[Csnd] spectral analysis of oboe multiphonics
| Date | 2012-10-11 00:37 |
| From | peiman khosravi |
| Subject | [Csnd] spectral analysis of oboe multiphonics |
| Dear all, Soon I'll be working on the spectral analysis of a few hundred and previously unknown Oboe multiphonics. Naturally I need a script to do this. The aim of the project (not mine) is to extract the most prominent frequencies rather than show the spectral evolution of the sound. My question is, would it better to extract the most prominent partials then average out their frequency and amplitude fluctuations or is it more feasible to first extract the steady-state of the note for analysis and simply use a frame from the middle of the analysis? ATS seems to be giving me good results and will also allow me to patch analyse everything. I can then read the analysis files into csound and convert the frequency to fractional midi values (with a resolution of 8th of a whole tone): this is the easy part. Any suggestion are really welcome. Many Thanks Peiman |
| Date | 2012-10-11 21:38 |
| From | Richard Dobson |
| Subject | Re: [Csnd] spectral analysis of oboe multiphonics |
For stable multiphonics with a clean steady-state (something the oboe can do very well), a single "representative" frame should be fine. There would not be much in the way of spectral evolution to worry about - no different from a normal fixed note. However, you can also get multiphonics that sound like a bluebottle trapped inside a snare drum, in which case there may not be a representative frame to pick - civil war inside the air column. You may have to admit defeat and listen to them all at least once. An alternative is to multisample: pick three frames from just-after-start, middle and just-before-end, and if these are significantly different, separate those files out for special treatment. I would read "previously unknown" as "undocumented". Oboists (and wind players generally) can be tricsky that way. One of the most famous flute "multiphonics" (listed in the Bartolozzi book) is nothing more than a standard top D, underblown (i.e. as played by a beginner). Richard Dobson On 11/10/2012 00:37, peiman khosravi wrote: > Dear all, > > Soon I'll be working on the spectral analysis of a few hundred and > previously unknown Oboe multiphonics. Naturally I need a script to do > this. The aim of the project (not mine) is to extract the most prominent > frequencies rather than show the spectral evolution of the sound. My > question is, would it better to extract the most prominent partials then > average out their frequency and amplitude fluctuations or is it more > feasible to first extract the steady-state of the note for analysis and > simply use a frame from the middle of the analysis? ATS seems to be > giving me good results and will also allow me to patch analyse > everything. I can then read the analysis files into csound and convert > the frequency to fractional midi values (with a resolution of 8th of a > whole tone): this is the easy part. Any suggestion are really welcome. > > Many Thanks > > Peiman |
| Date | 2012-10-11 22:03 |
| From | peiman khosravi |
| Subject | Re: [Csnd] spectral analysis of oboe multiphonics |
| Thanks Richard, that's really helpful. Indeed I have already come across a few of those bluebottle notes already where some kind of beating across between the partials that are constantly gliding. So I will take your advice and try the selective frame option. If that didn't work then I'll just have to listen to everything! In most cases you would be right but in this case they actually are previously unknown since this is a especially designed Oboe to allow all sorts of micro-tonal fingering (up to 8th note I think). http://www.christopherredgate.co.uk/ I remember Chris saying that the number of possible multiphonics on this oboe has increased up by a few hundred! Best, Peiman On 11 October 2012 21:38, Richard Dobson <richarddobson@blueyonder.co.uk> wrote: For stable multiphonics with a clean steady-state (something the oboe can do very well), a single "representative" frame should be fine. There would not be much in the way of spectral evolution to worry about - no different from a normal fixed note. However, you can also get multiphonics that sound like a bluebottle trapped inside a snare drum, in which case there may not be a representative frame to pick - civil war inside the air column. You may have to admit defeat and listen to them all at least once. |
| Date | 2012-10-11 22:17 |
| From | Tarmo Johannes |
| Subject | Re: [Csnd] spectral analysis of oboe multiphonics |
| Richard, you are very right about the flute d3 -fingering multiphonic (like all the 3rd octava fingerings produce multitones). Peiman, I appreciate very much the project that you are on now - to track down the most important fequencies of the spectra - the traditional notation is often so approximate and sometimes also misleading, the usual spectrograms too complex from the other hand. Of course, a multiphonic produced with one fingering can vary enourmosly, depending on the blowing pressure, instrument, qualities of the reed, pressure of the lips etc etc. Anyway, if you work out some Csound or other patch to track the frequencies of multitones, would you agree to share it? I would be most interested to try it out with flute. If the instrument or patch could work also in realtime, it were even more interesting - it would enable to also examine the stability and "trustability" of the fingerings. greetings,'tarmo 2012/10/11 peiman khosravi <peimankhosravi@gmail.com> Thanks Richard, that's really helpful. Indeed I have already come across a few of those bluebottle notes already where some kind of beating across between the partials that are constantly gliding. So I will take your advice and try the selective frame option. If that didn't work then I'll just have to listen to everything! |
| Date | 2012-10-11 22:23 |
| From | peiman khosravi |
| Subject | Re: [Csnd] spectral analysis of oboe multiphonics |
| Hi Tarmo, Of course I will share whatever I come up with. But it is likely to be none-real-time and based on ATS analysis. Due to the share volume of samples! Best, Peiman On 11 October 2012 22:17, Tarmo Johannes <tarmo.johannes@otsakool.edu.ee> wrote: Richard, |
| Date | 2012-10-11 22:36 |
| From | Richard Dobson |
| Subject | Re: [Csnd] spectral analysis of oboe multiphonics |
Ah, I see, interesting! I'm totally out of the loop these days, but I do have some history with oboists - my father (Michael Dobson) was a professional player, and teacher at the RAM (when not doing all the other stuff he did such as running an orchestra and fly-fishing), and my composition teacher and general mentor even before going to the RCM was Edwin Roxburgh. If you meet him, say "Hi" from me! Richard Dobson On 11/10/2012 22:03, peiman khosravi wrote: > Thanks Richard, that's really helpful. Indeed I have already come > across a few of those bluebottle notes already where some kind of > beating across between the partials that are constantly gliding. So I > will take your advice and try the selective frame option. If that didn't > work then I'll just have to listen to everything! > > In most cases you would be right but in this case they actually are > previously unknown since this is a especially designed Oboe to allow all > sorts of micro-tonal fingering (up to 8th note I think). > http://www.christopherredgate.co.uk/ > > I remember Chris saying that the number of possible multiphonics on this > oboe has increased up by a few hundred! > > Best, > Peiman > |
| Date | 2012-10-11 22:42 |
| From | peiman khosravi |
| Subject | Re: [Csnd] spectral analysis of oboe multiphonics |
| Of course will do! I didn't know you were at RCM! I have been to a few of Chris' workshops and the possibilities are apparently enormous. I'm really looking forward to receiving all the samples. BTW I should have said 8th tone rather than "note". Best, Peiman On 11 October 2012 22:36, Richard Dobson <richarddobson@blueyonder.co.uk> wrote: Ah, I see, interesting! I'm totally out of the loop these days, but I do have some history with oboists - my father (Michael Dobson) was a professional player, and teacher at the RAM (when not doing all the other stuff he did such as running an orchestra and fly-fishing), and my composition teacher and general mentor even before going to the RCM was Edwin Roxburgh. If you meet him, say "Hi" from me! |
| Date | 2012-10-12 07:36 |
| From | luis antunes pena |
| Subject | Re: [Csnd] spectral analysis of oboe multiphonics |
| Attachments | samp2fftall.sh |
|
Hello Peiman,
I know you're not looking for csound solutions, but I thought it could be interesting for you to take a look at a script I wrote some years ago to analyze sounds using the csound phase vocoder opcodes. The script creates three different files: a pvoc-file, a human readable text file with amp/freq pairs and a file to be imported by LISP/PWGL. The script SAMP2FFTALL.SH (in attachment) can analyze all sounds in a folder. You can look at it here: http://luisantunespena.eu/downloads.html Hope it helps. Best, Luís Am 11.10.12 23:42, schrieb peiman khosravi: Of course will do! I didn't know you were at RCM! I have been to a few of Chris' workshops and the possibilities are apparently enormous. I'm really looking forward to receiving all the samples. -- -- http://luisantunespena.eu |
| Date | 2012-10-12 09:29 |
| From | peiman khosravi |
| Subject | Re: [Csnd] spectral analysis of oboe multiphonics |
| Hi Luís, That's great, thanks very much. I'm going to try out your script and study it. Best, Peiman On 12 October 2012 07:36, luis antunes pena <k_o_m_p@yahoo.de> wrote:
|
| Date | 2012-12-18 16:11 |
| From | peiman khosravi |
| Subject | [OT] Re: [Csnd] spectral analysis of oboe multiphonics |
| Dear all, So I ended up making a standalone with max since it allows me to analyse and visualise 8th tones in real time. In case it is of any interest here is a link. The standalone is osx only but I've included the maxmsp patch, which requires the externals to run. Please see the readme.
To run the app: 1- turn DSP on 2- press the 'O' (not zero) key and select a sound file. It works best with sounds that have some pitch content.
3- press the 'S' key to enable the audition of the input sound. 4- press the space bar to play the file. 5- when the tone reaches a steady state press the 'A' key. This will analyse and notate the result.
6- press the 'C' key to hear the synthesis as chord. Or press the 'N' key to arpeggiate the partials. In the 'notes' mode you can also graphically select a note and press 'V' to hear the selected note.
The components can be organised by frequency, strength, or according to psychoacoustic prominence of the pitches. You can also view the list as frequency, amplitude, weight. To install the app you need to move the content of two folders to appropriate directories as instructed in the readme file.
Best, Peiman On 12 October 2012 09:29, peiman khosravi <peimankhosravi@gmail.com> wrote: Hi Luís, |
| Date | 2012-12-18 17:15 |
| From | Tarmo Johannes |
| Subject | Re: [OT] Re: [Csnd] spectral analysis of oboe multiphonics |
|
sounds great!
any chanve for PD version? As I remember you wrote in some point about going over to puredata?
congratulations, anyway! must have been quite a bit of work.
tarmo
On Tuesday 18 December 2012 16:11:35 peiman khosravi wrote: Dear all, So I ended up making a standalone with max since it allows me to analyse and visualise 8th tones in real time. In case it is of any interest here is a link. The standalone is osx only but I've included the maxmsp patch, which requires the externals to run. Please see the readme. https://dl.dropbox.com/u/47945259/analysis.zip To run the app: 1- turn DSP on 2- press the 'O' (not zero) key and select a sound file. It works best with sounds that have some pitch content. 3- press the 'S' key to enable the audition of the input sound. 4- press the space bar to play the file. 5- when the tone reaches a steady state press the 'A' key. This will analyse and notate the result. 6- press the 'C' key to hear the synthesis as chord. Or press the 'N' key to arpeggiate the partials. In the 'notes' mode you can also graphically select a note and press 'V' to hear the selected note. The components can be organised by frequency, strength, or according to psychoacoustic prominence of the pitches. You can also view the list as frequency, amplitude, weight. To install the app you need to move the content of two folders to appropriate directories as instructed in the readme file. Best, Peiman
On 12 October 2012 09:29, peiman khosravi <peimankhosravi@gmail.com> wrote: Hi Luís, On 12 October 2012 07:36, luis antunes pena <k_o_m_p@yahoo.de> wrote: Hello Peiman, Of course will do! I didn't know you were at RCM! I have been to a few of Chris' workshops and the possibilities are apparently enormous. I'm really looking forward to receiving all the samples. On 11 October 2012 22:36, Richard Dobson <richarddobson@blueyonder.co.uk> wrote: Ah, I see, interesting! I'm totally out of the loop these days, but I do have some history with oboists - my father (Michael Dobson) was a professional player, and teacher at the RAM (when not doing all the other stuff he did such as running an orchestra and fly-fishing), and my composition teacher and general mentor even before going to the RCM was Edwin Roxburgh. If you meet him, say "Hi" from me!
Thanks Richard, that's really helpful. Indeed I have already come Send bugs reports to the Sourceforge bug tracker --
--
http://luisantunespena.eu |
| Date | 2012-12-18 17:37 |
| From | peiman khosravi |
| Subject | Re: [OT] Re: [Csnd] spectral analysis of oboe multiphonics |
| The trouble is that the only thing I could find that would allow 8th tone notation (other than pwgl and openmusic) was in max. There are few externals actually but I ended up using this one http://www.bachproject.net/bach/home_page.html. It's really nice and easy to work with. I have ported a couple of my patches to pd and I actually really like it. But there are some GUI stuff missing that I don't think I can live without! I think cabbage will soon bridge the gap though!
Best, Peiman
On 18 December 2012 17:15, Tarmo Johannes <tarmo.johannes@otsakool.edu.ee> wrote:
|
| Date | 2012-12-19 10:08 |
| From | Tarmo Johannes |
| Subject | Re: [OT] Re: [Csnd] spectral analysis of oboe multiphonics |
|
Hello,
once again, just thinking out loud if such solution could be more cross platform (and not telling at all that you should do it or change anything):
there are arrowed accidentals in Lilypond http://lilypond.org/doc/v2.12/Documentation/user/lilypond/The-Feta-font.html#Accidental-glyphs
http://lilypond.org/doc/v2.12/Documentation/topdocs/NEWS.html
Also there are powerful possibilities of microtonal notation: http://x31eq.com/lilypond/
I have very little experience with Lilypond but I will keep it in mind - I think csound and lilypond can be powerful partners, free, beautiful and crossplatform
In your audio analyses did you use just MaxMSP or there is Csound object doing the work? Could you share the csd?
best regards, tarmo
On Tuesday, December 18, 2012 05:37:39 PM peiman khosravi wrote: The trouble is that the only thing I could find that would allow 8th tone notation (other than pwgl and openmusic) was in max. There are few externals actually but I ended up using this one http://www.bachproject.net/bach/home_page.html. It's really nice and easy to work with. I have ported a couple of my patches to pd and I actually really like it. But there are some GUI stuff missing that I don't think I can live without! I think cabbage will soon bridge the gap though! Best, Peiman On 18 December 2012 17:15, Tarmo Johannes <tarmo.johannes@otsakool.edu.ee> wrote: sounds great!
any chanve for PD version? As I remember you wrote in some point about going over to puredata?
congratulations, anyway! must have been quite a bit of work.
tarmo
On Tuesday 18 December 2012 16:11:35 peiman khosravi wrote: Dear all, So I ended up making a standalone with max since it allows me to analyse and visualise 8th tones in real time. In case it is of any interest here is a link. The standalone is osx only but I've included the maxmsp patch, which requires the externals to run. Please see the readme. https://dl.dropbox.com/u/47945259/analysis.zip To run the app: 1- turn DSP on 2- press the 'O' (not zero) key and select a sound file. It works best with sounds that have some pitch content. 3- press the 'S' key to enable the audition of the input sound. 4- press the space bar to play the file. 5- when the tone reaches a steady state press the 'A' key. This will analyse and notate the result. 6- press the 'C' key to hear the synthesis as chord. Or press the 'N' key to arpeggiate the partials. In the 'notes' mode you can also graphically select a note and press 'V' to hear the selected note. The components can be organised by frequency, strength, or according to psychoacoustic prominence of the pitches. You can also view the list as frequency, amplitude, weight. To install the app you need to move the content of two folders to appropriate directories as instructed in the readme file. Best, Peiman
On 12 October 2012 09:29, peiman khosravi <peimankhosravi@gmail.com> wrote: Hi Luís, On 12 October 2012 07:36, luis antunes pena <k_o_m_p@yahoo.de> wrote: Hello Peiman, Of course will do! I didn't know you were at RCM! I have been to a few of Chris' workshops and the possibilities are apparently enormous. I'm really looking forward to receiving all the samples. On 11 October 2012 22:36, Richard Dobson <richarddobson@blueyonder.co.uk> wrote: Ah, I see, interesting! I'm totally out of the loop these days, but I do have some history with oboists - my father (Michael Dobson) was a professional player, and teacher at the RAM (when not doing all the other stuff he did such as running an orchestra and fly-fishing), and my composition teacher and general mentor even before going to the RCM was Edwin Roxburgh. If you meet him, say "Hi" from me!
Thanks Richard, that's really helpful. Indeed I have already come Send bugs reports to the Sourceforge bug tracker --
--
http://luisantunespena.eu |
| Date | 2012-12-19 11:00 |
| From | peiman khosravi |
| Subject | Re: [OT] Re: [Csnd] spectral analysis of oboe multiphonics |
| Hi Tarmo, I did consider lilypond. It doesn't support 8th tones out of the box and although it would be possible to make custom notation, but being a pragmatist and this being a job, I wanted to save time. Also this is standalone and doesn't need any other applications to run. This took me a couple of hours to make, the lilypond options would probably have taken me over a week since I am not familiar with it at all. Also this is real-time, which means you have a notation in less than a second. You can even plug in a microphone and sing into the computer.
I used only maxmsp for this but the analysis part could be done with sound too. This was quicker as I used a maxmsp external that carries out a psychoacoustically informed spectral analysis out of the box.
I was born in the last days of a revolution, and brought up during an eight year war. I became a pragmatist when I realised that my grandfather's and father's revolutions were in vain! Cheers, Peiman
On 19 December 2012 10:08, Tarmo Johannes <tarmo.johannes@otsakool.edu.ee> wrote:
|
| Date | 2012-12-19 11:27 |
| From | Tarmo Johannes |
| Subject | Re: [OT] Re: [Csnd] spectral analysis of oboe multiphonics |
|
Thanks,
sure, all clear.
tarmo
On Wednesday, December 19, 2012 11:00:18 AM peiman khosravi wrote: Hi Tarmo, I did consider lilypond. It doesn't support 8th tones out of the box and although it would be possible to make custom notation, but being a pragmatist and this being a job, I wanted to save time. Also this is standalone and doesn't need any other applications to run. This took me a couple of hours to make, the lilypond options would probably have taken me over a week since I am not familiar with it at all. Also this is real-time, which means you have a notation in less than a second. You can even plug in a microphone and sing into the computer. I used only maxmsp for this but the analysis part could be done with sound too. This was quicker as I used a maxmsp external that carries out a psychoacoustically informed spectral analysis out of the box. I was born in the last days of a revolution, and brought up during an eight year war. I became a pragmatist when I realised that my grandfather's and father's revolutions were in vain! Cheers, Peiman On 19 December 2012 10:08, Tarmo Johannes <tarmo.johannes@otsakool.edu.ee> wrote: Hello,
once again, just thinking out loud if such solution could be more cross platform (and not telling at all that you should do it or change anything):
there are arrowed accidentals in Lilypond http://lilypond.org/doc/v2.12/Documentation/user/lilypond/The-Feta-font.html#Accidental-glyphs
http://lilypond.org/doc/v2.12/Documentation/topdocs/NEWS.html
Also there are powerful possibilities of microtonal notation:
I have very little experience with Lilypond but I will keep it in mind - I think csound and lilypond can be powerful partners, free, beautiful and crossplatform
In your audio analyses did you use just MaxMSP or there is Csound object doing the work? Could you share the csd?
best regards, tarmo
On Tuesday, December 18, 2012 05:37:39 PM peiman khosravi wrote: The trouble is that the only thing I could find that would allow 8th tone notation (other than pwgl and openmusic) was in max. There are few externals actually but I ended up using this one http://www.bachproject.net/bach/home_page.html. It's really nice and easy to work with. I have ported a couple of my patches to pd and I actually really like it. But there are some GUI stuff missing that I don't think I can live without! I think cabbage will soon bridge the gap though! Best, Peiman On 18 December 2012 17:15, Tarmo Johannes <tarmo.johannes@otsakool.edu.ee> wrote: sounds great!
any chanve for PD version? As I remember you wrote in some point about going over to puredata?
congratulations, anyway! must have been quite a bit of work.
tarmo
On Tuesday 18 December 2012 16:11:35 peiman khosravi wrote: Dear all, So I ended up making a standalone with max since it allows me to analyse and visualise 8th tones in real time. In case it is of any interest here is a link. The standalone is osx only but I've included the maxmsp patch, which requires the externals to run. Please see the readme. https://dl.dropbox.com/u/47945259/analysis.zip To run the app: 1- turn DSP on 2- press the 'O' (not zero) key and select a sound file. It works best with sounds that have some pitch content. 3- press the 'S' key to enable the audition of the input sound. 4- press the space bar to play the file. 5- when the tone reaches a steady state press the 'A' key. This will analyse and notate the result. 6- press the 'C' key to hear the synthesis as chord. Or press the 'N' key to arpeggiate the partials. In the 'notes' mode you can also graphically select a note and press 'V' to hear the selected note. The components can be organised by frequency, strength, or according to psychoacoustic prominence of the pitches. You can also view the list as frequency, amplitude, weight. To install the app you need to move the content of two folders to appropriate directories as instructed in the readme file. Best, Peiman
On 12 October 2012 09:29, peiman khosravi <peimankhosravi@gmail.com> wrote: Hi Luís, On 12 October 2012 07:36, luis antunes pena <k_o_m_p@yahoo.de> wrote: Hello Peiman, Of course will do! I didn't know you were at RCM! I have been to a few of Chris' workshops and the possibilities are apparently enormous. I'm really looking forward to receiving all the samples. On 11 October 2012 22:36, Richard Dobson <richarddobson@blueyonder.co.uk> wrote: Ah, I see, interesting! I'm totally out of the loop these days, but I do have some history with oboists - my father (Michael Dobson) was a professional player, and teacher at the RAM (when not doing all the other stuff he did such as running an orchestra and fly-fishing), and my composition teacher and general mentor even before going to the RCM was Edwin Roxburgh. If you meet him, say "Hi" from me!
Thanks Richard, that's really helpful. Indeed I have already come Send bugs reports to the Sourceforge bug tracker --
--
http://luisantunespena.eu |
| Date | 2012-12-19 12:03 |
| From | peiman khosravi |
| Subject | Re: [OT] Re: [Csnd] spectral analysis of oboe multiphonics |
| Having said this. It would be nice to implement the analysis algorithm in Csound. P
On 19 December 2012 11:27, Tarmo Johannes <tarmo.johannes@otsakool.edu.ee> wrote:
|