Sped up but non chipmunk audio

Factual Questions
1

My new DVD player(1) has a “1.5x” option where the audio (and video) is sped up but with normal pitch. Making “chipmunk” audio is easy (play a record(2) or tape at faster than the recorded speed). How is the rate of audio sped up with the pitch remaining OK?

Are just the pauses sped up? Is there some sort of instaneous pitch change circuit used?

Curious,
Brian

  1. actually it is a DVD player / recorder with hard drive
  2. younger folks: These are like big black CDs
2

Go up to 2X, 4X, 8X, 16X, 32X and you’ll see how it works.

Sections are skipped over - it’s more obvious at higher speeds, but it works the same way at low speeds, too.

3

My DVD recorder will only play sped-up sound in the 2X mode. It sounds like you only get to hear the first bit of each second of sound.

4

Except there is no audio at > 1.5x.
I didn’t do a thorough test, but there were no apparent words missing in the 1.5x playback.

Brian

5

I think it’s pretty straight-forward:

Instead of squeezing all the samples into 75% of the duration, which would increase the frequency and therefore the pitch, skip every third sample. The pitch remains the same, you just lose resolution and duration.

I am not an audio engineer – this is just an assumption.

6

IR audio engineer, and alas skipping a sample and closing the gaps just adds quantisation noise unless the sample rate is very large. To slow down the pitch of a fast-forward playback and maintain normal pitch you’d need to employ pitch-shifting devices. The 'puter in the DVD has access to the audio data anyway, and so a bit of digital signal processor coding later and you’ve got an inbuilt pitchshifter. There’s a small delay in the audio processing, but it’s not generally noticable.

My old Technical Drawing teacher at school designed an electronic pitch shifter to enable talking books for the blind to be got through a little faster. I wasn’t privy to how it worked, but I’d imagine a chunk of audio was sampled and stored in RAM, resampled at a different rate, and played back, with the inevitable discontinuities being fudged over.

There’s another technique used in the advertising industry, and that’s to electronically process speech to remove or shorten pauses. It makes for very fast speech that’s still intelligible.