Can you change a voice to make it sound GOOD?

[vandy12]

Is there an electronic device that could be developed which could automatically change the actual tone quality of a voice to turn someone's off key, terrible tone vocal into something in tune with a more pleasant tone? The non-singers dream come true? I also thought for an extra ten dollars we could add in the Elvis attachment. Or the Frank Sinatra or Joe Perry Attachment, and when you sing thru the mike the magic box makes you sound just like them. I do not know enough about the physics of sound to know if this is even possible. I thought a lot of people who have music in their heart might enjoy it more if they were able sing better if they are a non-singer.  In tune with a better tone than their natural voice, so they don't drive us us crazy at the Karaoke parties. Not talking about Autotune or Melodyne. My idea is way different. Maybe just another one of my hair-brained ideas.

[DIY Bass]

I know from experience that even drowning my voice in reverb does not improve anything.

[antonis]

the actual tone quality

Tone unit is Hertz but what's the quality unit..??

[anotherjim]

Is it the audio equivalent of the Deepfake video you want?

[vandy12]

Very good question Atonis!! How about alter the tone to pitch somehow automatically.
Actually you can do this right now and sing with autotune automatically tuning as you sing?
I have heard about this, never done it, and I use Melodyne. I don't think Melodyne has this
feature unless it is in upgrade. Been awhile since I used it. So you use something like an
auto autotune while singing and then for the quality it can only be some kind of filter right?
Is there any kind of a filter you could use to change the quality of the pitch? I have no clue
how to do this and don't know if it can be done.

[vandy12]

Have never heard of Deep Fake. Googled it. Pretty wild!! I guess you could call this an equivalent and more. Sounding like
Elvis would not be the main part. The main part would be able to correct your voice pitch and somehow alter the tone
quality to make it sound....better than it did before. Maybe not professional quality, but turning people who can't sing into those
who can.

[Radical CJ]

I think what you are talking about would be an extremely complicated form of envelope filter.

[Mark Hammer]

Making a voice that's off-pitch be more accurate (i.e., ON-pitch) was the basis for the original development of Auto-Tune.  Of course, we have since come to associate it with alternate uses in hip-hop, Cher, and other applications where it can be used much like a vocoder.  But fixing a singing voice that is a little "pitchy" (as former American Idol judge Randy Jackson used to say) was the impetus for creating Autotune.  It can now be purchased in a pedal from what I understand.

Some producers also use harmonizers/pitch-shifters to add just a touch of octave-down to voices that may otherwise sound a little reedy and thin, in order to make them sound a little fuller.  I suppose one could use an octave divider for that application, since voice is a monophonic signal.  But analog octave-dividers are notoriously fickle.  Moreover, they are best-behaved when the original source, be it guitar, bass, or voice, is pitch-stable.  A voice that wavers is unlikely to be a prime candidate for analog octave-division.

[Rob Strand]

Back around 2000 someone developed voice changing software.   People have been digitally modelling voice since the 70's the general idea is to map the features of one voice onto another.   IIRC the stuff around 2000 had a press release along the lines of *speak* like David Letterman.   There's the fun side to this stuff and an industry side but there's also the dangerous side of misuse with an intent to deceive.

[11-90-an]

A very amusing-to-use text to speech site that I sometimes use for school video projects when I don't feel like recording my voice is https://vo.codes/

All you have to do is select the person/character whos voice you want, type some words in the box, click "speak", then after a few minutes that person speaks that text to you back! Although not perfect, it has quite a resemblance and has lots of comedic potential

However, like what Rob said, sooner or later, the more sophisticated things like these get, it can be hard to distinguish from the truth, and that is where the chaos begins...

[iainpunk]

auto-tune, it makes everything in tune, and i almost bought an auto-tune rack mount unit once, since it was a funky vibrato effect when doing pitch bending or sliding on a guitar. a bit to expensive when i saw the price tag.
there is also standard auto tune in some DAW's, like FL Studio has it standard.

cheers

[vandy12]

Auto tune will help with pitch, but then how to do you take the new pitch that has a bad tone and a make it a good tone?

[mikepurcell]

Melodyne does time & pitch correction and it can also shift formants in real time. It's easy to make a male singer sound like a female and vice versa. As far as matching the Timbre of another singer, I'm sure it's possible if someone set their mind to it.

[GGBB]

No device necessary - just perform Bob Dylan covers. Bad voice and off-key will fit right in.

[dennism]

I have the TC Helicon Mic Mechanic on my pedal board.    It has some reverb and some auto tune and will correct you if you are slightly off pitch without any noticeable extreme auto pitch warbling.   It works well for me, but for my bandmate, who makes Bob Dylan sound like Pavoratti, it just didn't have enough to pull his vocals into the right key.    Best solution there is to just take their microphone away altogether....

[Mark Hammer]

No device necessary - just perform Bob Dylan covers. Bad voice and off-key will fit right in.

Or Neil Young:  https://en.wikipedia.org/wiki/Tears_Are_Not_Enough#Recording_process

[anotherjim]

Convolution can do it. It's used to reform sounds by capturing the response of the original source to an impulse. Done all the time for speaker cab and reverb simulation etc...

What you do is put a microphone (absolutely transparent omni calibration mic) in the singer's mouth and fire a starting pistol to provide the impulse. A recording of the sound inside the singer will carry the response to the impulse. Alternatively, if you like the singer, you can input a noise source at the opposite end to their mouth.

Yes, I'm being silly! 

[GGBB]

^
Makes you wonder how far we are away from voice modelling. Dial-up the patch for your favourite vocalist and sing into the microphone. Instant Aretha, Lennon, Garfunkel, Sinatra, ...

[vandy12]

Crack me up GGBB. I've got a couple recordings of Kris Kristofferson in the studio under some adult medication. His vocals were not so good. Bob Dylan used to get fruit and veggies thrown at him. Horrible singer AND guitar player. What was he thinking?

[amptramp]

Maybe the best advice is to have the offender take some singing lessons.  It might be cheaper than electronics and it can last a lifetime without any maintenance.  I sang in a church choir at one point and I ended up a bit better than when I started because the choir director gave specific advice to anyone who needed it.

This topic reminds me of some of the advice I see on miata.net about how to be a better driver - take a high-performance driving course because the money spent there is worth more than all the tires, wheels and aftermarket stuff you can put on a car.  I was at one autocross event where there were a lot of cars, some high-dollar and high horsepower in attendance and the best time before the last guy tried out was 51 seconds for the course.  The last guy was driving a Lotus Elan with the smallest engine, narrowest tires and smallest wheels of all the cars there.  But he was an absolute master at the wheel and he smoked it through the course at 47 seconds.

Never underestimate the value of getting competent at what you do.