'll find out if you need the name. It looks like a standard portable casette deck, but has a little slide pot that runs from something like 50% to maybe 250% speed. The principle intent of this was to allow you to play a lecture tape back *faster* than normal, to get more comprehension in a given period of time, but it can also sloowwwww things down, and have them still quite understandable. The signal procssor was mostly a single analog shift register chip made especially for this purpose... Dave dbell@cup.portal.com