Pitch Block - What Am I Missing? (Latency Issue)

Ugly Bunny

Power User
So all I've been reading over the past few weeks is how great the pitch shifter is and virtual capo this and that... Yet, I updated to 2.0 and it's, well, virtually unusable with latency; like playing through Bias FX on an iPad through bluetooth speaker bad. Like, I tap the strings and a few ms later I hear the signal. There's no way I can use it like this.

There doesn't appear to be any settings that I can adjust in the pitch block itself - is there somewhere else I should be looking? I thought maybe my preset was too CPU intensive so I made a new one with just an input, output, amp, cab, and pitch block (right after the input) and either way; definitely noticeable latency.

I don't know. I'm kinda let down; I'm open to suggestions - hopefully I'm just doing something wrong if other people are raving about it. Do I need to try one of the Whammies? They don't seem to be much better. Or maybe it's just a hardware limitation and in a full band situation it wouldn't be as noticeable... I've never used the Digitech Drop so I have little to judge it by, other than just knowing that is simply doesn't "feel" right. I almost never use these types of effects because of this reason but I had my hopes set high after all the hype :)
 
I'm having the same issue, very basic preset, in-pitch-amp-cab-out. Even with the shift set to 0, the latency is around 35ms. I added a mono delay set to 100% mix to try & gauge what the delay is in the pitch block & toggling back & forth, they feel about the same at 35ms. I would be thrilled if I was getting 2-3ms latency!
 
Last edited:
What's your CPU use in the preset? Post preset?
Do you really mean "a few" milliseconds? As in, 2-3 msecs? Because that's as good as anything gets and should be largely imperceptible.

I'm having the same issue, very basic preset, in-pitch-amp-cab-out. Even with the shift set to 0, the latency is around 35ms. (added a mono delay set to 100% mix to try & gauge what the delay is in the pitch block & toggling back & forth, they feel about the same).

This - exactly this. I'd say 35ms sounds about right. That's a great way to test the latency! I'll have to remember that!

But yeah, @iaresee - 2-3 would be pretty much imperceptible, but it's quite perceptible. Unusably perceptible. Maybe I'm just expecting too much. But to answer your question, the first preset was about 70-some percent with the pitch block, and the basic one was about 26% including the pitch block. I've pretty much ruled out CPU being overloaded. I'm sure there's either a setting I'm missing or I just can't be pleased by an effect like this that just isn't "there" yet, in terms of industry-wide tech/algorithm implementation.
 
I'm hoping I can find some issue that is causing this. It's no where near useable for me the way it is on my Axe. I don't have experience with any other pitch shifter other than the Axe Fx II & normally only use it for pitch detune chorus or harmonies but it would be nice to have the Virtual Capo available as a viable option.
 
Here it is
Everything seems fine here. I'm measuring 29 ms peak-to-peak for your preset between the shifted and unshifted tones. And ~23 ms if I push tracking down to 1.3. That's very good for polyphonic, pitch shifted cleans.

2-3 would be pretty much imperceptible
To be clear, I was asking for clarification on your "a few" statement -- I "a few" would be single digit delay, not 30+ ms.

Setup and measurements are visible here:

 
Well, I was pretty close in my guess of about 35ms. If this is normal & how the down tune/virtual capo is supposed to be, so be it. Not useable for me. I tried playing some funk rhythm with the click & it down tuned 1 whole step & it was impossible to lock in to the click. (And yes, I can lock into a click after 51 years of playing professionally.) Thanks for looking at the preset. At least i know there is not something wrong or setting overlooked.
 
I've been experiencing the same with the latency, I thought there was something wrong as well but it seems like that's how it's suppose to sound. I was hoping to use this to tune down half a step and a whole step which would save me taking another guitar to a gig but it's just not cutting it for me unfortunately.

Just my opinion but I wouldn't sell your Digitech Drop's just yet lol
 
I've been experiencing the same with the latency, I thought there was something wrong as well but it seems like that's how it's suppose to sound. I was hoping to use this to tune down half a step and a whole step which would save me taking another guitar to a gig but it's just not cutting it for me unfortunately.

Just my opinion but I wouldn't sell your Digitech Drop's just yet lol

Drop tuning / virtual capo can't be achieved without latency.

The improvement with the latest firmware (latency and also stability and non-warble) is so obvious to me, that I can't understand that not everyone is experiencing this.

What's important is to turn up the output level. The acoustic sound and attack of the guitar should not interfere with the shifted output signal.
 
Your measurements are faulty. The block's latency is measured between 5.2 and 8.2ms. I set up a simple preset with input going into the pitch block to an output and a parallel path bypassing the pitch block. The difference in signal travel between the 2 paths varies between 5.2 and 8.2ms. Not sure what accounts for the differences but my measurements were all in that bracket. which is less than a third the Drop's latency (about 17ms).
The overall latency of the III, going through AD-DA and other DSP is greater, but the pitch block's latency, which is what is being discussed here is 5.2 - 8.2ms. Which is absolutely ridiculously low.
 
  • Like
Reactions: yek
To me it is not in the ballpark of my Digitech Drop Tune pedal. No biggie, but would be killer if could eliminate that ONE pedal I own and use with the Axe Fx III.
 
I'll play with it a little more before I give up on it entirely. It seems it is the way it is and that's just a wall that the tech has run into. It's great if it's good enough for you guys, but yeah, definitely not even in the ballpark for me; I know we've all got different levels of sensitivity to these sorts of things, so it's not a dig at anyone. Maybe in a pinch with a full band and lots of distortion it could fly; but I feel like for home practice or to just avoid having to retune a guitar (or use a capo) I'd be constantly slowing down.

I once had to be an announcer for an event where they put me IN FRONT of the speakers about 100m away from them. You should have heard the garbled sentences that came out of my yap and words bled into other words as my brain and ears and mouth all basically melted like ice cream in the sun... It would have been hilarious were it not for a major military change of command ceremony! Anyway, that's what latency does to me. It's gotta be immediate or bust :)

Oh well. Thanks for confirming my suspicions. Love you guys, and the Axe :)
 
I'll play with it a little more before I give up on it entirely. It seems it is the way it is and that's just a wall that the tech has run into. It's great if it's good enough for you guys, but yeah, definitely not even in the ballpark for me; I know we've all got different levels of sensitivity to these sorts of things, so it's not a dig at anyone. Maybe in a pinch with a full band and lots of distortion it could fly; but I feel like for home practice or to just avoid having to retune a guitar (or use a capo) I'd be constantly slowing down.

I once had to be an announcer for an event where they put me IN FRONT of the speakers about 100m away from them. You should have heard the garbled sentences that came out of my yap and words bled into other words as my brain and ears and mouth all basically melted like ice cream in the sun... It would have been hilarious were it not for a major military change of command ceremony! Anyway, that's what latency does to me. It's gotta be immediate or bust :)

Oh well. Thanks for confirming my suspicions. Love you guys, and the Axe :)

There certainly is variability in latency awareness per individual. I’ve done a bit of research into things such as minimum auditory gap detection on normal and impaired hearing listeners and results vary a good bit. Around 10ms is undetectable by anyone and some people can’t even detect 30-40ms. With hearing damage temporal and frequency resolution generally decline, as such, hearing loss just isn’t reduced thresholds to soft sounds.
 
Drop tuning / virtual capo can't be achieved without latency.

The improvement with the latest firmware (latency and also stability and non-warble) is so obvious to me, that I can't understand that not everyone is experiencing this.

What's important is to turn up the output level. The acoustic sound and attack of the guitar should not interfere with the shifted output signal.

I can't turn it up that loud but I tried it with headphones my drummer uses to record with that have the same shells as airport ramp & baggage handlers use to block out the noise. Still unusable for me. I'm not saying it's not improved but I can't really say how much. This is something I've never found useable on the Axe & I haven't used any of the other units that do drop tuning so I don't have a frame of reference. I found the latency too much even with the pitch shift set at zero. It's not something I would use much anyway & my point was not to complain but to find out if there was something specifically wrong with a setting on my end. Apparently not.
 
Your measurements are faulty. The block's latency is measured between 5.2 and 8.2ms. I set up a simple preset with input going into the pitch block to an output and a parallel path bypassing the pitch block. The difference in signal travel between the 2 paths varies between 5.2 and 8.2ms. Not sure what accounts for the differences but my measurements were all in that bracket. which is less than a third the Drop's latency (about 17ms).
The overall latency of the III, going through AD-DA and other DSP is greater, but the pitch block's latency, which is what is being discussed here is 5.2 - 8.2ms. Which is absolutely ridiculously low.

This is making me thing there is something wrong on my end. I am definitely getting a distinct doubling affect with the virtual capo or the dual shift, even set at zero. No way would I hear that if I was getting 5.2-8.2ms latency.
 
I had to do the measurement again because I closed the session without saving it.
This one is 9.5ms.
Not sure why the variations.

Screen Shot 2018-11-22 at 1.04.35 PM.png
 
Back
Top Bottom