pitch shifting doesn't sound good

[brownt]

Why is it that when the frequency of of musical note is changed by electronic methods, it doesn't sound good? The further the note is changed the worse it sounds. Changes in lower notes are better than changes in higher notes.

For example, playing a note on a guitar sounds ok, and playing one octave up sounds ok, but if the first note is pitch shifted an octave electronically using an effects unit, although it is up an octave it sounds different, squeaking, a bit distorted and not good. Why is that?

[mmagin]

Conceptually, there are two ways you could be pitch shifting:  1. a linear change in frequency (such as you might achieve in a mixer in a superhet radio where you mix a signal with a carrier resulting in a sum and difference, then filter one out.)  2. a change in frequency where harmonic relationships are maintained (such as you might get by playing digital audio samples back at a different rate than they were recorded at.)

With the first, as you go up in frequency, two notes (or two partials) become notes which are a much shorter musical interval between each other than they were to start with.  With the second, you still have the same musical interval.

[DaJMasta]

The nature of acoustic instruments and voice make them sound slightly different over a range, changing timbre (other harmonic content) with different notes, strings, octaves, comfortable ranges, etc., and if you're used to the sound of the instrument in question, you can tell when the sound has been altered because the instrument playing in that range normally (if possible) would have a different timbre.


If you take a synth making some set waveform, adjusting the frequency (pitch) doesn't effect the timbre, so anything you notice would be limiting effects of the system, caused by uneven response in the electronics or, more likely, the speakers.  It's also worth mentioning that the frequency response of the human ear plays a significant role - a 1kHz square wave will sound different from a 4kHz square wave because we're able to hear much more of the odd harmonic content that characterizes the composition of the wave - the 1kHz we hear clearly 1,3,5,7,9,11,13,15,17kHz harmonics (well, you probably can't discern the last few given the power relative to the primary tone, but those can be discerned with the normal 20Hz to 20kHz range of hearing), but a 4kHz square, 2 octaves up, we can really only hear the 4kHz, 12kHz, and maybe 20kHz tones that comprise the odd harmonics of the fundamental.  The ear's low pass filtering cuts out the higher harmonics, thus changing the profile of the sound being perceived, though this is not likely to be noticeable unless fairly high in range for most musical instruments or singing.

[John B]

Why is it that when the frequency of of musical note is changed by electronic methods, it doesn't sound good? The further the note is changed the worse it sounds. Changes in lower notes are better than changes in higher notes.

For example, playing a note on a guitar sounds ok, and playing one octave up sounds ok, but if the first note is pitch shifted an octave electronically using an effects unit, although it is up an octave it sounds different, squeaking, a bit distorted and not good. Why is that?

What method are you using? Quasi-realtime hardware effects, post production software effects and plugins?

The first issue is that the harmonic content. Take your guitar example: if you play a note then play a fretted note an octave higher, it will have a different frequency envelope vs pitch shifting the lower note.

Secondly is aliasing, artefacts and distortion. Not all pitch shifting methods are equal, and one method may work better for one timbre vs another. Ie, vocals, strings, percussion etc.

Polyphonic audio is harder to pitch shift without artefacts, even 2 notes on a guitar at once vs a single (monophonic) note makes a difference.

Quasi-realtime hardware effects (<10ms latency) tend to give worse results than ones that are designed for post production use.

The technology is actually quite good now. When used correctly and within reason (+/- a tone perhaps) it can be imperceptible that the pitch has been altered.

[BrianHG]

Could you be talking about sound/audio/sampling aliasing?  This happen when cheaply down-sampling without proper filtering.
Does the same effect happen if you analog record the guitar on an analog tape and playback the tape at half speed?
What software are you using for the octave lowering?
What function are you using to do it?

[Audioguru]

Many singers cannot sing in tune (Drake is one) so they use a pitch corrector (Autotune) that corrects their pitch and it sounds good but is obvious to people like me.
Just now I watched a video showing many pop singers singing the same song with and without Autotune and they had difficulty finding anyone who can sing well without Autotune.  They found Christina Aguilera.

[David Hess]

If the sample rate is changed, it affects both pitch and duration but the fundamental and harmonics maintain the same relationship.  For instance a tone at 1kHz shifted to 1.5kHz has its second harmonic move from 2kHz to 3kHz or the same 50% change and it remains the second harmonic.

Mixing alters the relationship between the fundamental and harmonics.  A tone at 1kHz shifted to 1.5kHz has its second harmonic move from 2kHz to 2.5kHz which is going to be an issue.  Solving this is a fascinating problem.