Learning to sing by spectograph emulation ...

[six20aus]

I've been so busy I haven't yet had a chance to sing and use the spectograph software but I have been thinking about it a lot and this might be a really stupid question but here goes...

If you had an example soundfile and spectograph of a trained singer who could easily sing throughout their required range would it not be possible to train your muscle memory merely by matching the positioning of the formants as they do ?

So let say A4 is a troublesome note to you but you have the spectrogram of someone that can do it easily - you load it up on screen and sing around it, varying your setup and placement until you match it and in so doing train your own muscle memory and setup.

Would this work or are there other factors involved ?

[Steven Fraser]

I've been so busy I haven't yet had a chance to sing and use the spectograph software but I have been thinking about it a lot and this might be a really stupid question but here goes...

If you had an example soundfile and spectograph of a trained singer who could easily sing throughout their required range would it not be possible to train your muscle memory merely by matching the positioning of the formants as they do ?

So let say A4 is a troublesome note to you but you have the spectrogram of someone that can do it easily - you load it up on screen and sing around it, varying your setup and placement until you match it and in so doing train your own muscle memory and setup.

Would this work or are there other factors involved ?

six20aus: I think this question is being sent to me :-)

You can learn quite a bit about how other singers are managing their resonance, and if your ear is good, hear the vowel forms that they are using to accomplish the formant alignment with the harmonics. What the spectrograph does not tell you is what they are doing at the laryngeal level... with their phonation.

However, just sing into it a while, and then lets talk about what it is telling you.

[six20aus]

haha in a roundabout way Steve - trying to be inclusive !!

I guess what I have been wondering is if I could play scales or acapella sections from pro singers and see exactly what they're doing and just try to emulate it directly... - should also take my own hearing perceptions out of the loop (like having a good teacher).

What the spectrograph does not tell you is what they are doing at the laryngeal level

Interesting point. What does any singer do at the laryngeal level that's different ? He/She may have 'slightly' larger/smaller folds but about the only thing they could do 'differently' would be to raise or lower their larynx would it not ? (Stay with me here for a second...)

If I can 'match' the formant and harmonics and 'match' the tone (differences in genetic makeup aside) then by that logic the 'setup' at the laryngeal level must be the same must it not ?

[Steven Fraser]

I guess what I have been wondering is if I could play scales or acapella sections from pro singers and see exactly what they're doing and just try to emulate it directly... - should also take my own hearing perceptions out of the loop (like having a good teacher).

Interesting point. What does any singer do at the laryngeal level that's different ? He/She may have 'slightly' larger/smaller folds but about the only thing they could do 'differently' would be to raise or lower their larynx would it not ? (Stay with me here for a second...)

If I can 'match' the formant and harmonics and 'match' the tone (differences in genetic makeup aside) then by that logic the 'setup' at the laryngeal level must be the same must it not ?

six20aus: You can certainly see what they are doing with vowel pronunciation, twang, etc. And, the display will be objective, as you say. You can even see the effects of larynx raising, or lowering techniques, and compare your overall tone quality with some other singer's. If you are interested to do that, then the tool can help you.

But what the tool does not know is the ways in which your voice is naturally different than another singer, nor how to instruct you in what you should change to go from where you are currently to this tone quality you wish to emulate. Its like the speedometer on your car... it displays your rate of speed, but does not know anything about how the engine works, or that the engine speed increases when you press the gas pedal more. The speedometer reading is just a measurement... feedback to the driver, of how the car is responding to the instruction given by that driver in the pressing of the pedal.

So, for example, you are working on the notes from C4 to G4, you will be able to see the displayof the sounds produced by the singer, but you will not know how the singer did it. That you will have to discover, or learn from a teacher, or a combo of both.

But circling back to the original concept for a moment... the tool will be like another pair of ears in the room, ears you can trust to tell you the truth about the sound you are making. One way you can do what you've said you'd like to is this: there is a control called 'capture' on the bottom-right control panel. That will take a 'snapshot' of whatever sound is in the air in the room, that the mic can pick up. If, for example, you are playing a recording of an artist on your CD player, or even over Youtube while the spectrogram is running, you can see the display of it. Press the 'capture' when you want to get to a note to examine for a minute.

The captured image will be in green, and the current sound in the room, for example, with you singing, will be traced in Blue. As long as you do not press the capture button again, you can do some playing around with your vowel forms, twang, vibrato, whatever, and see your own blue trace and the artist's green trace at the same time.

I hope this helps.

[Robert Lunte]

Ive been taking lessons from Steve on the spectragraph science. I intend to continue this education and use it to help TVS singers. I think I understand how to read the application, more or less... Im still working on how the information will apply to helping singers. As Steve points out above, I think its essentially, getting practice mapping your preferred vocal tone with what that looks like on the graph. If the vocal tone is not your preference, then the graph will indicate that as well. So, in a sense, the spectrograph reinforces what your your ear and intuition is telling you on what sounds "good" or "better".

It can also be applied to introduce formant and "dominant harmonic" talk-track to the pedagogy. Any time a student of singing can gleem a little bit of surface level science on how the voice works in terms of physics and physiology, the learning process is greatly enhanced.

I think what the application could use is another visual that shows human anatomy, specifically, the upper vocal tract and formants. If you could see both what the tongue, upper palette, pharynx, lips, jaw, etc... look like in a simulation and then cross reference that to the wave signatures of the spectragraph... you have the makings of a very powerful tool. More on this later...

[chanteurmoderne]

That's an interesting idea, Robert. The only issue I see with it is that you can affect the harmonics by moving one of a number of elements in the vocal tract - so that the extrapolation of the vocal tract cross section would be unlikely to be accurate (I know there are a few people working on vocal tract mapping, but it's proving very difficult to make reliable) - for example, I teach a form of medialisation of vowels to most of my singers - this means that vowel formant differences are far more heavily made with the tongue than with the rest of the vocal tract, so it kind of messes up the data a bit for vocal tract mapping.

It sure would be an excellent tool if we could get it going though, eh?! :)

[Estef]

Hi guys, sorry to intrude into Six's post but I've been reading this with a lot of interst.. Maybe Allan or Steve (or anyone who knows really) could share any sites or books where we can learn more about reading spectrographs? Thanks!