|
Dear Adrian - dear list members
thanks for your comment!
- Carl Seashore (1938) published graphical pictures of sung
melodies by professional singers. I always wondered how this was
achieved. But Seashore did not give the slightest reference!
Your comment and hint to EGG (electroglottography) being the most
accurate reference for measuring vocal pitch is very interesting!
Is it possible that Seashore and others used EGG to produce those
pictures of sung melodies?
It is amazing that even nowadays, with all our knowledge and
technology, the analysis of pitch is still a difficult enterprise,
when it comes about the analysis of big data, e.g., large amounts of
sung melodies or songs. In this context, I often see information on
pitch reduced to unreliable and useless "data" by inacceptable,
quick-and-dirty procedures ...
Could you please tell the advantages of the non-invasive EGG? Could
this be a method to cope with big data? And maybe even solve the
problem with the interferences in acoustic signals due to polyphonic
music making?
Thanks and best wishes, Stefanie
Prof. Dr. Dr. h.c. Stefanie Stadler Elmer
- University of Zürich http://www.psychologie.uzh.ch/institut/angehoerige/dozierende/stadler.html
- Schwyz University of Teacher Education, www.phsz.ch
Am 06.05.17 um 16:14 schrieb Adrian
Fourcin:
Dear Stefanie,
Yes, the book by Wolfgang Hess is still a mine of information
and his recommendation that the EGG signal provides the most
accurate reference for voice "pitch" measurements still holds --
because the percept of voice pitch is primarily dependent on the
acoustic excitation of the vocal tract that is derived from
vocal fold contact.
Best wishes,
Adrian
Dear list members
Ani, thanks for sharing this fascinating website with Indian
Ragas on Febr. 28 2017, where the singing is analyzed and
visualized in real time!
It is very nicely done, indeed. Yet, different from you, Jim,
I do not think that these analyses are "error free", since
they use praat to calculate pitch out of the signal. This is
one way to do it, yet, there are other possibilities, see
e.g.,
Hess, W. (1983). Pitch determination of speech
signals: Algorithms and devices. New York: Springer.
Praat is a fantastic and powerful tool! It is
optimized for speech sounds, and pitch analysis is
done with one single algorithm! Depending on the
signal, there could be more suitable ones.
My impression is that some people are not aware
that pitch is not very well defined. Various
algorithms may yield various results. Any calculation
and visualization of pitch in a signal is an
approximation by the use of a specific algorithm.
In our research on children's singing, we analyze pitch at
the basis of a program that offers two different
algorithms, both selected for the typical sound quality of
singing of and for children, e.g., without vibrato. This
program and related instructions are available here for
free:
http://mmatools.sourceforge.net/
Have fun and enjoy!
Stefanie Stadler Elmer
Am 01.03.17 um 17:09 schrieb James W. Beauchamp:
Dear Ani,
Thanks for informing the Auditory list about the Music in Motion
web site!
It's nice to see (and hear) these results. They seem to be error free.
I'm assuming that the drone accompaniment is on a different track;
otherwise it would interfere with the voice pitch trace. Looking the
site's Introduction pull-down, it appears that the program is using
the Praat algorithm for pitch detection. I would think that the same
program would work for western singing except that the vertical scale
should ideally correspond to standard equal-tempered (ET) notes.
The software that I've been developing over many years, SNDAN,
includes a program, that uses the Two-Way-Mismatch method (written by
Robert C. Maher*), that makes pitch-vs.-time plots with the ET scale,
but it's not error free and it's not real time. The program was
written in C for Unix (e.g., Mac terminal or Linux). However, in 2008
Ugur Guney wrote an offshoot of this program that runs in real time,
called Freqazoid. It is written in Java and is a free download at:
<https://code.google.com/archive/p/freqazoid/downloads>
Ugur is now (in his spare time) working on a real-time web-based
version of freqazoid.
Best,
Jim
James W. Beauchamp
Research Professor
Professor Emeritus of Music and Electrical & Computer Engineering
University of Illinois at Urbana-Champaign
email: jwbeauch@xxxxxxxxxxxx (also: jwbeauch@xxxxxxxxxxxxxxxxxxxxxx)
WWW: http://ems.music.uiuc.edu/beaucham
http://www.ece.illinois.edu/directory/profile/jwbeauch
* References on the Two-Way-Mismatch method:
R. C. Maher and J. W. Beauchamp, "Fundamental frequency estimation
of musical signals using a two-way mismatch procedure", J. Acoust.
Soc. Am. 95(4), 2254-2263 (1994).
J. W. Beauchamp, "Analysis and Synthesis of Musical Instrument Sounds",
in Analysis, Synthesis, and Perception of Musical Sounds: Sound of
Music, J. W. Beauchamp, Ed., Springer, N. Y., pp. 1 - 89 (2007).
***********************************************************************
Original message:
From: "Patel, Aniruddh D." <a.patel@xxxxxxxxx>
Date: Tue, 28 Feb 2017 15:02:13 +0000
To: AUDITORY@xxxxxxxxxxxxxxx
Subject: Melodic movies
Dear List,
I have a question... there is a website where you can hear professional
Indian classical singing and watch the pitch of the singer's voice on
a scrolling graph which captures all the nuances in vivid detail.
https://autrimncpa.wordpress.com/
(Click on a Raga on the left menu bar, then after the page loads, scroll
down and watch the associated movie.)
My question is: does anyone know of a similar resource for western
classical singing?
Thanks,
Ani Patel
Aniruddh D. Patel
Professor
Dept. of Psychology
Tufts University
490 Boston Ave.
Medford, MA 02155
Senior Fellow
Canadian Institute for Advanced Research (CIFAR)
Azrieli Program in Brain, Mind, & Consciousness
a.patel@xxxxxxxxx<mailto:a.patel@xxxxxxxxx>
http://ase.tufts.edu/psychology/people/patel/
|