Does Pitch block latency vary based on interval?

Wolfenstein98k

Wolfenstein98k

Power User
Howdy,

I need to play a song which involves riffs around A1 (55Hz).
Should I play it on the E string and pitch down seven semitones, or play it on the A and pitch down one octave?

It seems intuitive to me that pitch shifting an octave would be lowest latency, as the maths would be simplest - but I don't actually know how the Pitch block works its magic.

And yes I know, "use your ears", but I'm curious for a theoretical answer here. Anecdotal evidence welcome too!
 
Forget about latency. Its not the main factor here.

The farther you pitch-shift an instrument, the less natural it sounds. Try it with your voice. Sing into a mic and shift it up by an octave. It won't sound like you singing in a higher register. It'll sound like a cartoon chipmunk.

I always go with the least amount of shift that will get me where I need to be.


Anyway, latency is caused by pitch detection. It's the frequency you play at, not the frequency you shift to.
 
Rex said:
The farther you pitch-shift an instrument, the less natural it sounds. [...] I always go with the least amount of shift that will get me where I need to be.
Click to expand...
Interesting - this would lead me to think that the answer is playing on the E string, as it's shifting a smaller distance.

However...

Rex said:
Anyway, latency is caused by pitch detection. It's the frequency you play at, not the frequency you shift to.
Click to expand...
...This makes me think that playing on the A string is the answer, as the lower the pitch the harder to track.

Dang.
 
IIRC, it's not so much the frequency you play at; it's more about the lowest frequency that the pitch shifter can detect. The pitch detector has to hear a significant portion of a single wave to figure out what the frequency is. The lower the frequency, the longer the wave.

Wolfenstein98k said:
Interesting - this would lead me to think that the answer is playing on the E string, as it's shifting a smaller distance.

However...


...This makes me think that playing on the A string is the answer, as the lower the pitch the harder to track.

Dang.
Click to expand...
And that's the root of the problem. There's no such thing as perfect pitch shifting. The problem is that you're not just shifting the pitch of the strings. You're also shifting the resonances of the instrument, and changing its voice.

So it (just like most everything else) really comes down to using your ears. Listen to it each way, and decide which imperfection is easier to live with.
 
Last edited:
Rex said:
So it (just like most everything else) really comes down to using your ears. Listen to it each way, and decide which imperfection is easier to live with.
Click to expand...
Noooo that's exactly the answer I didn't want. It's also obviously the right answer. :p I'm just interested in the theoretical considerations here, and I think the strongest argument is that the harmonics & timbre being shifted forms the biggest negative factor on the sound.

Thanks for the input Rex.
 
  • Like
Reactions: Rex
Rex said:
The pitch detector has to hear a significant portion of a single wave to figure out what the frequency is. The lower the frequency, the longer the wave.
Click to expand...

I'm not sure if that has much of an effect on latency.

With tracking at 5 (mono mode) both shift amounts seem to average about 10-15 ms latency. The octave shift had a higher maximum (around 25 vs. 20 ms) but I only looked at about 10 test notes with each.

At 100% mix the smaller interval will probably sound better. You might want to also try octave shift plus some of the original note.
 
Bakerman said:
I'm not sure if that has much of an effect on latency.

With tracking at 5 (mono mode) both shift amounts seem to average about 10-15 ms latency. The octave shift had a higher maximum (around 25 vs. 20 ms) but I only looked at about 10 test notes with each.

At 100% mix the smaller interval will probably sound better. You might want to also try octave shift plus some of the original note.
Click to expand...
I think you've looked into it more deeply than I have.

Are you referring to the Octave Div algorithm vs. Fixed Harmony, or two different amounts of shift in Fixed Harmony?
 
I wonder if the latency is any lower using the Ring Mod block and the octave down settings.

@Bakerman am I right in reading your comment as saying that the octave is measurably and notably slower than a several-semitone shift? That's essentially exactly what I was after, but I don't have the facilities to do a remotely scientific test.
 
Wolfenstein98k said:
am I right in reading your comment as saying that the octave is measurably and notably slower than a several-semitone shift?
Click to expand...

On average it looked that way, but that was only based on about 10 notes.

An octave-down ring mod shouldn't have any real latency. The tone would be different from fixed harmony though, and you'd have to be careful about chords interfering with pitch detection.
 
Semi related, this weekend I took my AXE home from the practice room and wanted to record some bass parts and some harmonies. I tried with the Fixed harmony mode but found much more pleasing results with the adv Whammy and playing with the mix level, keeping some of the original note helps a lot of the fluidity. It seemed to keep more attack and felt a little more natural for mine. Obviously this was recording only not live playing but maybe worth a try if you haven't already. Also I'd play it on the A, punchier and more definition but enough lows. That plus and EQ block and I think you might get close to what your looking for.
 
The bigger the interval the higher the latency. And the more artificial it sounds. You can test that by setting the mix to 100%.

Also I think setting the pitch to negative (lower note than the original) pitch has higher latency than positive pitch. But not totally sure about the last thing.
 
You must log in or register to reply here.
Back
Top Bottom