Harmonizer sounds bad

[b418me]

Hello everyone, at the moment I am using the HX Effects in front of an older Vox AC30 which sounds good. Only the Harmonizer Sound is pretty bad. The harmonics/notes sound pretty harsh almost distorted (also when the amp is clean). Is this normal for this setup (no FXLoop)? Thanks Marc  

[silverhead]

The harmonizer generally works reasonably well for single notes but not for chords. It’s a mono- rather than poly - phonic FX. Works best for lead solos rather than rhythm. Does it work acceptably well for your solos?

[b418me]

11 minutes ago, silverhead said:

The harmonizer generally works reasonably well for single notes but not for chords. It’s a mono- rather than poly - phonic FX. Works best for lead solos rather than rhythm. Does it work acceptably well for your solos?

No I mean single notes. I don't think it sounds well enough to play live unfortunately 

[SaschaFranck]

While plenty of folks seem to disagree, I think that all pitch shifting effects in the Helix sound pretty bad, so IMO there's not much you can do about it.

[PierM]

44 minutes ago, b418me said:

Hello everyone, at the moment I am using the HX Effects in front of an older Vox AC30 which sounds good. Only the Harmonizer Sound is pretty bad. The harmonics/notes sound pretty harsh almost distorted (also when the amp is clean). Is this normal for this setup (no FXLoop)? Thanks Marc  

 

Harmonizer, as for any other pitch tracking/detection algo, it's sensible to signal quality, strength and clarity. If you are 100% sure your guitar tone is clean (and I mean, no fret buzz, clean playing, clean phrasing, no clipping, in tune etc..) then I'd try placing some compressor before the harmonizer, and of course, harmonizer before any other effect and amp/cab.

 

IME, if the input is clean and my playing style is clean, these effects are working pretty good and in a predictable way.

[craiganderton]

 

1 hour ago, PierM said:

 

Harmonizer, as for any other pitch tracking/detection algo, it's sensible to signal quality, strength and clarity. If you are 100% sure your guitar tone is clean (and I mean, no fret buzz, clean playing, clean phrasing, no clipping, in tune etc..) then I'd try placing some compressor before the harmonizer, and of course, harmonizer before any other effect and amp/cab.

 

In addition to that excellent advice, I'd also recommend pulling back your guitar's tone control. It's easier to track a signal with fewer high frequencies.

 

At least for me, where to insert the harmonizing effect seems to vary on a case-by-case basis. Some people have reported better results placing the harmonizing effect between the amp and cab. I usually insert the harmonizing effect after an amp if it distorts heavily. Even though you're playing single notes, as far as the amp is concerned you're playing at least two notes, so there's intermodulation distortion. If you place the harmonizing effect after the amp, then the harmonized output won't have the intermodulation distortion issues. But, in other cases, harmonizing effects sound better before the amp. I haven't really found a "rule" for this, it depends on your preset, the amp you're using, and the amount of distortion. Try both pre and post amp to see what sounds best. For a relatively clean preset, you'll probably want the harmonizer at the beginning of the chain.

 

FWIW I've used harmonizing effects live and no one threw anything at me, so I guess it was acceptable :)

[tjbassoon]

I sometimes will run the harmonizer on a parallel path to the original signal, and put an EQ block after it. Then return the path to the original signal. You just have to set the Mix in the actual harmonizer to be 100% wet, and then use the moment you merge the signals to actually blend the original and the pitch shifted signal together. You don't turn the effect on or off with this setup, but set a footswitch to engage or disengage the path/split, like having the shifter and EQ blocks in a true bypass looper on a normal pedalboard.

[jester700]

The above good advice hints at the harmonizer paradox: they typically TRACK better with a clean signal, but if used with distortion they SOUND better placed after the distortion. That's why Digitech's old HarmonyMan had a tracking loop; you could feed a clean signal to the detector, but apply the effect to the main signal. Of course, this added wiring complexity to your rig.

 

tjbassoon's cool idea does a similar thing, with the bonus of being able to shape the paths sounds differently. 

 

To sound best, IME any distortion generator (be it amp or pedal block) needs to be placed on both paths before mixing. This uses more DSP but allows your "harmony voice" to have a different amp.

 

I still think the HarmonyMan is slightly better sounding, but I waffle on it when I compare. They're close enough that I never use the pedal these days.

[b418me]

Thanks for the answers. I will try a little bit playing around with it.

[theElevators]

If you want to do the "queen" multitracked solo sound, you need to harmonize the distorted signal as opposed to do it the other way around.  Also, mix the harmonized note down to about 90%.  It should not be exactly as loud as the actual not you are playing -- that's the secret to it not sounding artificial IMO.

[somebodyelse]

My own experience...

I've been using pitch shifting since the 80s - MXR m129 and then later an Eventide H3000.

The Helix shifter/harmonisers are not bad, but they do take a bit of work. Of course, it does depend on the sound you're after - Brian May harmonies are a very different sound to Big Country bagpipe guitars.

I have so many does and don'ts, I don't know where to start...

One thing that did make a big improvement on the Helix for me. The MXR m129 has an inherent latency of around 40ms - yeah, I know that seems massive. When I first started trying to get a decent shifted sound on the Helix, there was something not quite right. I suddenly remembered this latency, dialled in a 40ms predelay on the Helix and it was thereabouts.

That may be because that's what my ear expects to hear after 35+ years or that when humans harmonise with each other there is always a certain amount of natural latency between them...?

[SaschaFranck]

9 hours ago, somebodyelse said:

That may be because that's what my ear expects to hear after 35+ years or that when humans harmonise with each other there is always a certain amount of natural latency between them...?

 

Why would there be a "natural latency"? Choir voices or bigband sections aren't 40ms apart from each other. And once I manually harmonize a line (as in actually playing a hamonization), there's no 40ms of latency, either.

[somebodyelse]

17 hours ago, SaschaFranck said:

 

Why would there be a "natural latency"? Choir voices or bigband sections aren't 40ms apart from each other. And once I manually harmonize a line (as in actually playing a hamonization), there's no 40ms of latency, either.

I didn't say they were 40ms apart, but I guarantee they aren't dead on in sync - neither are you, by the way.

There is 1 ms of 'latency for every foot of air between a sound source and your ear. 

[codamedia]

22 hours ago, SaschaFranck said:

 

Why would there be a "natural latency"? Choir voices or bigband sections aren't 40ms apart from each other. And once I manually harmonize a line (as in actually playing a hamonization), there's no 40ms of latency, either.

 

40ms is pretty extreme, but yes, there is "natural latency". Considering notes can be pushed and pulled by a certain tolerance.... 40ms difference is not out of the question during a live performance.... in a studio that might get tracked again to be a little tighter... but it will rarely be "dead on". 

 

For reference... with a mid tempo strum on an acoustic, 40ms is close to the time difference between the sound of the low E and high E ringing out. It's natural in that context, but would sound a little off as part of a double track. 

 

Latency is a term I prefer to leave to software and hardware... it's always late from the original... the "delay". 

Humans can be early or late... it's a big part of what makes us sound like humans rather than machines. 

[theElevators]

I made a little video how I do it...

[SaschaFranck]

5 hours ago, somebodyelse said:

I didn't say they were 40ms apart, but I guarantee they aren't dead on in sync - neither are you, by the way.

 

I'm absolutely familiar with latencies, their typical values and what not. Those 40ms was the number you came up with. And I'd never try to delay any harmonising part that much in case I was recording it manually.

[spaceatl]

7 hours ago, somebodyelse said:

I didn't say they were 40ms apart, but I guarantee they aren't dead on in sync - neither are you, by the way.

There is 1 ms of 'latency for every foot of air between a sound source and your ear. 

 

You certainly inferred 40ms of latency in your post....At sea level it takes sound 0.887ms to travel a foot...You latency variables make no sense at all bro...lol!

On 1/5/2022 at 2:18 AM, somebodyelse said:

My own experience...

I've been using pitch shifting since the 80s - MXR m129 and then later an Eventide H3000.

The Helix shifter/harmonisers are not bad, but they do take a bit of work. Of course, it does depend on the sound you're after - Brian May harmonies are a very different sound to Big Country bagpipe guitars.

I have so many does and don'ts, I don't know where to start...

One thing that did make a big improvement on the Helix for me. The MXR m129 has an inherent latency of around 40ms - yeah, I know that seems massive. When I first started trying to get a decent shifted sound on the Helix, there was something not quite right. I suddenly remembered this latency, dialled in a 40ms predelay on the Helix and it was thereabouts.

That may be because that's what my ear expects to hear after 35+ years or that when humans harmonise with each other there is always a certain amount of natural latency between them...?

 

 

[Mutonium]

1 hour ago, spaceatl said:

 

You certainly inferred 40ms of latency in your post....At sea level it takes sound 0.887ms to travel a foot...You latency variables make no sense at all bro...lol!

 

 

 

Hmmm...  They made sense to me.  I read @somebodyelse's post as using 40ms because that is what he is used to with the MXR m129 and did not see an inference that 40ms happens with choir voices, big band sections , or other physical world.  And 0.887ms rounds close enough to 1ms from my perspective for general discussion.

[spaceatl]

9 minutes ago, Mutonium said:

 

Hmmm...  They made sense to me.  I read @somebodyelse's post as using 40ms because that is what he is used to with the MXR m129 and did not see an inference that 40ms happens with choir voices, big band sections , or other physical world.  And 0.887ms rounds close enough to 1ms from my perspective for general discussion.

my poor attempt at a bad joke...I did not see that either....I just thought about the Helix being around 40' behind an MXR....Those things were slow...SPX-90 was kinda slow too...It's all just time alignment.....gotta deal with it on everything...What has always been interesting to me about sound traveling just under 1 foot in 1ms is that 1kHz propagates to just under 1 foot in length...

[b418me]

Thanks for all the input. I played around with it. I guess the harmonizer just sounds bad. Unfortunately not something Line6 seems to be good at. Oh well, I now use my harmonizer from the 1990s (via send/return) and its great :-)