Re: Interpolation schemes and SB Live!

Date	1998-10-22 15:40
From	Hans Mikelson
Subject	Re: Interpolation schemes and SB Live!
	Hi, >The problem is that the first order filters that are implemented in the waveguide opcodes >don't use fractional delay, so the pitch of high notes is not precise. There is a paper on fractional waveguide filters at: http://www.hut.fi/Yksikot/Akustiikka/publications/vesa_phd.html Bye, Hans Mikelson Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa08334; 22 Oct 98 15:54 BST Received: from pat.bath.ac.uk by stork.maths.Bath.AC.UK id aa12907; 22 Oct 98 15:54 BST Received: (qmail 21989 invoked from network); 22 Oct 1998 14:54:21 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by pat.bath.ac.uk with SMTP; 22 Oct 1998 14:54:21 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (PAA27156); Thu, 22 Oct 1998 15:43:05 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Thu, 22 Oct 1998 15:42:56 +0100 Received: from root@lix.intercom.es [194.179.21.2] by hermes via ESMTP (PAA00016); Thu, 22 Oct 1998 15:42:50 +0100 (BST) Received: from intercom.es (iv1-233.intercom.es [195.76.154.233]) by lix.intercom.es (8.7.3/8.6.12) with ESMTP id QAA25474; Thu, 22 Oct 1998 16:41:29 +0100 Message-ID: <362F4396.CA213143@intercom.es> Date: Thu, 22 Oct 1998 16:39:18 +0200 From: Josep M Comajuncosas X-Mailer: Mozilla 4.05 [en] (Win95; I) MIME-Version: 1.0 To: Gabriel Maldonado CC: csound@maths.ex.ac.uk Subject: Re (II): Interpolation schemes and SB Live! References: <362E5DB3.43AB74F6@intercom.es> <362F1E6D.746B372E@agora.stm.it> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable X-MIME-Autoconverted: from 8bit to quoted-printable by exeter.ac.uk id PAA00016 Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk Gabriel Maldonado wrote: > Where is it possble to get the algorithm to implement 8 point interpola= tion? If I could > see an example of it I will implement it in Csound. Hi, as I see you=B4re interested in the subject, here it is the general formula for a Nth order Lagrange interpolator, it works as a FIR structure and has an inherent lowpass behaviour, but implements very accurately the fractional delay, further references in T.Tolonen & Vesa V=E4lim=E4ky works. Josep M Comajuncosas This is a PS to text translation of their Thesis: h(n) =3D mult (from k=3D0, k=3D!n to N) [(D-k)/(n-k)] for n=3D 0,1,2,...N= (3.63) where D is the fractional delay and N the order of the FIR filter. Computation of Coefficients for a Lagrange Interpolator The first=ADorder Lagrange interpolator corresponds to the well=ADknown linear interpola=AD tion. The coefficients are obtained from Eq. (3.63) with N =3D 1 as h(0) =3D 1- D , h(1) =3D D (3.72) Note that in the first=ADorder case D =3D d since the integer part of D i= s zero. The formulas for the filter coefficients of the Lagrange interpolator with N =3D 2 are h(0) =3D (1/2)(D -1)(D - 2) h(1) =3D -D(D - 2) h(2) =3D (1/2)D(D -1) (3.73) and with N =3D 3, h(0) =3D -(1/6)(D -1)(D - 2)(D - 3) h(1) =3D (1/2)D(D - 2)(D - 3) h(2) =3D -(1/2)D(D -1)(D - 3) h(3) =3D (1/6)D(D -1)(D - 2) (3.74) In general, the coefficients h(n) of the Lagrange interpolator are determined by an Nth=ADorder polynomial in D. This implies that N additions and N multiplications per coefficient are required to evaluate its value for a given D. Altogether, it takes N(N +1) =3D N 2 + N additions and N 2 + N multiplications to update all the N coeffi=AD cients. An efficient procedure to update the coefficients of a Lagrange interpolator was introduced in V=E4lim=E4ki (1992, Appendix A): 1) evaluate the differences D -- 1, D -- 2, ..., D -- N; 2) evaluate the products D(D -- 1), (D -- 2)(D -- 3), ..., (D -- N + 1)(D -- N); 3) evaluate the coefficients h(0), h(1), ..., h(N) by multiplying the results of Steps 1 and 2 and constant coefficients. Using this technique, the computational burden of evaluating the coefficients of an Nth=AD order interpolator consists of N additions (subtractions) and less than N 2 + N multipli=AD cations. For example, computing the values of the four coefficients of a third=ADorder Lagrange interpolator using this method requires 3 additions and 10 multiplications. This is a remarkable saving when compared to 12 additions and 12 multiplications that would be required were the straightforward polynomial evaluation used. The most efficient practical solution to the updating of interpolating coefficients is to use table lookup. In software implementations this is usually the fastest way, and thus most recommendable, provided that the required memory space is available. > > Another question related to waveguide opcodes. Does deltapi (and dela= yw > > btw) accept fractional sample delay lengths (well, fractional 1/kr > > units) ? > > Yes it does. > > The problem is that the first order filters that are implemented in the= waveguide opcodes > don't use fractional delay, so the pitch of high notes is not precise. Hmm a serious handicap. Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa08495; 22 Oct 98 16:56 BST Received: from pat.bath.ac.uk by stork.maths.Bath.AC.UK id aa18968; 22 Oct 98 16:56 BST Received: (qmail 29610 invoked from network); 22 Oct 1998 15:46:34 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by pat.bath.ac.uk with SMTP; 22 Oct 1998 15:46:34 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (QAA03863); Thu, 22 Oct 1998 16:40:08 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Thu, 22 Oct 1998 16:39:55 +0100 Received: from howl.werewolf.net [206.103.224.20] by hermes via ESMTP (QAA04760); Thu, 22 Oct 1998 16:39:48 +0100 (BST) Received: from default (dial150.werewolf.net [206.103.225.160]) by howl.werewolf.net (8.9.0/8.9.0) with SMTP id KAA21279 for ; Thu, 22 Oct 1998 10:35:31 -0500 (CDT) From: Hans Mikelson To: Csound Subject: Create Halloween Sound Effects with Spatializer Orc/Sco Date: Thu, 22 Oct 1998 10:43:02 -0500 Message-ID: <01bdfdd2$a7433140$a0e167ce@default> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 4.71.1712.3 X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3 Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk Hidyho Csounders, If you are like me then you have been busy decorating the outside of your house with cob-webs, skeletons, dummies, transforming the front yard into a graveyard, installing motion sensor triggered special effects, smoke machines, black lights and strobe lights. No Halloween house would be complete without some spooky sound effects. First go pick some spooky sounds from your local internet haunted sound depot: http://www.geocities.com/Heartland/Meadows/6158/story/hhsound.html This is a good start but mono sounds are not going to cut it for Halloween. You want owls hooting in the tree above your head, wolves howling all around you, cackling witches whizzing past on broomsticks, organs playing in a cathedral and beasts trying to bite your head off. Now you can animate those dull lifeless Halloween sounds with my handy-dandy spatial audio generator to make truly bone-chilling, spine-tingling effects. Merely insert your soundin.? file into the spatializer, specify x, y coordinate paths and let your sounds come alive: ; ORCHESTRA ;--------------------------------------------------------------------------- ------- ; Spatial Audio ; Coded by Hans Mikelson October 1998 ; This spatial audio system included the following features: ; ; 1. Moving sound source with table supplied moving X and Y coordinates ; 2. Volume decreases as 1/r^2 ; 3. Filtering of sound due to air absorption ; 4. Doppler effect due to moving sound ; 5. Simplified HRTF for head shadow filtering ; 6. Inter ear time delay ; 7. Speaker cross-talk cancellation based on both inter ear time delay & head shadowing ;--------------------------------------------------------------------------- ------- sr=44100 kr=22050 ksmps=2 nchnls=2 zakinit 30,30 ;--------------------------------------------------------------------------- ------- ; Disk Input Mono ;--------------------------------------------------------------------------- ------- instr 4 iamp = p4 irate = p5 isndin = p6 ioutch1 = p7 ain1 diskin isndin, irate zawm ain1, ioutch1 endin ;--------------------------------------------------------------------------- ------- ; Disk Input Stereo ;--------------------------------------------------------------------------- ------- instr 5 iamp = p4 irate = p5 isndin = p6 ioutch1 = p7 ioutch2 = p8 ain1, ain2 diskin isndin, irate zawm ain1, ioutch1 zawm ain2, ioutch2 outs ain1iamp, ain2iamp endin ;--------------------------------------------------------------------------- ------- ; Spatial Audio ; Coded by Hans Mikelson October 1998 ;--------------------------------------------------------------------------- ------- instr 11 idur = p3 ; Duration iamp = p4 ; Amplitude iexl = p5 ; Fixed left ear location X coordinate iexr = iexl+.23 ; Fixed right ear location X coordinate iey = p6 ; Fixed ear location Y coordinate (facing +Y axis) isxtab = p7 ; Moving source table for X coordinate in m isytab = p8 ; Moving source table for Y coordinate in m iinch = p9 ; Audio input channel ihrtabl = p10 ; Head shadowing filter for left ear ihrtabr = p11 ; Head shadowing filter for right ear imax = 500 ; Maximum delay time in ms ipi = 3.14159265 ; Pi ics = .333 ; Approx. speed of sound in m/ms kamp linseg 0, .002, iamp, idur-.004, iamp, .002, 0 ; Declick Enevelope ksx oscil 1, 1/idur, isxtab ; Moving Source X ksy oscil 1, 1/idur, isytab ; Moving Source Y ksmexl = ksx-iexl ; Delta X left ksmexr = ksx-iexr ; Delta X right ksmey = ksy-iey ; Delta Y kdistl = sqrt(ksmexlksmexl+ksmeyksmey) ; Distance from left ear to source kdistr = sqrt(ksmexrksmexr+ksmeyksmey) ; Distance from right ear to source kdistlnz= (kdistl==0 ? 1 : kdistl) ; Avoid divide by zero problem kdistrnz= (kdistr==0 ? 1 : kdistr) ; Avoid divide by zero problem kangl = cosinv(ksmexl/kdistlnz) ; Get the magnitude of the angle (0 to Pi) kangr = cosinv(ksmexr/kdistrnz) ; Get the magnitude of the angle (0 to Pi) ksmeynz = (ksmey==0 ? 1 : ksmey) ; Make sure it is not zero ksign = ksmeynz/abs(ksmeynz)/2/ipi ; Get the sign kanglel = kanglksign ; Get the true angle -Pi to Pi kangler = kangrksign ; Get the true angle -Pi to Pi kdd4l = kdistl/4 ; Factor for distance and air absorption kdd4r = kdistr/4 ; Factor for distance and air absorption ain zar iinch ; Input audio adoppl vdelay ain/(1+kdd4lkdd4l), kdistl/ics, imax ; Volume decreases as distance^2. adoppr vdelay ain/(1+kdd4rkdd4r), kdistr/ics, imax ; Volume decreases as distance^2. ; arrival time is delayed by speed of sound, doppler effect afltail butterlp adoppl, 20000/(1+kdd4l) ; Air absorption of distant sound afltair butterlp adoppr, 20000/(1+kdd4r) ; Air absorption of distant sound khrtfl tablei kanglel, ihrtabl, 1, .5 ; Get the "poor man's" HRTF for left ear khrtfr tablei kangler, ihrtabr, 1, .5 ; Get the "poor man's" HRTF for right ear ahrtfl butterlp afltail, khrtfl ; Apply the filter for head shadowing left ear ahrtfr butterlp afltair, khrtfr ; Apply the filter for head shadowing right ear asctcl butterlp ahrtfl, 8000 ; Simplified head shadowing for speaker cross-talk asctcr butterlp ahrtfr, 8000 ; Simplified head shadowing for speaker cross-talk asctcdl delay asctcl, .0003 ; Speaker cross-talk cancellation delay left asctcdr delay asctcr, .0003 ; Speaker cross-talk cancellation delay right outs (ahrtfl-asctcdr)kamp, (ahrtfr-asctcdl)kamp ; Cancel speaker cross-talk and ouput zacl iinch, iinch ; Clean up the zak channel endin ; SCORE ; Sta Dur Amp Pitch SoundIn OutCh i4 0.0 2.61 0.0 1 1 1 i4 3.0 1.52 0.0 1 3 2 i4 4.0 4.38 0.0 1 2 1 i4 8.0 1.86 0.0 1 4 1 i4 6.0 5.44 0.0 1 5 2 f4 0 1024 -7 16000 128 12000 384 3000 128 8000 256 20000 256 16000 ; HRTF Left f5 0 1024 -7 8000 128 3000 384 12000 128 16000 256 20000 256 8000 ; HRTF Right f10 0 1024 -7 12 512 25 512 15 ; Moving Source X f11 0 1024 -7 7 256 7 256 8 256 9 256 15 ; Moving Source Y f12 0 1024 -7 18 512 23 512 15 ; Moving Source X f13 0 1024 -7 18 256 9 256 8 256 9 256 15 ; Moving Source Y ; 2D Echos ; Sta Dur Amp EarX EarY SourceX SourceY InCh HRTFL HRTFR i11 0.0 10.0 .5 20 10 10 11 1 4 5 i11 3.0 10.0 .5 20 10 10 12 2 4 5 Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa08777; 22 Oct 98 19:55 BST Received: from pat.bath.ac.uk by stork.maths.Bath.AC.UK id aa05690; 22 Oct 98 19:55 BST Received: (qmail 20514 invoked from network); 22 Oct 1998 18:55:51 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by pat.bath.ac.uk with SMTP; 22 Oct 1998 18:55:51 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (TAA04008); Thu, 22 Oct 1998 19:50:54 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Thu, 22 Oct 1998 19:50:41 +0100 Received: from kmt.hku.nl [192.87.219.66] by hermes via ESMTP (TAA24009); Thu, 22 Oct 1998 19:50:40 +0100 (BST) Received: from student-kmt.hku.nl (student-kmt.hku.nl [192.87.219.67]) by kmt.hku.nl (980427.SGI.8.8.8/8.8.8) with ESMTP id UAA00564 for ; Thu, 22 Oct 1998 20:50:09 +0200 (MDT) Received: from pc-6457.hku.nl (pc-6457.hku.nl [192.87.216.62]) by student-kmt.hku.nl (980427.SGI.8.8.8/8.8.8) with SMTP id UAA04270 for ; Thu, 22 Oct 1998 20:50:09 +0200 (MDT) Message-Id: <199810221850.UAA04270@student-kmt.hku.nl> From: Dave Perry To: csound@maths.ex.ac.uk Date: Thu, 22 Oct 1998 20:51:26 -0700 MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Subject: ANNOUNCE : Visual Orchestra no longer available from Demon Priority: normal X-mailer: Pegasus Mail for Win32 (v3.01b) Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk This is to announce that the new download site for Visual Orchestra, the Windows Csound GUI is: The old download site will be offline in the next few days. thanks, Dave Perry email : perry@student-kmt.hku.nl Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa08783; 22 Oct 98 19:56 BST Received: from pat.bath.ac.uk by stork.maths.Bath.AC.UK id aa05747; 22 Oct 98 19:56 BST Received: (qmail 20565 invoked from network); 22 Oct 1998 18:56:33 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by pat.bath.ac.uk with SMTP; 22 Oct 1998 18:56:33 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (TAA08852); Thu, 22 Oct 1998 19:53:18 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Thu, 22 Oct 1998 19:53:06 +0100 Received: from mailhub.dartmouth.edu [129.170.16.6] by hermes via ESMTP (TAA23371); Thu, 22 Oct 1998 19:53:04 +0100 (BST) Received: from donner.Dartmouth.EDU (donner.dartmouth.edu [129.170.208.3]) by mailhub.dartmouth.edu (8.8.8+DND/8.8.8) with SMTP id OAA23159 for ; Thu, 22 Oct 1998 14:51:41 -0400 (EDT) Message-id: <14319411@donner.Dartmouth.EDU> Date: 22 Oct 98 14:51:41 EDT From: "Matt J. Ingalls" Subject: Csound 3.49.1 PPC To: "Sarah E. Meyers" , "Peter M. Traub" , "Owen R. Grace" , "Michael C. Frengel" , "Douglas I. Repetto" , "Lily Anne N. Perez" , Taehong Park , "Charles M. Dodge" , "Douglas I. Repetto" , Larry Polansky , Kui Dong , csound@maths.ex.ac.uk MIME-Version: 1.0 Content-Type: text/plain Content-Disposition: inline Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk hi everyone, i just put a new version of mac PPC csound on the ftp: ftp://mills.edu/ccm/csound.ppc/Csound3.49.1PPC.sit.hqx a bit of the README below... because of upgrade to CWP4, some things have changed most notably the MIDI code - which hasnt been tested - so if you could report any bugs that would be a great help. -matt ======================================= oct22.98 Csound for Power Macintosh ======================================= This Version: fFitch code base 3.49 Perf "engine" 3.49.1 Csound "front end" 1.1.1 Available From: ftp://mills.edu/ccm/csound.ppc ======================================= changes from 3.485 (front end 1.0.0) ======================================= Perf: - Upgraded source to 3.49 - ReEnabled User-Settable screen buffer - Fixed Saved Command Line bug losing last char - Fixed AIFF-C bug not compatible with Soundhack (thanks to tom erbe for help) - Internal MIDI changes with merge may have improved MIDI or made implementation worse(?) Front End: - Fixed bug in score generator autogen - Added option to not display output window when saving as a listing file - Reflected newly added ability to use a "Goggins-Style" one file (.csd) instead of orc/sco combination - Due to compiler upgrade, now an optimiztion conflict in score generator - as a result front end is not entirely optimized - noticeable effect is unknown. ======================================= Please email bugs to csound-dev@mills.edu Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa09251; 22 Oct 98 20:59 BST Received: from pat.bath.ac.uk by stork.maths.Bath.AC.UK id aa11320; 22 Oct 98 20:59 BST Received: (qmail 25431 invoked from network); 22 Oct 1998 19:59:47 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by pat.bath.ac.uk with SMTP; 22 Oct 1998 19:59:47 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (UAA13609); Thu, 22 Oct 1998 20:56:38 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Thu, 22 Oct 1998 20:56:25 +0100 Received: from exim@wallace.maths.bath.ac.uk [138.38.100.104] by hermes via ESMTP (UAA08486); Thu, 22 Oct 1998 20:56:23 +0100 (BST) Received: from [138.38.99.25] (helo=maths.Bath.AC.UK ident=mmdf) by wallace.maths.bath.ac.uk with smtp (Exim 1.92 #2) for csound@maths.ex.ac.uk id 0zWQqM-0000jl-00; Thu, 22 Oct 1998 20:56:22 +0100 Date: Thu, 22 Oct 98 20:56:21 BST From: jpff@maths.bath.ac.uk Subject: Csound versi0on 3.49 To: csound@maths.ex.ac.uk Message-Id: Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk Message written at 21 Oct 1998 10:01:43 +0100 I placed the sources, binaries and release notes on ftp://ftp.maths.bath.ac.uk/pub/dream/newest/ late on Tuesday as I knew that I would be out of contact for a few days () and as the system seems reasonably robust, and there are real bugs in at least the SGI version for 3.48x, I thought I would slip them out and see who notices. The release notes are more or less as in the draft, except I completed some opcode descriptions. The system also has a scheme for accepting a single unified file of orchestra and score, but that is still rather incomplete -- it works but the code for setting options and such is not written yet. The design follows that of Michaeel Gogins, which he uses for AXCsound. I suggest that you use this facility with care. For 3.50 (yes there will be one) I intend completing this as well as some other things currently in draft here. I will place the SGI binary on the server as soon as I can (Friday?) and also the SUN one. I expect MAC and Linux systems to be ready by then also. I hope you all enjoy this release, and make good music with it ==John ffitch * looking for a job; anyone have a vacancy for an aging expensive academic/hacker? Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa09287; 22 Oct 98 21:02 BST Received: from pat.bath.ac.uk by stork.maths.Bath.AC.UK id aa11509; 22 Oct 98 21:02 BST Received: (qmail 25637 invoked from network); 22 Oct 1998 20:02:30 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by pat.bath.ac.uk with SMTP; 22 Oct 1998 20:02:30 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (UAA06774); Thu, 22 Oct 1998 20:59:21 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Thu, 22 Oct 1998 20:59:08 +0100 Received: from exim@wallace.maths.bath.ac.uk [138.38.100.104] by hermes via ESMTP (UAA23618); Thu, 22 Oct 1998 20:59:07 +0100 (BST) Received: from [138.38.99.25] (helo=maths.Bath.AC.UK ident=mmdf) by wallace.maths.bath.ac.uk with smtp (Exim 1.92 #2) for csound@maths.ex.ac.uk id 0zWQt0-0000kK-00; Thu, 22 Oct 1998 20:59:06 +0100 Date: Thu, 22 Oct 98 20:59:06 BST From: jpff@maths.bath.ac.uk Subject: SGI operating system query To: csound@maths.ex.ac.uk Message-Id: Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk Message written at 22 Oct 1998 15:42:18 +0100 The computer support staff in Bath have proposed that all out SGIs be upgraded to Irix6.3 in order to solve the Y2K problem. This would mean that I would be unable to generate an Irix5.3 binary (although a 4.0.5F is likely to continue!). My question is would the lack of a 5.3 system inconveninece anyone? I may still me able to influence changes here, but it is not worth doing it unless there is a real need. It may be that everyone with SGI also has compilers in which case I do not have to be too concerned with providing binaries anyway. Or it could be a real pain for someone. Let me know please. ==John ffitch Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa09460; 22 Oct 98 22:26 BST Received: from mercury.bath.ac.uk by stork.maths.Bath.AC.UK id aa18863; 22 Oct 98 22:26 BST Received: (qmail 16810 invoked from network); 22 Oct 1998 21:26:35 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by mercury.bath.ac.uk with SMTP; 22 Oct 1998 21:26:35 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (WAA20171); Thu, 22 Oct 1998 22:23:11 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Thu, 22 Oct 1998 22:22:59 +0100 Received: from gabriel.utsa.edu [129.115.51.41] by hermes via ESMTP (WAA17621); Thu, 22 Oct 1998 22:22:57 +0100 (BST) Received: by gabriel.utsa.edu with Internet Mail Service (5.5.1960.3) id <4944W5RH>; Thu, 22 Oct 1998 16:26:11 -0500 Message-ID: <0204B9309979D11194AA00805FFEDE3F01A806A3@gabriel.utsa.edu> From: Scott Lipscomb To: 'CSound Discussion List' Subject: psychoacoustic demonstrations Date: Thu, 22 Oct 1998 16:26:10 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.1960.3) Content-Type: text/plain Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk Greetings, folks! I have been searching all of the various CSound-related sights hoping to find an answer in the Discussion List FAQ or on the variety of Orchestra/Score sites, but have come up empty handed ... that brings me to you, collectively. I teach a variety of classes here at UTSA, many of which address the topic of psychoacoustics. I am attempting to find a repository of sound files (public domain) that provide aural demonstrations of some of the more common auditory illusions (e.g., pitch circularity as demonstrated in the Shepard Scale [Shepard, 1964], Risset's gliding tone that illustrates the same concept, etc.) and other related psychoacoustic phenomena. Is there a collection of orchs & scores anywhere in existence that would contain some of these materials? either in sound file format or in orch/score combinations? Thanks in advance for your response to a novice CSound user! Sincerely, Dr. Scott D. Lipscomb Assistant Director, Assistant Professor, and Undergraduate Advisor of Record UTSA Division of Music 6900 N. Loop 1604 West San Antonio, TX 78249 (210) 458-4354 (210) 458-4381 FAX lipscomb@utsa.edu http://music.utsa.edu/~lipscomb/ Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa09845; 23 Oct 98 1:53 BST Received: from mercury.bath.ac.uk by stork.maths.Bath.AC.UK id aa07797; 23 Oct 98 1:53 BST Received: (qmail 27545 invoked from network); 23 Oct 1998 00:53:22 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by mercury.bath.ac.uk with SMTP; 23 Oct 1998 00:53:22 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (BAA08169); Fri, 23 Oct 1998 01:50:52 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Fri, 23 Oct 1998 01:50:39 +0100 Received: from camel8.mindspring.com [207.69.200.58] by hermes via ESMTP (BAA26511); Fri, 23 Oct 1998 01:50:37 +0100 (BST) Received: from axe (user-38ld095.dialup.mindspring.com [209.86.129.37]) by camel8.mindspring.com (8.8.5/8.8.5) with SMTP id UAA11847; Thu, 22 Oct 1998 20:50:30 -0400 (EDT) Message-ID: <000c01bdfe20$b701d0c0$258156d1@axe> From: Michael Gogins To: Gabriel Maldonado Cc: Csound list , "Boulanger, Richard" Subject: Re: Interpolation schemes and SB Live! Date: Thu, 22 Oct 1998 21:01:47 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 4.72.3110.5 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3 Content-Transfer-Encoding: quoted-printable X-MIME-Autoconverted: from 8bit to quoted-printable by exeter.ac.uk id BAA26511 Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk Gabriel, there was a presentation on this at the ICMC that you can find i= n your Proceedings: Roger Dannenberg, "Interpolation Error in Waveform Table Lookup", page 24= 0. You could email Dannenberg to see if he has suggestions on the best way t= o do this in Csound. By the way, I had a chat with Jim Wright from IBM about using TCP/IP or U= DP as a protocol for MIDI. He was against it because delivery latency cannot= be guaranteed with straight IP. He suggested that FireWire be used, and said= he was looking into it. Coincidentally, I ran into the following Web page when doing my semi-regu= lar Web search for Csound related stuff: http://www.pavo.com/index.htm. This is a very interesting site because this outfit appears to be working with Analog Devices on an Extended Csound related project, and with Yamah= a on a FireWire protocol for both multi-channel real-time digital audio, an= d MIDI information. I think we should talk to the Pavo people and see what the specification = of their FireWire protocol is, and whether any manufacturers (Yamaha?) are going to make controllers for it. If so, it might be a good choice to put into Csound. -----Original Message----- From: Gabriel Maldonado To: Josep M Comajuncosas Cc: csound@maths.ex.ac.uk Date: Thursday, October 22, 1998 8:17 AM Subject: Re: Interpolation schemes and SB Live! Josep M Comajuncosas wrote: > > Hi all, > I recently saw the promotional prospectus of a low-cost PC sound card > (called MaxiSound or so) and I was really disapointed with its specs. > compared to Csound. Specifically, it is said to use a 3rd order > polynomial scheme for wavetable interpolation, when Csound only accepts > linear (1st order) interpolation. > As you may agree this cannot be accepted ;-), I=B4d suggest at least 3 = new > > opcodes, called maybe > oscilp > tablep > deltapp > with an additional parameter to choose the polynomial degree used for > interpolated readout. > I=B4m not very sure about the improvement in the s/n ratio and in the > frequency response in relation to the degree of the Lagrange > interpolator. Maybe some of you have any clues. But it would be nice to > implement such feature. The problem is: does it improve the sound quality? How much does it cost = in processing speed? My new SoundBlaster Live has an internal wavetable synthesizer who does 8 point interpolation in realtime. Actually the sound quality is excellent. I don= 't know if it depends of the interpolation algorithm or of the hardware analog converte= r. Where is it possble to get the algorithm to implement 8 point interpolati= on? If I could see an example of it I will implement it in Csound. > Another question related to waveguide opcodes. Does deltapi (and delayw > btw) accept fractional sample delay lengths (well, fractional 1/kr > units) ? Yes it does. > If not, is it expected to implement it soon? Otherwise it could > be difficult to tune a delay line at a precise high frequency, esp. at > low k-rates. > This is specially noticeable when developing waveguide models. The problem is that the first order filters that are implemented in the waveguide opcodes don't use fractional delay, so the pitch of high notes is not precise. -- Gabriel Maldonado http://www.agora.stm.it/G.Maldonado/home2.htm Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa09872; 23 Oct 98 2:09 BST Received: from pat.bath.ac.uk by stork.maths.Bath.AC.UK id aa09277; 23 Oct 98 2:09 BST Received: (qmail 12672 invoked from network); 23 Oct 1998 01:09:43 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by pat.bath.ac.uk with SMTP; 23 Oct 1998 01:09:43 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (CAA07442); Fri, 23 Oct 1998 02:06:53 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Fri, 23 Oct 1998 02:06:40 +0100 Received: from root@westnet.com [206.24.6.2] by hermes via ESMTP (CAA12974); Fri, 23 Oct 1998 02:06:39 +0100 (BST) Received: from goodguy (port10.ts1.westnet.com [206.24.7.231]) by westnet.com (8.9.1/8.9.1) with SMTP id VAA08946; Thu, 22 Oct 1998 21:06:24 -0400 (EDT) Message-ID: <362FD7C5.525ABF77@westnet.com> Date: Fri, 23 Oct 1998 01:11:33 +0000 From: Larry Troxler X-Mailer: Mozilla 3.01 (X11; I; Linux 2.0.31 i586) MIME-Version: 1.0 To: Michael Gogins CC: Gabriel Maldonado , Csound list , "Boulanger, Richard" Subject: Re: Interpolation schemes and SB Live! References: <000c01bdfe20$b701d0c0$258156d1@axe> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk A couple of observations/questions: - When I was using Csound, I would simply use a large enough table size, so that the error created by using only linear (or none at all) interpolation, was less then the resolution of the output file type (in my case, 16 bits). On today's machines, using 64K for a wavetable isn't a big deal. I suppose the tradeoff is that there the RAM cache will be overwhelmed on certain processors, hence making interpolation preferable. - At what point, does Sinc interpolation (which, is the theoretically correct approach) become more cost effective (in terms of fidelity per computation time) then polynomial interpolation? Without really analyzing the tradeoffs, my initial response to a proposal to a proposal to use say, eight-point poly interpolation, would be that surely a Sinc interpolation with lesser points would give better results. But perhaps not, especially given that the implementation of a Sinc interpolation might need to involve a second table lookup to efficiently compute the sinc function! Well, in summary, (a) Why not just use larger table sizes where the interpolation method isn't an issue, and (b) what are the tradeoff points with polynomial vs. Sinc interpolations? Larry -- Larry Troxler -- lt@westnet.com -- Patterson, NY USA -- Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa09900; 23 Oct 98 2:17 BST Received: from mercury.bath.ac.uk by stork.maths.Bath.AC.UK id aa09893; 23 Oct 98 2:17 BST Received: (qmail 28381 invoked from network); 23 Oct 1998 01:17:43 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by mercury.bath.ac.uk with SMTP; 23 Oct 1998 01:17:43 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (CAA25939); Fri, 23 Oct 1998 02:15:17 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Fri, 23 Oct 1998 02:15:05 +0100 Received: from nmol.com [206.162.11.2] by hermes via SMTP (CAA06849); Fri, 23 Oct 1998 02:15:03 +0100 (BST) X-ROUTED: Thu, 22 Oct 1998 19:33:58 -0600 X-TCP-IDENTITY: Mikeb Received: from nmol.com [206.162.11.157] by nmol.com with smtp id BDCBBMAA ; Thu, 22 Oct 1998 19:33:28 -0600 Message-ID: <362FDA91.B5891C4F@nmol.com> Date: Thu, 22 Oct 1998 19:23:30 -0600 From: Mike Berry X-Mailer: Mozilla 4.5 (Macintosh; U; PPC) X-Accept-Language: en MIME-Version: 1.0 To: Michael Gogins , Csound list Subject: Re: Interpolation schemes and SB Live! References: <000c01bdfe20$b701d0c0$258156d1@axe> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk I have recently implemented UDP control of GrainWave using the Open Sound Control (OSC) kit from CNMAT. We are still getting a few bugs out, but the system is able to provide quite solid performance, even over a 33.6 modem line. OSC has a timing system in it, so everything is timetagged so that it can be played on schedule. We are presently simply encapsulating MIDI messages inside OSC packets, but the facility is there to expand to other protocols. OSC is platform independant - all you need to do is add a UDP receiver and sender. You can get info and the SDK at: http://cnmat.CNMAT.Berkeley.EDU/OSC -- Mike Berry mikeb@nmol.com http://www.nmol.com/users/mikeb Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa10003; 23 Oct 98 3:16 BST Received: from mercury.bath.ac.uk by stork.maths.Bath.AC.UK id aa16433; 23 Oct 98 3:16 BST Received: (qmail 297 invoked from network); 23 Oct 1998 02:16:05 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by mercury.bath.ac.uk with SMTP; 23 Oct 1998 02:16:05 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (DAA04002); Fri, 23 Oct 1998 03:13:32 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Fri, 23 Oct 1998 03:13:19 +0100 Received: from smtp2.jps.net [209.63.224.235] by hermes via ESMTP (DAA26542); Fri, 23 Oct 1998 03:13:14 +0100 (BST) Received: from hml (209-142-35-188.stk.jps.net [209.142.35.188]) by smtp2.jps.net (8.9.0/8.8.5) with SMTP id TAA22904 for ; Thu, 22 Oct 1998 19:15:31 -0700 (PDT) Message-ID: <000b01bdfb2c$423a6d20$bc238ed1@hml> From: nunativs To: Csound List Subject: MADSR Date: Sun, 18 Oct 1998 23:46:43 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 4.72.2106.4 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4 Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk Greetings Upon testing the new midi envelope I've found that it dosen't work. I'm assuming that there is a note off sensor for the release stage, but it dosen't seem to trigger. Also the Winsound interface dosen't seem to remember if one sets the "T" flag or not. Good Luck, Ken Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa10054; 23 Oct 98 3:44 BST Received: from pat.bath.ac.uk by stork.maths.Bath.AC.UK id aa18803; 23 Oct 98 3:44 BST Received: (qmail 15933 invoked from network); 23 Oct 1998 02:44:04 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by pat.bath.ac.uk with SMTP; 23 Oct 1998 02:44:04 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (DAA01573); Fri, 23 Oct 1998 03:40:52 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Fri, 23 Oct 1998 03:40:39 +0100 Received: from root@proxy3.ba.best.com [206.184.139.14] by hermes via ESMTP (DAA13130); Fri, 23 Oct 1998 03:40:37 +0100 (BST) Received: from charlieb.com (IDENT:baker@baker.vip.best.com [206.86.232.121]) by proxy3.ba.best.com (8.9.0/8.9.0/best.out) with ESMTP id TAA17065; Thu, 22 Oct 1998 19:38:44 -0700 (PDT) Message-ID: <362FC0F3.8288B51B@charlieb.com> Date: Thu, 22 Oct 1998 19:34:11 -0400 From: Charles Baker X-Mailer: Mozilla 4.05 [en] (X11; I; Linux 2.0.34 i686) MIME-Version: 1.0 To: Larry Troxler , csound Subject: Re: Interpolation schemes and SB Live! References: <000c01bdfe20$b701d0c0$258156d1@axe> <362FD7C5.525ABF77@westnet.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk Larry Troxler wrote: ..some helpful, insightful stuff, then questions himself, saying >perhaps not, especially given that the implementation of a Sinc interpolation >might need to involve a second table lookup to efficiently compute the sinc function! Well, absolutely, you need the second lookup... but the second lookup still is much faster even though you will still have to window the Sinc interpolation function, anyway... as to polynomial interpolation.. as I understand it (and you evidently also opine) the polynomial calculation expense grows quite high for orders that produce useful results, and if you can afford the memory 'cost' of large table size & the sinc & window tables for filtering, you can easily beat polynomial interpolation. In fact , I don't even think the sinc 'filter' has to be that many points, although again to get enough 'resolution' in the appropriatly sampled sinc func, you should use a large enough sinc function or use some sort or (usually linear ;) ) interpolation in reading the sinc function! And still, I understand this algorithm clocks out faster than equivalently clean polynomial interpolation algorithms!! charlieb Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa10758; 23 Oct 98 11:13 BST Received: from pat.bath.ac.uk by stork.maths.Bath.AC.UK id aa00305; 23 Oct 98 11:13 BST Received: (qmail 15197 invoked from network); 23 Oct 1998 10:13:49 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by pat.bath.ac.uk with SMTP; 23 Oct 1998 10:13:49 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (LAA24246); Fri, 23 Oct 1998 11:08:11 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Fri, 23 Oct 1998 11:07:50 +0100 Received: from agora.stm.it [195.62.32.1] by hermes via ESMTP (LAA20384); Fri, 23 Oct 1998 11:07:46 +0100 (BST) Received: from agora.stm.it (ppp05-14.dial-access.stm.it [195.62.38.14]) by agora.stm.it (8.8.8/8.8.5) with ESMTP id MAA24199; Fri, 23 Oct 1998 12:07:16 +0200 (ITADST) Message-ID: <36304D95.A6E7689B@agora.stm.it> Date: Fri, 23 Oct 1998 11:34:13 +0200 From: Gabriel Maldonado X-Mailer: Mozilla 4.06 [en] (Win95; I) MIME-Version: 1.0 To: Larry Troxler CC: Csound list Subject: Re: Interpolation schemes and SB Live! References: <000c01bdfe20$b701d0c0$258156d1@axe> <362FD7C5.525ABF77@westnet.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk I think you are right about the fact that a big table can make multi-point interpolation useless. Actually, in the case of a sine wave I find that even a 1024 point table, readed by the oscili opcode, sounds very very good at any frequency. The problem rises when using sound samples (for example with loscil). In this case the table is filled by a set of samples already existent, and the user cannot modify them. Also in this case the computer memory is not so big to expand the wavefile length and interpolate the samples by means of good a wave editor. BTW What is the difference beetween Sinc interpolation and polynomial interpolation (excuse my math ignorance). Gabriel Larry Troxler wrote: > > A couple of observations/questions: > > - When I was using Csound, I would simply use a large enough table > size, so that the error created by using only linear (or none at all) > interpolation, was less then the resolution of the output file type (in > my case, 16 bits). On today's machines, using 64K for a wavetable isn't > a big deal. I suppose the tradeoff is that there the RAM cache will be > overwhelmed on certain processors, hence making interpolation > preferable. > > - At what point, does Sinc interpolation (which, is the theoretically > correct approach) become more cost effective (in terms of fidelity per > computation time) then polynomial interpolation? Without really > analyzing the tradeoffs, my initial response to a proposal to a proposal > to use say, eight-point poly interpolation, would be that surely a Sinc > interpolation with lesser points would give better results. But perhaps > not, especially given that the implementation of a Sinc interpolation > might need to involve a second table lookup to efficiently compute the > sinc function! > > Well, in summary, (a) Why not just use larger table sizes where the > interpolation method isn't an issue, and (b) what are the tradeoff > points with polynomial vs. Sinc interpolations? > > Larry > > -- Larry Troxler -- lt@westnet.com -- Patterson, NY USA -- -- Gabriel Maldonado http://www.agora.stm.it/G.Maldonado/home2.htm Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa11103; 23 Oct 98 13:03 BST Received: from pat.bath.ac.uk by stork.maths.Bath.AC.UK id aa11574; 23 Oct 98 13:03 BST Received: (qmail 316 invoked from network); 23 Oct 1998 12:03:07 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by pat.bath.ac.uk with SMTP; 23 Oct 1998 12:03:07 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (MAA26424); Fri, 23 Oct 1998 12:54:37 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Fri, 23 Oct 1998 12:54:23 +0100 Received: from exim@wallace.maths.bath.ac.uk [138.38.100.104] by hermes via ESMTP (MAA21449); Fri, 23 Oct 1998 12:54:22 +0100 (BST) Received: from [138.38.99.25] (helo=maths.Bath.AC.UK ident=mmdf) by wallace.maths.bath.ac.uk with smtp (Exim 1.92 #2) for csound@maths.ex.ac.uk id 0zWfnN-0001US-00; Fri, 23 Oct 1998 12:54:17 +0100 Date: Fri, 23 Oct 98 12:54:15 BST From: jpff@maths.bath.ac.uk Subject: Re: Interpolation schemes To: csound@maths.ex.ac.uk Message-Id: Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk Message written at 22 Oct 1998 22:25:37 +0100 --- Copy of mail to gelida@intercom.es --- Improved intepolation was on my list after ICMC; have not got round to it yet. ==John ffitch Received: from stork.maths.bath.ac.uk by omphalos.maths.Bath.AC.UK id aa11115; 23 Oct 98 13:04 BST Received: from pat.bath.ac.uk by stork.maths.Bath.AC.UK id aa11749; 23 Oct 98 13:04 BST Received: (qmail 499 invoked from network); 23 Oct 1998 12:04:43 -0000 Received: from hermes.ex.ac.uk (HELO exeter.ac.uk) (144.173.6.14) by pat.bath.ac.uk with SMTP; 23 Oct 1998 12:04:43 -0000 Received: from noether [144.173.8.10] by hermes via SMTP (MAA19412); Fri, 23 Oct 1998 12:54:58 +0100 (BST) Received: from exeter.ac.uk by maths.ex.ac.uk; Fri, 23 Oct 1998 12:54:47 +0100 Received: from exim@wallace.maths.bath.ac.uk [138.38.100.104] by hermes via ESMTP (MAA04572); Fri, 23 Oct 1998 12:54:46 +0100 (BST) Received: from [138.38.99.25] (helo=maths.Bath.AC.UK ident=mmdf) by wallace.maths.bath.ac.uk with smtp (Exim 1.92 #2) for csound@maths.ex.ac.uk id 0zWfnt-0001Uf-00; Fri, 23 Oct 1998 12:54:49 +0100 Date: Fri, 23 Oct 98 12:54:49 BST From: jpff@maths.bath.ac.uk Subject: Reverb question To: csound@maths.ex.ac.uk Message-Id: Sender: owner-csound-outgoing@maths.ex.ac.uk Precedence: bulk Message written at 23 Oct 1998 11:32:21 +0100 This is a question expecting the answer no (as I remember from learning latin). Does anyone have the impulse-response of the Cathedral of the Dormition in Smolensk? I ask as it seems to have a very sweet sound judging by a recording of the USSR Ministry of Culture Chamber Choir I have, and I would like to use it. ==John ff