Female speaking voice down in pitch

I would like to put a female speaking voice down in pitch. I am not sure which program would be best to use. The choices are Autotune, Cubase, Wavelab and Spectralayers.
Which program would you recommend for this and how would you proceed.
So far I’ve tried a few things but haven’t been satisfied.

What, specifically, have you tried? Which programs? What, exactly, was wrong with the results you got?

Frankly, it sounds like you’re getting ready to waste a lot of money.

1 Like

Oh no, I already have these programs. I’m certainly not going to spend extra money because of this.

My wife is doing some kind of radio play, and wants to make her own voice seem masculine. You know, to achieve that it is necessary to go more than two semitones lower.
With SpectraLayers it sounds horrible. I am probably doing something wrong. Wavelab with Autotune sounds reasonably good, but with artifacts. With Cubase I don’t yet know how to lower an entire recording, and before I pursue this further, I just wanted to ask if anyone has any experience with transposing speaking voices using these programs.

Formant shifting is usually the way to go with this, something like SoundToys Little Altar Boy…

1 Like

+1. Little AlterBoy is great for quick voice changes like this.

Or, since you already have Melodyne, you could do this manually by pitching down everything a couple of notes (maybe 3–5 semitones?) and also adjust the global formant settings in Melodyne down a bit.

You can try this technique with Cubase and adapt it to a simple vocal track:

@Knopf : However, this method does not take into account the formants
@eirik_myhr : Unfortunately, I only own the programs listed above. Maybe Audacity. But I expected that it should be possible with my Steinberg programs
@matjones : I will try again with SpectraLayers by manually separating the formant area. Possibly that was my mistake.

Well, ok. I was thinking you needed a simple transposing. Maybe you are needing that the result is really 100% convincing. I have no idea how to achieve it.

Ah, my mistake, I wrongly read that you own Melodyne.

The same thing can be done within Cubase’s own VariAudio. Pitch the vocal down some semitones and adjust the formant to your liking. Melodyne might sometimes sound even more convincing but VariAudio is close, and also very convenient since it’s already a part of every Cubase audio event, once you activate it. I’m sure you’ll find great tutorial videos of VariAudio on Steinberg’s YouTube channel.

Formant shifting, in very simple layman’s terms, 'cos i’m a very simple layman! :rofl: is essentially, in the case of a human voice, changing the length of the throat, longer = lower tone and shorter = higher in tone, a REALLY famous example is 'Kissing with confidence by Will Powers, it states on the wikipedia article that a ‘voice recorder’ was used (lol no sh*t!) in the production, i think they’re alluding to tape speed manipulation, but it sounds more like formant shifting to me. https://www.youtube.com/watch?v=AO128UhPivE

I work with a contralto singer and often use Little Altar Boy to get her voice to sound a bit more ‘girly’, it works a treat.
Melodyne also works really nicely as stated above and Waves Torque has formant shifting, but i’ve never used it for this purpose so can’t comment on that.
It seems there are a few free plugs which might be useful if you have a quick google for formant shifter plugin.


Yes thank you, for the information about the formants. However, I have known all this for years.
At the moment I’m just trying out whether I can produce a reasonably good result with the software available to me.
Autotune sounds good at first, but if you go down too many semitones it gets very scratchy.
I think it is also a significant difference whether I raise or lower the voice. If I raise a voice, the overtones eventually disappear into the inaudible range. A piccolo simply doesn’t have as many audible overtones as a bass clarinet, for example. If I make a voice lower, then the vibrations become slower and slower, and lose their mathematical support points in the recording. The samples are pulled apart. The software would then have to artificially fill the “free” area. The software has to add something that wasn’t there before. That is why it is so difficult (not only in music) to make a man out of a woman.

I would be interested to know if anyone here has experience with the programs I use and can give me tips for a better result.

Or try the PitchCorrect plugin that comes with cubase.
It uses a Yamaha algorithm, that can sound better than autotune in some circumstances.


@peakae Thank you for this tip. PitchCorrect plugin seems to produce, in a certain range, really a very good sound. Let’s see if I can still eliminate the artifacts that are generated anyway.

I would think that VariAudio would also produce better results than Autotune for this kind of task.
You could also experiment with changing the algorithm that VariAudio uses, to find the one that sounds the most convincing.

I wanted to give cubase another chance, so I tried again with VariAudio. Apparently Cubase is more capable of varying the pitch of a singing voice than a speaking voice. As you can see on the ScreenShot, parts of the words are not analyzed at all. These remain then in the original pitch. Furthermore, when all events are selected, it is very difficult to move them up or down with the mouse. This can only be done with the arrow keys. Furthermore, the possible zoom range is much too small to get a good overview in the vertical editing.

Ah. Right. Then PitchCorrect would be a better and simpler tool for this, if it’s anything like Little AlterBoy.

The best thing I know of for this sort of thing is Flux IRCAM Trax. It’s incredible.


All the Flux stuff is incredible, if not a bit complex, interface-wise. Things get clearer after a YouTube or two.


Unfortunately, I can only use the programs mentioned.

So if you are familiar with these programs and have a suggestion how to solve the problem with their help, please let me know.

I ought to have acknowledged your stipulations before posting the suggestion. Sorry about that.

My thinking put into an extended post, would go like this:

You’ve mentioned you can only use certain programs, but if you ever reconsider, there is a tool literally MADE for this, that was actually used in several motion pictures (it’s of that quality); so, if you get exhausted from efforts at tweaking the applications on your list to the limits of their functionality (we all enjoy a bit of that now and again), you might try a demo of the TRAX. :slight_smile:

Good luck with everything!

Little AlterBoy by Soundtoys is currently on sale for $29, so it might be a good time to get it. It allows for easy pitch and formant shifting which makes changing a voice from female to male or vice versa pretty straight forward.