Carotinho wrote: The developement version of swami (http://swami.sourceforge.net/) has introduced sume basic support for the gigasample format. I've never used it for this, but I've succesfully produced some custom sf2 samplebanks with it and used them with fluidsynth (via qsynth, http://qsynth.sourceforge.net). If you couple one of the tools above with Pure Data, you can reach an umbelieveble level of flexibility. For instance, I'm making a pd patch that is basically a virtual drummer, with different ranges of random variation, both for timing and for velocity, routed to linuxsampler with a very good quality drum bank I've found at http://www.naturalstudio.co.uk/ns_kit7freedownload.html Another patch I'm working on can make a good piano soundbank sound even more realistic, implementing some of the intrinsicnesses of a real piano, like the sympathetic vibration of a string when the key is pressed and I play the same key at an octave above, or the fact that if I've just hit a key and I hit it again when the hammer hasn't had the time to go back in the rest position, and so on. In bocca al lupo. c. -- www.cesaremarilungo.com

On 3/10/06, Cesare Marilungo <cesare(a)poeticstudios.com> wrote: MHO: There's no point in wasting your time. There is a vast commercial market of fine samples out there to be had, and I don't see the open source community jumping on the task of creating their own, at least at the rate that commercial developers do. Unfortunately, not all hardware samplers import all formats, so going with the hardware solution will probably be just as frustrating as the software one. It's a catch-22, with no real good solution. The only viable option is to try to run something like Native Instruments' Kontakt on Linux, which a lot of people do, from what I understand. I would love to see Linuxsampler try to acheive the same level of functionality as Kontakt, which is a fine sample program. Why they are focusing so much on the Giga library and totally ignoring the Akai library is a mystery to me. Linux Audio needs a full-featured sample program *badly*, and Linuxsampler (from looking at the spec sheets) just ain't it (yet). -- Josh Lawrence http://www.hardbop200.com

Carlo Capocasa wrote: errr, wtf? Why is that then? Both SDS and SMDI are well-documented and common. It's a piece of piss to write an SDS transfer program, and would be even easier if ALSA had a slightly less horrible API for handling sysex (that said, it's the best of a bad bunch). Incidentally, I've written a .wav to .syx converter to turn a sample into a sysex file that you can just fire into the venerable Ensoniq Mirage sampler. Great for crunchy gritty drums, especially with the analogue filters. Gordon.

Erik de Castro Lopo wrote: I'm using libsndfile to read the audio files. The main reason I'm using .wav at the moment is that I know exactly how many samples long the sound is when I edit in Audacity (or whatever). The Mirage can only accept a maximum of 64kB per dump (one half of the memory - yes, 128kB x 8 bit sampling). Gordon.