Wednesday 26 November 2003 10.56 skrev Daniel James: I think the live 512 has the same drivers as the audigy, hence it's quite probable it has the same problems as the Audigy with dropped notes. (It's probably only a problem if you have very fast compositions.) /Robert

Mark Constable wrote I'm not quite sure that these are about SF2. The first one uses Timidity, the second one is library, which gives support of Gigasampler files and is used in LinuxSampler. -- Alexandre Prokoudine | "When you set yourself on fire and aim ALT Linux Documentation Team | for the sky, you hope to leave behind E-mail: avp(a)altlinux.ru | some sparks of heat and light" JabberID: avp(a)altlinux.org | Neil Peart

On Wed, 26 Nov 2003 09:57:12 -0500 Dave Phillips <dlphilp(a)bright.net> wrote: There's a new version of this nice Soundfont? -- music: http://www.soundclick.com/bands/9/florianschmidt.htm

On Wed, 26 Nov 2003 10:25:08 -0500 Dave Phillips <dlphilp(a)bright.net> wrote: Ah, i see. Thanks for the link. Flo -- music: http://www.soundclick.com/bands/9/florianschmidt.htm

Wednesday 26 November 2003 22.17 skrev Florian Schmidt: I have run sfark under 'some' version of Wine... I don't think it works with the current wine though :/ ... maybe with some tweaks... sfpack on the other hand works flawlessly under wine, maybe there is a version packed with that somewhere? /Robert

On Wed, 26 Nov 2003 16:54:49 -0500 Dave Phillips <dlphilp(a)bright.net> wrote: Yes, and in the zipfile is what? ;) Yes, an sfArk file... Flo -- music: http://www.soundclick.com/bands/9/florianschmidt.htm

On Thu, 27 Nov 2003 09:38:57 +0000 Steve Harris <S.W.Harris(a)ecs.soton.ac.uk> wrote: Hmm, ok so it is the same version that i have. still with vastly different volumes for some instruments, etc.. [the link you gave above is where i got my copy from]. the "20" in the filename "FluidR3_20GM.sf2" made me believe there was another version, like 2.0 of this soundfont... I guess not.. FLo -- music: http://www.soundclick.com/bands/9/florianschmidt.htm

On Wed, 26 Nov 2003, Dave Phillips wrote: I would use TiMidity++. The problem is: In contrast to FluidSynth TiMidity loads the instruments not until they are really needed. That is: If the clarinet keeps silent for - say - 20 measures the clarinet waves are not loaded. In measure 20 TiMidity starts loading the clarinet. This requires processor time and leads to a short gap. (Also with my 2.4 GHz Intel) Does anybody know how to change this behaviour? And: Does anybody know how to recorord the FluidSynth aoutput into a WAV file ? -- J.Anders, Chemnitz, GERMANY (ja(a)informatik.tu-chemnitz.de)

Greetings: For all purposes these days TiMidity == TiMidity++. The original app is no longer developed by its original coder, nor has it been developed for years. There's no practical reason to go calling it TiMidity++. But if it makes you happy... ;) Best regards, == dp Joerg Anders wrote:

On Fri, 28 Nov 2003, Julien Claassen wrote: "-j" ??? This has no influence! -- J.Anders, Chemnitz, GERMANY (ja(a)informatik.tu-chemnitz.de)

On Tue, 2 Dec 2003, Takashi Iwai wrote: it could load all WAVes into the RAM. FluidSynth apparently does. I tried to use TiMidity++-2.11.3, but: gcc -DHAVE_CONFIG_H -I. -I. -I.. -I.. -I../timidity -I../libarc -I../utils -I/usr/local/include -I/usr/X11R6/include -DAU_ALSA -DPKGDATADIR=\"/usr/local/share/timidity\" -DPKGLIBDIR=\"/usr/local/lib/timidity\" -DDEFAULT_PATH=\"/usr/local/share/timidity\" -DTIMIDITY_OUTPUT_ID=\"s\" -O2 -c `test -f alsaseq_c.c || echo './'`alsaseq_c.c alsaseq_c.c: In function `ctl_pass_playing_list': alsaseq_c.c:267: storage size of `pinfo' isn't known alsaseq_c.c:278: `SND_SEQ_OPEN' undeclared (first use in this function) alsaseq_c.c:278: (Each undeclared identifier is reported only once alsaseq_c.c:278: for each function it appears in.) alsaseq_c.c:278: too few arguments to function `snd_seq_open' alsaseq_c.c:287: storage size of `pinfo' isn't known alsaseq_c.c:292: `SND_SEQ_GROUP_DEVICE' undeclared (first use in this function) alsaseq_c.c: In function `do_sequencer': alsaseq_c.c:472: `SND_SEQ_EVENT_PORT_USED' undeclared (first use in this function) alsaseq_c.c:486: `SND_SEQ_EVENT_PORT_UNUSED' undeclared (first use in this function) -- J.Anders, Chemnitz, GERMANY (ja(a)informatik.tu-chemnitz.de)

On Tue, 2 Dec 2003 09:18 pm, Takashi Iwai wrote: Is this documented somewhere ? Generic question: can individual midi tracks determine their own level of effects on a track by track basis (or channel by channel perhaps). If yes then what midi events can trigger this ? --markc

On Tue, 2 Dec 2003, Takashi Iwai wrote: You are right! I could compile the pre-release TiMidity++-2.12.0-pre1 and indeed, the sound is better. But please have a look at: http://rnvs.informatik.tu-chemnitz.de/TIMIDITY/timidity.html Here you find a MIDI file, the hardware synthesis and the TiMidity synthesis. Ok this is a sensitive example. But make your own attempts: In my opinion there is no similarity between hardware and soft synthesis. The timidity command was: timidity -Ow1l -s 44100 -EFdelay=0 -EFchorus=0 -EFreverb=0 -o <file>.wav <file>.mid Again: I would use the emu10k1 hardware synthesizer if it wouldn't create lot of gaps: http://rnvs.informatik.tu-chemnitz.de/DROP/gap.html -- J.Anders, Chemnitz, GERMANY (ja(a)informatik.tu-chemnitz.de)

On Wed, 26 Nov 2003, Dave Phillips wrote: j.anders(a)informatik.tu-chemnitz.de wrote: At the end of this email I've included the Timidity Mailing List address. I am pretty sure Timidity will pre-load instruments but I don't remember how. If you're referring to converting a midi file to WAV, Timidity does this. In fact (I beleive) that is one of it's purposes when it was developed. I have done it many times in the past. In fact, You can convert a whole directory of .mid to .wav. If You configure everything right, Timidiy sounds _great_ and works well with MuSE. And that's good because MuSE allows sycronizing of (live) audio and midi. I'm going to download the newest version of MuSE. I'm not sure if it has added (recording) of live audio yet. But even with My old version of MuSE, I've been able to sycronize (live) audio with midi. There's probly going to be some people on this list that will not like to hear this, but with the (above) combination, I was able to avoid all complexities of Jack. I eliminated some major steps to syncronizing midi and audio. For Me it's just a more stripped down way to get things done. I've been away on a long vacation and this is My first day back, so I don't recall how to do any of the above (by memory). Here's the Timidity Mailing List... ================================================ (I don't remember if You have to "sign up" to post, but here's the welcome message I received when I joined...) ================================================ Welcome to the Timidity and Eawpats mailing list. Feel free to ask questions or reply to questions. Please adderess all mail to the list as follows. timidity(a)stardate.ca ================================================ Keep in mind that Timidity is also a Windows program. So there are (very minor) differences when a Windows Person is answering Your question. Rocco

On Mon, 1 Dec 2003 07:13 pm, Joerg Anders wrote: Is it possible that Timidity does not sound quite as clean as the hardware version because you have it's effects on by default ? To compare Timidity you really need to -EFreverb=0 -EFdelay=0 -EFchorus=0 I've been tempted to hack Timidity to have sane defaults and to accept stdin.. and to not use less for paging the help. Thanks to your comparison mp3s I would not bother using FluidSynth for anything. These and your other examples indicate this obnoxious attack and again, it sounds nearly a semi-tone too high. BTW how did you record the output from the SBlive to wav ? --markc

On Mon, 1 Dec 2003 08:57 pm, Joerg Anders wrote: Apologies.. maybe the extra volume in the Timidity sample delueded me into hearing things that were not there. I do notice a very subtle initial attack on the timidity example strings which is not there in the hardware example... but one has to "listen very hard" to notice it. FluidSynth is not even in the race. For me, except for the above slight defect of timdidty I can hardly pick the difference and considering the hardware could drop fast notes (as per your gap examples) then it seems Timidity is still quite usable. I am quite amazed how "right" it is compared to authentic Creative hardware. Sorry to be boring... you mean without plugging a lead from your analog out into your analog in and then simply recording as above ? I've got an old SBlive but sfxload won't work (for me) on a 2.6 kernel... so I can't test this right now. --markc

On Wed, 3 Dec 2003 09:50 am, LinuxMedia wrote: (about capturing SBlive hardware synth output directly to a wav file) So you've developed your own soundfont ? Is it something you can share with others ? --markc

On Thu, 4 Dec 2003 01:17 pm, LinuxMedia wrote: So would this still apply, if say Timidity was in alsaseq mode with MusE playing back thru it, and ALSO using a keyboard at the same time aconnected to Timidity, that BOTH MusE putput and keyboard input could be recorded to a wav file via arecord and/or ecasound ? --markc

lunedì, 01 dicembre 2003 alle 20:47:03, Mark Constable ha scritto: I would also add "-s 44100" :) Ciao -- Emiliano Grilli Linux user #209089 http://www.emillo.net

On Mon, 1 Dec 2003 10:13:25 +0100 (MET), Joerg Anders wrote / a écrit: Hi all just my own opinion, but I think I should, for what it's worth: I use timidity for main midi synth (with muse). But when I want to play with a soundfont, there's no question: fluidsynth is better. The Roland_Piano and the Rhodes_mkII soundfonts definetely are better with fluidsynth. Especially the Rhodes samples, because timidity doesn't handle at all the builtin vibrato effect. Fluidsynth does. Just one thing, I don't know how to change the buffers size in fluid synth. And as another poster said, the samples on the web page above have not been equally done (Timidity effects, volume, sample rate 44100 and fluidsynth pitch because sample rate 48000). cheers all. GuyCLO~

On Wed, 26 Nov 2003, Robert Jonsson wrote: On Wed, 26 Nov 2003, Alexandre Prokoudine wrote: I would use TiMidity++ But in contrast to FluidSynth this program loads every instrument (program) immediately before it is needed. This leads to gaps if a new instrument begins to play. FluidSynth (seems to) load(s) the whole soundfont into the ram AT program start. -- J.Anders, Chemnitz, GERMANY (ja(a)informatik.tu-chemnitz.de)

On Wed, 26 Nov 2003, Hartmut Z Noack wrote: ... because everybody recommended Audigy and SB Live. I assume they spoke about the PCM output of this card. My main point is the MIDI synthesizer. ... I have enough. How is it loaded ? With "sfxload" ? .... it is supported. But the new computers have no ISA slots. ... "much better" is not so interesting to me (if you think about noise and latency or so... ) All I want is: The SF2 instruments should be correctly sythesised (unlike FluidSynth string creation) and the wavetable synthesizer should not omit some notes (unlike ALSA emu10k1 driver). Both offers AWE-64 with 24 MB RAM extention. Ok! I'll make a try. -- J.Anders, Chemnitz, GERMANY (ja(a)informatik.tu-chemnitz.de)