Hi Dave and JM, I built a .lib for arm architecture. I work on VS2005, and it's actually easy to build. What you have to do is to change the platform on which you want to build your lib. Open the libspeex project in "speex-1.2beta1\win32\VS2005\libspeex" In "Build" menu, choose "Configuration Manager" under "Active solution platform" select "<New...>" choose the platform you want (...ARM...) and under "copy setting from" select "Win32". I think you have to use FIXED_POINT mode as well because pocket pc doesn't really support FLOAT_POINT. To do that, go in libspeex properties, then in "C/C++, Preprocesor" add FIXED_POINT in "Preprocesor Definitions" Compile and you should have libspeex.lib in " speex-1.2beta1\win32\VS2005\libspeex\Pocket PC 2003 (ARMV4)\Release" I've good results with that. I can record and encode the sound picked up from PPC's microphone. Then I send it to an another PPC on which I decode and play. (Actually, I send, through an AP, to itself or to a PC because I have only one PPC). My next step is to adapt the sound's quality to the congestion's network. What I'm trying to say is, if the network is congested, I would modify dynamically the quality by reducing it in order to send less datas to keep a fluent conversation. (and if the network is not congested, possibly increase the quality). Do you understand what I mean ? And so, I've a question for JM or anybody who could help me. Is it possible to modify dynamically the quality/complexity/... or anything else to reduce what I send ? Regards, Patrick. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.xiph.org/pipermail/speex-dev/attachments/20061114/b92f9f4c/attachment.html