Good source for detailed formant data?

Anything modular synth related that is not format specific.

Moderators: lisa, Kent, Joe.

Post Reply
wechard
Common Wiggler
Posts: 146
Joined: Tue Oct 24, 2017 7:54 pm

Good source for detailed formant data?

Post by wechard » Sun Feb 02, 2020 12:45 pm

Hi Everyone -- Can anyone recommend a good source for detailed data about vowel formants? I've found a lot of sites, including some mentioned in other Muffs threads, but they all seem to be limited in one way or another. What I'd love is to find data for a wide range of vowels that ticks all the following boxes: (i) 3-4 formant frequencies given for each vowel; (ii) also data for relative amplitudes and bandwidths of all formants; (iii) ideally, vowels named using IPA, or at least really unambiguous casual descriptions (not just something like 'a', 'e', etc.). Thanks...

User avatar
3hands
Veteran Wiggler
Posts: 530
Joined: Mon Nov 19, 2018 6:36 pm
Location: Ottawa Ontario

Re: Good source for detailed formant data?

Post by 3hands » Sun Feb 02, 2020 8:37 pm

You could always just put yourself in front of your synth, and start experimenting?
Gum is fun, but not on a cat.

My minds an art gallery.

wechard
Common Wiggler
Posts: 146
Joined: Tue Oct 24, 2017 7:54 pm

Re: Good source for detailed formant data?

Post by wechard » Sun Feb 02, 2020 8:47 pm

Sure, the experiments are ongoing (have been for years). But I also enjoy finding and thinking about information that I wouldn’t necessarily discover on my own.

User avatar
3hands
Veteran Wiggler
Posts: 530
Joined: Mon Nov 19, 2018 6:36 pm
Location: Ottawa Ontario

Re: Good source for detailed formant data?

Post by 3hands » Sun Feb 02, 2020 8:55 pm

Completely fair point!!!

I’m curious as to what you’re doing to achieve your formants, as I haven’t played around with that stuff since the waldorf microwave 2!! (I’m showing my age). However, I am fascinated with the possibility of formants. Sorry I couldn’t be of any help!
Gum is fun, but not on a cat.

My minds an art gallery.

electricanada
Ultra Wiggler
Posts: 988
Joined: Wed Nov 29, 2017 12:26 am
Location: Norfolk, VA

Re: Good source for detailed formant data?

Post by electricanada » Sun Feb 02, 2020 9:02 pm

3hands wrote:
Sun Feb 02, 2020 8:55 pm
Completely fair point!!!

I’m curious as to what you’re doing to achieve your formants, as I haven’t played around with that stuff since the waldorf microwave 2!! (I’m showing my age). However, I am fascinated with the possibility of formants. Sorry I couldn’t be of any help!
Easiest place to start is two bandpass filters in parallel.
Eléctrica (electric) Nāda (the yoga of sound).

User avatar
mookmoof
Common Wiggler
Posts: 233
Joined: Thu Oct 18, 2012 10:59 pm
Location: tampa

Re: Good source for detailed formant data?

Post by mookmoof » Sun Feb 02, 2020 9:43 pm

I thought it was 3 bandpass in series :?

User avatar
3hands
Veteran Wiggler
Posts: 530
Joined: Mon Nov 19, 2018 6:36 pm
Location: Ottawa Ontario

Re: Good source for detailed formant data?

Post by 3hands » Sun Feb 02, 2020 9:53 pm

Oh here we go!

Haha!

I will try both and see where both experiments lead me!
Gum is fun, but not on a cat.

My minds an art gallery.

User avatar
cptnal
Super Deluxe Wiggler
Posts: 3591
Joined: Wed Jun 14, 2017 2:48 am
Location: People's Republic of Scotland

Re: Good source for detailed formant data?

Post by cptnal » Mon Feb 03, 2020 3:54 am

Sounds like you'd have to dive into the academic literature for the level of detail the OP's looking for. How about checking out the sources in the reference sections on Wikipedia? Search for formant synthesis or suchlike...

User avatar
Muff McMuff
Wiggling with Experience
Posts: 417
Joined: Sun Jul 29, 2018 7:20 am
Location: CNX and MAN

Re: Good source for detailed formant data?

Post by Muff McMuff » Mon Feb 03, 2020 4:40 am

The Doepfer A-127 manual has some of this kind of info if you have not seen it already.
Screen Shot 2020-02-03 at 16.37.28.png
You do not have the required permissions to view the files attached to this post.

wechard
Common Wiggler
Posts: 146
Joined: Tue Oct 24, 2017 7:54 pm

Re: Good source for detailed formant data?

Post by wechard » Mon Feb 03, 2020 11:54 am

cptnal wrote:
Mon Feb 03, 2020 3:54 am
Sounds like you'd have to dive into the academic literature for the level of detail the OP's looking for. How about checking out the sources in the reference sections on Wikipedia? Search for formant synthesis or suchlike...
Yes, I think the academic stuff is what I'm looking for. Although even there, it's interesting how partial the data actually is in a lot of cases (I did browse most of the articles linked from the Wikipedia entry). Maybe relative amplitude and bandwidth are more variable, and therefore less important to dial in precisely. I was just wondering if anyone has suggestions for especially rich data sets buried in the literature, as a time-saver.
Muff McMuff wrote:
Mon Feb 03, 2020 4:40 am
The Doepfer A-127 manual has some of this kind of info if you have not seen it already.
Thanks, I hadn't seen that. Also interesting that DD cites the Strange book -- I'll have to check that too.
3hands wrote:
Sun Feb 02, 2020 8:55 pm
I’m curious as to what you’re doing to achieve your formants, as I haven’t played around with that stuff since the waldorf microwave 2
Just about anything that will let you do several parallel bandpass filters will work. I started out playing with multing the signal to different individual filters, whatever I had on hand, and then mixing them back together. I also played around with using a Mutable Shelves, but I find the interactivity of the channels on that filter hard to predict and control (probably just need more practice). Most recently, I picked up a Doepfer A-104 Trautonium filter, and that's what's gotten me back into it.

User avatar
oldenjon
Veteran Wiggler
Posts: 676
Joined: Sun Feb 26, 2012 1:28 am
Location: Sacramento, CA

Re: Good source for detailed formant data?

Post by oldenjon » Mon Feb 03, 2020 12:00 pm

Idk if you’ve seen this article already
https://www.soundonsound.com/technique ... -synthesis
"68% of all statistics are made up on the spot"

http://synthscalledbeasts.tumblr.com/

Fabong
Common Wiggler
Posts: 101
Joined: Sun Dec 10, 2017 10:07 am

Re: Good source for detailed formant data?

Post by Fabong » Mon Feb 03, 2020 4:44 pm

I wonder if Doepfer's A-128 Fixed Filter Bank would be accurate enough to get in the ballpark with regards to creating these vowel formants.
Avant Garde a clue

wechard
Common Wiggler
Posts: 146
Joined: Tue Oct 24, 2017 7:54 pm

Re: Good source for detailed formant data?

Post by wechard » Mon Feb 03, 2020 5:53 pm

oldenjon wrote:
Mon Feb 03, 2020 12:00 pm
Idk if you’ve seen this article already
https://www.soundonsound.com/technique ... -synthesis
Yes, I've seen that (but thanks). It was actually the later part of that article, where he gives just a little bit of amplitude and bandwidth information, that made me wonder why most other sources don't go into that much detail.
Fabong wrote:
Mon Feb 03, 2020 4:44 pm
I wonder if Doepfer's A-128 Fixed Filter Bank would be accurate enough to get in the ballpark with regards to creating these vowel formants.
Sort of, but not quite (I have an A-128, and have tried it for formants). One issue is that the frequency bands aren't tunable, so they often aren't quite where you want them to be for a particular vowel. Also, the bandwidths are quite a bit wider than the relatively narrow bandwidths of many formants. So you can sometimes get close, but ultimately it's not the right tool for that particular job. Great filter for other things though...

KSS
Ultra Wiggler
Posts: 858
Joined: Mon Jan 25, 2016 7:28 am

Re: Good source for detailed formant data?

Post by KSS » Mon Feb 03, 2020 6:23 pm

Would expect Parametric Equalizers to be far better than FFBs for this. Plenty of 3 and 4 band rack and standalone modules can be found.

For the data tables, would call a local ENT doctor and ask. Ear, Nose, Throat specialist.
Next would ask a vocal coach for resources.

Vocal coach and ENT may not have the data you seek, but would probably be well-set to provide useful leads.

Then to local university speech lab resources. Search for premier speech labs around the world. What you're asking for is what they do.

Finally, Bell labs did quite a bit of research into this in the late 50 and early 60's. And ongoing after that. Their in-house newsletter had articles dealing with this, can remember reading a couple of these in a university library *many* years ago.
edit: this last was triggered by a Polyphony Magazine article which was not about formants, but made me aware that the Bell labs had an internal newsletter. So went looking for it. And also found the JAES, where moog first published his module circuits in '64. The Audio Engineering Society's Journal would be another resource expecvted to have discussed this.

All of which points to a good first stop being the reference librarian at a *large* university or government library. They were the kings and queens of information before the net existed, and the good ones still are. /edit

electricanada
Ultra Wiggler
Posts: 988
Joined: Wed Nov 29, 2017 12:26 am
Location: Norfolk, VA

Re: Good source for detailed formant data?

Post by electricanada » Mon Feb 03, 2020 9:40 pm

Image
Eléctrica (electric) Nāda (the yoga of sound).

electricanada
Ultra Wiggler
Posts: 988
Joined: Wed Nov 29, 2017 12:26 am
Location: Norfolk, VA

Re: Good source for detailed formant data?

Post by electricanada » Mon Feb 03, 2020 9:43 pm

More useful charts in this one, including amplitude numbers in dbs:
https://www.jvoice.org/article/S0892-1 ... 4/fulltext
Eléctrica (electric) Nāda (the yoga of sound).

Fabong
Common Wiggler
Posts: 101
Joined: Sun Dec 10, 2017 10:07 am

Re: Good source for detailed formant data?

Post by Fabong » Tue Feb 04, 2020 9:20 am

wechard wrote:
Mon Feb 03, 2020 5:53 pm
Fabong wrote:
Mon Feb 03, 2020 4:44 pm
I wonder if Doepfer's A-128 Fixed Filter Bank would be accurate enough to get in the ballpark with regards to creating these vowel formants.
Sort of, but not quite (I have an A-128, and have tried it for formants). One issue is that the frequency bands aren't tunable, so they often aren't quite where you want them to be for a particular vowel. Also, the bandwidths are quite a bit wider than the relatively narrow bandwidths of many formants. So you can sometimes get close, but ultimately it's not the right tool for that particular job. Great filter for other things though...
Thanks for responding, I go back and forth over whether I want the A-128 FFB or the A-104 Trautonium Filter so it's nice to hear these opinions.
Avant Garde a clue

jorg
Wiggling with Experience
Posts: 422
Joined: Fri Apr 03, 2015 9:38 am
Location: East Coast USA

Re: Good source for detailed formant data?

Post by jorg » Tue Feb 04, 2020 9:24 am

Years ago I had quite a bit of success with a Reaktor formant toy. I just threw in everything I could find. I put a big knob on the front to select any one of about a dozen sets of formants, drawn from different sources (different languages, accents, etc). I also added knobs to adjust age and sex by pushing everything up or down in specific ways.

I have an A-127 and dearly love it. I haven't had a ton of success getting it to be very vocal-ish, but haven't invested a lot of time either.

One thing I found in the Roland VP-330 (though not actually their invention) is a series pair of bandpass filters, tuned 20% apart and set to a Q of 6.5. Makes a very nice formant: just enough width to capture a decent chunk of energy, but steep enough slopes to be distinctly separated from the next (similar) band. They use this filter in the analysis and synthesis banks of the vocoder, and also in the bank that shapes the choir sounds.

User avatar
oberdada
Common Wiggler
Posts: 208
Joined: Tue Nov 01, 2016 12:06 pm

Re: Good source for detailed formant data?

Post by oberdada » Tue Feb 04, 2020 12:20 pm

You might find the Praat software useful.

Ladefoged has written some relevant books. I have only read Vowels and Consonants which covers languages all over the world and is quite interesting, though not sufficiently detailed for you purposes.
The Csound book actually has a formant table with five formants for vocal ranges from soprano to bass.
Johan Sundberg has studied the singing voice extensively and written papers about it.
Lots of the papers can be found in the Journal of the Acoustic Society of America.

wechard
Common Wiggler
Posts: 146
Joined: Tue Oct 24, 2017 7:54 pm

Re: Good source for detailed formant data?

Post by wechard » Tue Feb 04, 2020 1:58 pm

Thanks to everyone for that latest round of suggestions -- lots of good information here...

Post Reply

Return to “Modular Synth General Discussion”