[Cs-dev] tests and examples for Cs5
Date | 2005-02-22 15:33 |
From | Dave Phillips |
Subject | [Cs-dev] tests and examples for Cs5 |
Greetings: After reading Rick B's latest message re: Cs5 I started working on a collection of files and examples for beginners in Cs5. I've completed material for realtime audio, realtime MIDI, FLTK opcode usage, and connecting to JACK. I intend to provide examples for plugin usage and for adding new opcodes, and I will try to address Rick's other concerns regarding example files. Btw, am I correct to assume that the majority of Csound users have abandoned the use of text mode ? It certainly looks that way from here. I consider this assumption to be a critical point, one that perhaps should be discussed a little more thoroughly by the developers. Also, am I correct to assume that files written for MacCsound are incompatible with CsoundAV and/or Winsound (and vice-versa) ? I have no idea what might be required to reconcile files from those versions of Csound with Cs5 (or any other version of Csound). I cannot supply flags and option switches for those platforms (many ofwhich have been designed with no regard at all towards compatibility with Cs4), and I invite any developers or users on those platforms to contribute the needed flags/switches to the examples set. At this time I'm supplying CsOptions only for Cs5 under Linux. To Dr. B: Rick, how did you try running Cs5 ? Your message was short on details, and I need details. You didn't mention on what platform you tried it, with what files you tested it, or with what options. As far as examples go, I use your very own Trapped In Convert to test realtime audio output, with these options: csound -+rtaudio=alsa -odac2 -d -m0 trappedrt.csd Actually I can dispense with the -+rtaudio switch, it's already defined in my $CSRTAUDIO setting. Best, dp ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-22 15:47 |
From | steven yi |
Subject | Re: [Cs-dev] tests and examples for Cs5 |
Hi Dave, By "text mode", what is it you are asking? Are you saying "run on the commandline on files"? If so, blue uses csound this way. steven Dave Phillips wrote: > Btw, am I correct to assume that the majority of Csound users have > abandoned the use of text mode ? It certainly looks that way from > here. I consider this assumption to be a critical point, one that > perhaps should be discussed a little more thoroughly by the developers. ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-22 16:09 |
From | "Richard Boulanger" |
Subject | Re: [Cs-dev] tests and examples for Cs5 |
Steven and Dave, Behind the scenes, I think that most versions of Csound formulate a command-line. But, I think that apart from the Linux Users of Csound, I think that most beginners/intermediates/composers/students and most of the general users of Csound - i.e. MacCsounders, CsoundAvers, SuperMills Csounders, WInSounders, Cecilia users, and Blue users rarely type the command line themselves to the program. There are a lot of Linux users of the program for sure, and I think that almost all are command-line users, but if you are not using Csound under Linux, then I think you have not been using the command-line. Every now and then, I modify a command line - especially when using an analysis utility. I am not sure that I should be admitting this, but I have not typed a command-line to Csound (except to experiment with some of the Csound5 stuff and the Utils) since about 1990/91. This means that many of us will not only need the test orchestras, but a nice set of command-lines to use with them. OR Better - When Csound5 is released, it would be great to have a nice cross-platform front-end that included a menu of default command lines and help the user set up the directories and preferences from a GUI. This may all already be there in the form of CsoundVST5. As I am just beginning my testing, I am probably way off base here. I look forward to David's collection, instructions, and advice and am happy that we will be hosting this at cSounds.com! -Dr. B. on 2/22/05 10:47 AM, steven yi at stevenyi@csounds.com wrote: > Hi Dave, > > By "text mode", what is it you are asking? Are you saying "run on the > commandline on files"? If so, blue uses csound this way. > > steven > > Dave Phillips wrote: > >> Btw, am I correct to assume that the majority of Csound users have >> abandoned the use of text mode ? It certainly looks that way from >> here. I consider this assumption to be a critical point, one that >> perhaps should be discussed a little more thoroughly by the developers. > > > > > ------------------------------------------------------- > SF email is sponsored by - The IT Product Guide > Read honest & candid reviews on hundreds of IT Products from real users. > Discover which products truly live up to the hype. Start reading now. > http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click > _______________________________________________ > Csound-devel mailing list > Csound-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/csound-devel _______________________________________________________________________ + Dr. Richard Boulanger, Professor + Music Synthesis Department, Berklee College of Music + 1140 Boylston Street - Boston, MA 02215-3693 + Office Phone: (617) 747-2485 Office Fax: (617) 747-2564 + eMail: rboulanger@csounds.com or rboulanger@berklee.edu + WebPage: http://csounds.com/boulanger/ ________________________________________________________________________ + Almost Everything Csound @ http://csounds.com/ + The Csound Catalog with Audio @ http://csounds.com/catalog/ + The Csound Book @ http://csounds.com/book/ + The Csound Magazine @ http://csounds.com/ezine/ + CsoundForums @ http://csounds.com/phpBB2/ ________________________________________________________________________ ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-22 18:00 |
From | "Richard Boulanger" |
Subject | Re: [Cs-dev] tests and examples for Cs5 |
David, Most of the composing is still done with a text editor - on the Mac it is usually bbedit, on the pc I have recommended editpad, - MacCsound and CsoundAV (and I think WinSound too) have built-in text editors. -Rick on 2/22/05 1:11 PM, Dave Phillips at dlphilp@bright.net wrote: > Hi Steven: > > Text-mode as in composing Csound instruments and scores with a text > editor. > > I've had no trouble using blue as a text editor for Olde Skool orc/sco > files or shiny new CSDs. Works great either way. > > Best, > > dp > > > > steven yi wrote: > >> Hi Dave, >> >> By "text mode", what is it you are asking? Are you saying "run on the >> commandline on files"? If so, blue uses csound this way. >> >> steven >> >> Dave Phillips wrote: >> >>> Btw, am I correct to assume that the majority of Csound users have >>> abandoned the use of text mode ? It certainly looks that way from >>> here. I consider this assumption to be a critical point, one that >>> perhaps should be discussed a little more thoroughly by the developers. >> >> >> >> >> >> ------------------------------------------------------- >> SF email is sponsored by - The IT Product Guide >> Read honest & candid reviews on hundreds of IT Products from real users. >> Discover which products truly live up to the hype. Start reading now. >> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click >> _______________________________________________ >> Csound-devel mailing list >> Csound-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/csound-devel >> > > > > > ------------------------------------------------------- > SF email is sponsored by - The IT Product Guide > Read honest & candid reviews on hundreds of IT Products from real users. > Discover which products truly live up to the hype. Start reading now. > http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click > _______________________________________________ > Csound-devel mailing list > Csound-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/csound-devel _______________________________________________________________________ + Dr. Richard Boulanger, Professor + Music Synthesis Department, Berklee College of Music + 1140 Boylston Street - Boston, MA 02215-3693 + Office Phone: (617) 747-2485 Office Fax: (617) 747-2564 + eMail: rboulanger@csounds.com or rboulanger@berklee.edu + WebPage: http://csounds.com/boulanger/ ________________________________________________________________________ + Almost Everything Csound @ http://csounds.com/ + The Csound Catalog with Audio @ http://csounds.com/catalog/ + The Csound Book @ http://csounds.com/book/ + The Csound Magazine @ http://csounds.com/ezine/ + CsoundForums @ http://csounds.com/phpBB2/ ________________________________________________________________________ ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-22 18:11 |
From | Dave Phillips |
Subject | Re: [Cs-dev] tests and examples for Cs5 |
Hi Steven: Text-mode as in composing Csound instruments and scores with a text editor. I've had no trouble using blue as a text editor for Olde Skool orc/sco files or shiny new CSDs. Works great either way. Best, dp steven yi wrote: > Hi Dave, > > By "text mode", what is it you are asking? Are you saying "run on the > commandline on files"? If so, blue uses csound this way. > > steven > > Dave Phillips wrote: > >> Btw, am I correct to assume that the majority of Csound users have >> abandoned the use of text mode ? It certainly looks that way from >> here. I consider this assumption to be a critical point, one that >> perhaps should be discussed a little more thoroughly by the developers. > > > > > > ------------------------------------------------------- > SF email is sponsored by - The IT Product Guide > Read honest & candid reviews on hundreds of IT Products from real users. > Discover which products truly live up to the hype. Start reading now. > http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click > _______________________________________________ > Csound-devel mailing list > Csound-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/csound-devel > ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-22 18:25 |
From | Dave Phillips |
Subject | Re: [Cs-dev] tests and examples for Cs5 |
Greetings, Doctor: Cool, thanks for the heads-up. Best, dp Richard Boulanger wrote: >David, > >Most of the composing is still done with a text editor - on the Mac it is >usually bbedit, on the pc I have recommended editpad, - MacCsound and >CsoundAV (and I think WinSound too) have built-in text editors. > >-Rick > >on 2/22/05 1:11 PM, Dave Phillips at dlphilp@bright.net wrote: > > > >>Hi Steven: >> >>Text-mode as in composing Csound instruments and scores with a text >>editor. >> >>I've had no trouble using blue as a text editor for Olde Skool orc/sco >>files or shiny new CSDs. Works great either way. >> >>Best, >> >>dp >> >> >> >>steven yi wrote: >> >> >> >>>Hi Dave, >>> >>>By "text mode", what is it you are asking? Are you saying "run on the >>>commandline on files"? If so, blue uses csound this way. >>> >>>steven >>> >>>Dave Phillips wrote: >>> >>> >>> >>>>Btw, am I correct to assume that the majority of Csound users have >>>>abandoned the use of text mode ? It certainly looks that way from >>>>here. I consider this assumption to be a critical point, one that >>>>perhaps should be discussed a little more thoroughly by the developers. >>>> >>>> >>> >>> >>> >>>------------------------------------------------------- >>>SF email is sponsored by - The IT Product Guide >>>Read honest & candid reviews on hundreds of IT Products from real users. >>>Discover which products truly live up to the hype. Start reading now. >>>http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click >>>_______________________________________________ >>>Csound-devel mailing list >>>Csound-devel@lists.sourceforge.net >>>https://lists.sourceforge.net/lists/listinfo/csound-devel >>> >>> >>> >> >> >>------------------------------------------------------- >>SF email is sponsored by - The IT Product Guide >>Read honest & candid reviews on hundreds of IT Products from real users. >>Discover which products truly live up to the hype. Start reading now. >>http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click >>_______________________________________________ >>Csound-devel mailing list >>Csound-devel@lists.sourceforge.net >>https://lists.sourceforge.net/lists/listinfo/csound-devel >> >> > >_______________________________________________________________________ > + Dr. Richard Boulanger, Professor > + Music Synthesis Department, Berklee College of Music > + 1140 Boylston Street - Boston, MA 02215-3693 > + Office Phone: (617) 747-2485 Office Fax: (617) 747-2564 > + eMail: rboulanger@csounds.com or rboulanger@berklee.edu > + WebPage: http://csounds.com/boulanger/ >________________________________________________________________________ > + Almost Everything Csound @ http://csounds.com/ > + The Csound Catalog with Audio @ http://csounds.com/catalog/ > + The Csound Book @ http://csounds.com/book/ > + The Csound Magazine @ http://csounds.com/ezine/ > + CsoundForums @ http://csounds.com/phpBB2/ >________________________________________________________________________ > > > >------------------------------------------------------- >SF email is sponsored by - The IT Product Guide >Read honest & candid reviews on hundreds of IT Products from real users. >Discover which products truly live up to the hype. Start reading now. >http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click >_______________________________________________ >Csound-devel mailing list >Csound-devel@lists.sourceforge.net >https://lists.sourceforge.net/lists/listinfo/csound-devel > > > ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-22 19:08 |
From | jpff@codemist.co.uk |
Subject | Re: [Cs-dev] tests and examples for Cs5 |
>>>>> "Dave" == Dave Phillips |
Date | 2005-02-22 19:15 |
From | jpff@codemist.co.uk |
Subject | Re: [Cs-dev] tests and examples for Cs5 |
WinSound does not create a command line -- I now think that was a mistake but it is what i did! ==John ffitch ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-22 19:56 |
From | "Richard Boulanger" |
Subject | Re: [Cs-dev] tests and examples for Cs5 |
Many files run in both CsoundAV and MacCsound. Then entire Csound Catalog does as do ALL the examples and chapters from The Csound Book. And many many more... FLTK/OpenGL orchestras don't work in MacCsound or WinSound (not sure about CsoundVST). MacCsound GUI orchestras don't run in CsoundAV. WinSound, or CsoundVST. Also, a large number non-FLTK/OpenGL orchestras developed in CsoundAV don't run in MacCsound or WinSound because of those 80 or so unique opcodes that could be, but have not yet been, ported to MacCsound or WinSound. I've been working on that - as you know. There is more, but others might have a better report for you on this. As I said, there are many command-line users of Csound - especially in the developer's circle.... So... There is a lot of expertise there to be shared with the community. Still, bringing Csound to the larger electronic music community of composers, producers, sound designers, performers, etc.. Has benefited immensely from simple graphical launcher/editor/front-ends such as John ffitch's WinSound launcher. Rick on 2/22/05 2:08 PM, jpff@codemist.co.uk at jpff@codemist.co.uk wrote: >>>>>> "Dave" == Dave Phillips |
Date | 2005-02-22 19:58 |
From | "Richard Boulanger" |
Subject | Re: [Cs-dev] tests and examples for Cs5 |
Interesting... I thought it did. Greetings John! And best wishes. Rick on 2/22/05 2:15 PM, jpff@codemist.co.uk at jpff@codemist.co.uk wrote: > WinSound does not create a command line -- I now think that was a > mistake but it is what i did! > ==John ffitch > > > ------------------------------------------------------- > SF email is sponsored by - The IT Product Guide > Read honest & candid reviews on hundreds of IT Products from real users. > Discover which products truly live up to the hype. Start reading now. > http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click > _______________________________________________ > Csound-devel mailing list > Csound-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/csound-devel _______________________________________________________________________ + Dr. Richard Boulanger, Professor + Music Synthesis Department, Berklee College of Music + 1140 Boylston Street - Boston, MA 02215-3693 + Office Phone: (617) 747-2485 Office Fax: (617) 747-2564 + eMail: rboulanger@csounds.com or rboulanger@berklee.edu + WebPage: http://csounds.com/boulanger/ ________________________________________________________________________ + Almost Everything Csound @ http://csounds.com/ + The Csound Catalog with Audio @ http://csounds.com/catalog/ + The Csound Book @ http://csounds.com/book/ + The Csound Magazine @ http://csounds.com/ezine/ + CsoundForums @ http://csounds.com/phpBB2/ ________________________________________________________________________ ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-22 20:17 |
From | Iain Duncan |
Subject | Re: [Cs-dev] tests and examples for Cs5 |
> Btw, am I correct to assume that the majority of Csound users have > abandoned the use of text mode ? It certainly looks that way from here. > I consider this assumption to be a critical point, one that perhaps > should be discussed a little more thoroughly by the developers. I'm not sure what you mean by this Dave. I still use csound quite often with nothing going except text, and for gigs I may even use it from a text only console to increase stability and performance. Does that count? Iain ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-22 20:19 |
From | Andres Cabrera |
Subject | Re: [Cs-dev] tests and examples for Cs5 |
Hi Dr. B, > Also, a large number non-FLTK/OpenGL orchestras developed in CsoundAV don't > run in MacCsound or WinSound because of those 80 or so unique opcodes that > could be, but have not yet been, ported to MacCsound or WinSound. I'm working on this (and others have also cleaned op things), and about 75% are already ported, so many CsoundAV orchestras (without openGL) should now work OK with csound5. I'm starting on the documentation for them, becuase I think opcodes are useless unless thay are properly documented. It's going very slow, though.... Cheers, Andrés ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-22 20:24 |
From | Iain Duncan |
Subject | Re: [Cs-dev] tests and examples for Cs5 |
> Text-mode as in composing Csound instruments and scores with a text > editor. > Please don't consider text mode abandoned then. That is all I do now, text editing hand in hand with my real time sequencer midi input. ( I can edit stuff using midi hardware and dump to text files and vice versa ) I realize probably no one has had time to look at the MakeScore utility I just released, but it is a high level tool for text editing of scores making text editing actually pretty quick, and I also have more text editing tools in the pipeline as python extensions for gvim. Between makescore and the python extensions I would say that my writing is already way faster working in all text from gvim than it ever was using Cakewalk etc, and definitely *better* with the extra control and resolution of csound unlimited pfields and higher bit depth. This is of course more true on linux ( and I guess could be on OSX? ) than on windows because of all the great linux text tools. Iain ( utilities at www.xornot.com ) ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-22 20:42 |
From | "Richard Boulanger" |
Subject | Re: [Cs-dev] tests and examples for Cs5 |
Andres, This is fantastic news. Thanks sooooooooooooo much for working on this! Without every realizing it, this will me a ton to future users! Dr. B. on 2/22/05 3:19 PM, Andres Cabrera at andres@geminiflux.com wrote: > Hi Dr. B, >> Also, a large number non-FLTK/OpenGL orchestras developed in CsoundAV don't >> run in MacCsound or WinSound because of those 80 or so unique opcodes that >> could be, but have not yet been, ported to MacCsound or WinSound. > I'm working on this (and others have also cleaned op things), and about > 75% are already ported, so many CsoundAV orchestras (without openGL) > should now work OK with csound5. I'm starting on the documentation for > them, becuase I think opcodes are useless unless thay are properly > documented. > It's going very slow, though.... > > > Cheers, > Andrés > > > ------------------------------------------------------- > SF email is sponsored by - The IT Product Guide > Read honest & candid reviews on hundreds of IT Products from real users. > Discover which products truly live up to the hype. Start reading now. > http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click > _______________________________________________ > Csound-devel mailing list > Csound-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/csound-devel _______________________________________________________________________ + Dr. Richard Boulanger, Professor + Music Synthesis Department, Berklee College of Music + 1140 Boylston Street - Boston, MA 02215-3693 + Office Phone: (617) 747-2485 Office Fax: (617) 747-2564 + eMail: rboulanger@csounds.com or rboulanger@berklee.edu + WebPage: http://csounds.com/boulanger/ ________________________________________________________________________ + Almost Everything Csound @ http://csounds.com/ + The Csound Catalog with Audio @ http://csounds.com/catalog/ + The Csound Book @ http://csounds.com/book/ + The Csound Magazine @ http://csounds.com/ezine/ + CsoundForums @ http://csounds.com/phpBB2/ ________________________________________________________________________ ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_ide95&alloc_id396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-23 06:25 |
From | Anthony Kozar |
Subject | [Cs-dev] Re: [Cs-dev]CsoundAV opcodes docs |
Andres, Thanks again for doing all of this work. I see that John ffitch added the opcodes to the Csound 4 module as well and that is cool. It should save myself and Matt some time in creating Csound 4 plugins. I wanted to make sure though that you do have Gab's original docs, right ?? http://www.csounds.com/maldonado/Manual/index.htm CsoundAV Manual Anthony On 2/22/05 3:19 PM, Andres Cabrera |
Date | 2005-02-23 07:35 |
From | steven yi |
Subject | [Cs-dev] Tablei - profiling |
Hi All, Something I've been trying to figure out for a long time is why the Russel Pinkston DX7 patches are very cpu-intensive. I enabled gprof when compiling CS5 (thanks Istvan for putting that in!) and when profiling, it seems that a great bulk of the processing time is spent in ktabli and tabli, the k-rate and a-rate functions for the tablei opcode. Taking a look at those two functions I didn't see anything out of the ordinary, but I thought I'd ask if anyone could take a look and see if they see anything that might catch their attention as to why this might be slow, or if it simply is as fast as it's going to get. Thanks!! steven ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-23 10:25 |
From | Istvan Varga |
Subject | Re: [Cs-dev] Tablei - profiling |
steven yi wrote: > Something I've been trying to figure out for a long time is why the > Russel Pinkston DX7 patches are very cpu-intensive. I enabled gprof > when compiling CS5 (thanks Istvan for putting that in!) and when > profiling, it seems that a great bulk of the processing time is spent in > ktabli and tabli, the k-rate and a-rate functions for the tablei > opcode. Taking a look at those two functions I didn't see anything out > of the ordinary, but I thought I'd ask if anyone could take a look and > see if they see anything that might catch their attention as to why this > might be slow, or if it simply is as fast as it's going to get. The high CPU usage may be due to the casts from float to int. I have only tested on old P3 CPUs (that's what I have), but such casts take something like 50 CPU cycles to compute. The only thing more expensive is trigonometric functions like sin(). Furthermore, tablei may unnecessarily do the cast more than once; tablexkt is more optimized in this respect, so you may try replacing ar tablei andx, ifn [, ixmode] [, ixoff] [, iwrap] with ar tablexkt xndx, kfn, 1, 2 [, ixmode] [, ixoff] [, iwrap] and see if that reduces CPU usage. ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-23 14:26 |
From | Andres Cabrera |
Subject | Re: [Cs-dev] Re: [Cs-dev]CsoundAV opcodes docs |
Hi Anthony, Yes, I'm basing the documentation on gab's manual. Just one question, I was writing that the opcodes were available only on csound5. From what 4 version would the opcodes be available? Cheers, Andres Anthony Kozar wrote: > Andres, > > Thanks again for doing all of this work. I see that John ffitch added the > opcodes to the Csound 4 module as well and that is cool. It should save > myself and Matt some time in creating Csound 4 plugins. > > I wanted to make sure though that you do have Gab's original docs, right ?? > http://www.csounds.com/maldonado/Manual/index.htm > CsoundAV Manual > > Anthony > > On 2/22/05 3:19 PM, Andres Cabrera |
Date | 2005-02-23 15:21 |
From | Dave Phillips |
Subject | [Cs-dev] Re: [Cs-dev]CsoundAV opcodes docs |
Greetings: > Anthony Kozar wrote: > >> Andres, >> Thanks again for doing all of this work. I see that John ffitch >> added the >> opcodes to the Csound 4 module as well and that is cool. It should save >> myself and Matt some time in creating Csound 4 plugins. >> >> I wanted to make sure though that you do have Gab's original docs, >> right ?? >> http://www.csounds.com/maldonado/Manual/index.htm >> CsoundAV Manual >> >> Anthony >> >> On 2/22/05 3:19 PM, Andres Cabrera |
Date | 2005-02-23 15:31 |
From | steven yi |
Subject | Re: [Cs-dev] Tablei - profiling |
Hi Istvan, Thanks for that information! I will try your recommendation later today and report back any results. steven Istvan Varga wrote: > steven yi wrote: > >> Something I've been trying to figure out for a long time is why the >> Russel Pinkston DX7 patches are very cpu-intensive. I enabled gprof >> when compiling CS5 (thanks Istvan for putting that in!) and when >> profiling, it seems that a great bulk of the processing time is spent >> in ktabli and tabli, the k-rate and a-rate functions for the tablei >> opcode. Taking a look at those two functions I didn't see anything >> out of the ordinary, but I thought I'd ask if anyone could take a >> look and see if they see anything that might catch their attention as >> to why this might be slow, or if it simply is as fast as it's going >> to get. > > > The high CPU usage may be due to the casts from float to int. > I have only tested on old P3 CPUs (that's what I have), but such > casts take something like 50 CPU cycles to compute. The only thing > more expensive is trigonometric functions like sin(). > Furthermore, tablei may unnecessarily do the cast more than once; > tablexkt is more optimized in this respect, so you may try replacing > ar tablei andx, ifn [, ixmode] [, ixoff] [, iwrap] > with > ar tablexkt xndx, kfn, 1, 2 [, ixmode] [, ixoff] [, iwrap] > and see if that reduces CPU usage. ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-23 15:33 |
From | Andres Cabrera |
Subject | Re: [Cs-dev] Re: [Cs-dev]CsoundAV opcodes docs |
Hi Dave, The ported csoundav opcodes which have been ported to csound5 (I don't know about the state for csound 4) are built with the scons system, and they build properly (and automatically- you don't need to do anything). At least, I haven't heard of any problems. The opcodes are untested, so some might not work properly yet. During documentation I will do this testing. Cheers, Andres Dave Phillips wrote: > Greetings: > >> Anthony Kozar wrote: >> >>> Andres, >>> Thanks again for doing all of this work. I see that John ffitch >>> added the >>> opcodes to the Csound 4 module as well and that is cool. It should save >>> myself and Matt some time in creating Csound 4 plugins. >>> >>> I wanted to make sure though that you do have Gab's original docs, >>> right ?? >>> http://www.csounds.com/maldonado/Manual/index.htm >>> CsoundAV Manual >>> >>> Anthony >>> >>> On 2/22/05 3:19 PM, Andres Cabrera |
Date | 2005-02-23 20:04 |
From | Anthony Kozar |
Subject | [Cs-dev] Re: CsoundAV opcodes docs |
On 2/23/05 10:21 AM, Dave Phillips |
Date | 2005-02-23 20:34 |
From | Dave Phillips |
Subject | Re: [Cs-dev] Re: CsoundAV opcodes docs |
Anthony Kozar wrote: >>Further, I tried compiling some of the opcodes according to the >>instructions in the (admittedly outdated) csound_manual.pdf and ran into >>failures. >> >> > >That manual is for Csound 5. The opcode sources in the Plugin Pack are for >Csound 4. They would require some modification to compile for Cs5. Victor >already has Cs5 sources for his opcodes on his web site. > Ah, okay, that's what I really need. I'm chasing the wrong tiger, sorry. >>I suggest to any & all plugin developers that you not release >>files without a standard make procedure, or at least don't call them 1.0 >>releases. >> >> > >Makefiles aren't so "standard" from my perspective. And I do not know the >procedures for compiling opcode plugins on other platforms since I have not >tried it yet. I made the .tgz release of the Plugin Pack available only as >a convenience for non-Mac people since the Mac disk image is unreadable on >other platforms. Creating makefiles for those platforms will have to be >someone else's responsibility. Sorry to be unhelpful. > No, no, I think I came on a little too strongly. You do state that the user is expected to write his own makefiles, I was just surprised by that. >I am going to be managing a "Plugin Repository" on Csounds.com. If you come >up with Linux makefiles for these plugins I would appreciate if you could >share them with me. Then I will make sure that they are distributed with >the sources. > > Alas, I'm focused on Cs5 right now, but I'll try building them again with a Cs4 tree when I have a chance. Best, dp ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-23 21:23 |
From | jpff@codemist.co.uk |
Subject | Re: [Cs-dev] Tablei - profiling |
Looking at tablei I see amongst other stuff a test on p->wrap for every sample, which could be a test on a local, or even better tested outside the ksmps loop rather than every time. This is Robin Wittle code who was keen that his code not be changed, and so I ignored it for a long time. ==John ffitch ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-23 22:13 |
From | Istvan Varga |
Subject | Re: [Cs-dev] Tablei - profiling |
jpff@codemist.co.uk wrote: > Looking at tablei I see amongst other stuff a test on p->wrap for > every sample, which could be a test on a local, or even better tested > outside the ksmps loop rather than every time. > This is Robin Wittle code who was keen that his code not be changed, > and so I ignored it for a long time. I was able to improve the performance of tablei by replacing lines like this indx = (long) floor((double)ndx); with indx = (ndx >= FL(0.0) ? (long) ndx : (long) ((double) ndx - 0.999999)); in OOps/ugens2.c. Some testing would be needed to find out how this change affects the speed of the opcode on various machines, but I got an improvement of a factor of more than 1.5. Moving code out of the loop should help too, I will test that now. ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-23 22:40 |
From | Dave Phillips |
Subject | Re: [Cs-dev] Tablei - profiling |
Greetings: The pvanal of the linus.wav yielded a normal PV data file, but the CSD test below doesn't work. Csound just hangs after here: orch now loaded audio buffered in 1024 sample-frame blocks writing 2048-byte blks of shorts to /home/dlphilp/soundfiles/linus-rebuilt.wav WAV SECTION 1: Nothing happens after that. The original file is only 5 secs long. Pvoc writes an empty header (44 bytes) and that's all. I have to kill -9 to close the session. Here's the test file, maybe I'm doing something wrong here: |
Date | 2005-02-23 22:51 |
From | Istvan Varga |
Subject | Re: [Cs-dev] Tablei - profiling |
Now with restructuring the code I got further speedup, but not quite as significant. The results for a simple test are: original tablei: ~2.2 seconds replaced floor(): ~1.4 seconds + restructured code: ~1.3 seconds Here is the faster version of tabli() (not tested extensively): int tabli(ENVIRON *csound, TABLE *p) { FUNC *ftp; long indx, mask, length; int nsmps = ksmps; MYFLT *rslt, *pxndx, *tab; MYFLT fract, v1, v2, ndx, xbmul, offset; ftp = p->ftp; if (ftp==NULL) { return perferror(Str("tablei: not initialised")); } rslt = p->rslt; length = ftp->flen; pxndx = p->xndx; xbmul = (MYFLT)p->xbmul; offset = p->offset; mask = ftp->lenmask; tab = ftp->ftable; if (!p->wrap) { /* limit mode */ do { ndx = (*pxndx++ * xbmul) + offset; if (ndx <= FL(0.0)) { *rslt++ = *tab; continue; } indx = (long) ndx; if (ndx >= length) { *rslt++ = *(tab + length); continue; } fract = ndx - indx; /* interpolate */ v1 = *(tab + indx); v2 = *(tab + indx + 1); *rslt++ = v1 + (v2 - v1)*fract; } while (--nsmps); } else { do { /* wrap mode */ ndx = (*pxndx++ * xbmul) + offset; indx = (ndx >= FL(0.0) ? (long) ndx : (long) ((double) ndx - 0.999999)); fract = ndx - indx; indx &= mask; /* interpolate */ v1 = *(tab + indx); v2 = *(tab + indx + 1); *rslt++ = v1 + (v2 - v1)*fract; } while(--nsmps); } return OK; } ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-23 22:55 |
From | Dave Phillips |
Subject | PVOC problem, was Re: [Cs-dev] Tablei - profiling |
Sorry, I meant to retitle the subject. Dave Phillips wrote: > Greetings: > > The pvanal of the linus.wav yielded a normal PV data file, but the > CSD test below doesn't work. Csound just hangs after here: > > orch now loaded > audio buffered in 1024 sample-frame blocks > writing 2048-byte blks of shorts to > /home/dlphilp/soundfiles/linus-rebuilt.wav > WAV > SECTION 1: > > Nothing happens after that. The original file is only 5 secs long. > Pvoc writes an empty header (44 bytes) and that's all. I have to kill > -9 to close the session. > > Here's the test file, maybe I'm doing something wrong here: > > |
Date | 2005-02-23 23:05 |
From | "Richard Boulanger" |
Subject | Re: [Cs-dev] Tablei - profiling |
Let's make it more efficient! -Rick on 2/23/05 4:23 PM, jpff@codemist.co.uk at jpff@codemist.co.uk wrote: > Looking at tablei I see amongst other stuff a test on p->wrap for > every sample, which could be a test on a local, or even better tested > outside the ksmps loop rather than every time. > This is Robin Wittle code who was keen that his code not be changed, > and so I ignored it for a long time. > > ==John ffitch > > > ------------------------------------------------------- > SF email is sponsored by - The IT Product Guide > Read honest & candid reviews on hundreds of IT Products from real users. > Discover which products truly live up to the hype. Start reading now. > http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click > _______________________________________________ > Csound-devel mailing list > Csound-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/csound-devel _______________________________________________________________________ + Dr. Richard Boulanger, Professor + Music Synthesis Department, Berklee College of Music + 1140 Boylston Street - Boston, MA 02215-3693 + Office Phone: (617) 747-2485 Office Fax: (617) 747-2564 + eMail: rboulanger@csounds.com or rboulanger@berklee.edu + WebPage: http://csounds.com/boulanger/ ________________________________________________________________________ + Almost Everything Csound @ http://csounds.com/ + The Csound Catalog with Audio @ http://csounds.com/catalog/ + The Csound Book @ http://csounds.com/book/ + The Csound Magazine @ http://csounds.com/ezine/ + CsoundForums @ http://csounds.com/phpBB2/ ________________________________________________________________________ ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-23 23:55 |
From | steven yi |
Subject | Re: [Cs-dev] Tablei - profiling |
Thanks Istvan for taking a look at this code! I'm looking forward to going home tonight to see what kind of difference this will make for the DX7 emulation patches in realtime for polyphony. steven Istvan Varga wrote: > Now with restructuring the code I got further speedup, but not quite > as significant. The results for a simple test are: > original tablei: ~2.2 seconds > replaced floor(): ~1.4 seconds > + restructured code: ~1.3 seconds > > Here is the faster version of tabli() (not tested extensively): > > int tabli(ENVIRON *csound, TABLE *p) > { > FUNC *ftp; > long indx, mask, length; > int nsmps = ksmps; > MYFLT *rslt, *pxndx, *tab; > MYFLT fract, v1, v2, ndx, xbmul, offset; > > ftp = p->ftp; > if (ftp==NULL) { > return perferror(Str("tablei: not initialised")); > } > rslt = p->rslt; > length = ftp->flen; > pxndx = p->xndx; > xbmul = (MYFLT)p->xbmul; > offset = p->offset; > mask = ftp->lenmask; > tab = ftp->ftable; > if (!p->wrap) { > /* limit mode */ > do { > ndx = (*pxndx++ * xbmul) + offset; > if (ndx <= FL(0.0)) { > *rslt++ = *tab; > continue; > } > indx = (long) ndx; > if (ndx >= length) { > *rslt++ = *(tab + length); > continue; > } > fract = ndx - indx; > /* interpolate */ > v1 = *(tab + indx); > v2 = *(tab + indx + 1); > *rslt++ = v1 + (v2 - v1)*fract; > } while (--nsmps); > } > else { > do { > /* wrap mode */ > ndx = (*pxndx++ * xbmul) + offset; > indx = (ndx >= FL(0.0) ? (long) ndx : (long) ((double) ndx - > 0.999999)); > fract = ndx - indx; > indx &= mask; > /* interpolate */ > v1 = *(tab + indx); > v2 = *(tab + indx + 1); > *rslt++ = v1 + (v2 - v1)*fract; > } while(--nsmps); > } > return OK; > } > > > ------------------------------------------------------- > SF email is sponsored by - The IT Product Guide > Read honest & candid reviews on hundreds of IT Products from real users. > Discover which products truly live up to the hype. Start reading now. > http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click > _______________________________________________ > Csound-devel mailing list > Csound-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/csound-devel > > ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2005-02-24 21:27 |
From | Anthony Kozar |
Subject | Re: [Cs-dev] Re: Plugin makefiles |
I found that there is an example *nix makefile for a Cs4 plugin in Steven's testOpcode package. I think that I could adapt this for my Plugin Pack -- maybe even produce a simple shell script to produce makefiles for future plugins. (I am not saying that this will happen soon -- just a possibility :) Anthony >> I am going to be managing a "Plugin Repository" on Csounds.com. If you come >> up with Linux makefiles for these plugins I would appreciate if you could >> share them with me. Then I will make sure that they are distributed with >> the sources. >> >> > Alas, I'm focused on Cs5 right now, but I'll try building them again > with a Cs4 tree when I have a chance. ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |