Re: [Musicpd-dev-team] [PATCH/RFC] Fix DSD-over-USB playback
On 2012/04/06 12:52, Jurgen Kramer gtmkra...@xs4all.nl wrote: The attached patch fixes playback for DSD-over-USB. Without this patch MPD uses a frame size of 2 (stereo files) while the actual needed frame size for the used output sample format is 8 (sample formats S24_P32 and S32). No, 8 is the wrong value for in_frame_size. The problem is not a wrong value for in_frame_size, but that in_frame_size is used where this variable is not appropriate. Your patch trades one breakage for another. The right approach would be to realize that two different in_frame_size variables are needed: one for the buffer passed to play(), and another one for the pcm_export result. This checks shift8 twice: + if (ad-export.shift8) { + g_warning(Adjusting in_frame_size for DSD-over-USB output\n); + ad-in_frame_size = ad-export.shift8 ? 8 : ad-in_frame_size; Why this warning message? Does the user care? -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH] Add support for DSF files to DSDIFF decoder
On 2012/04/14 17:05, Jurgen Kramer gtmkra...@xs4all.nl wrote: Attached patch adds support for DSF files to the DSDIFF decoder. It can be used in conjunction with DSD2PCM and DSD-over-USB. This looks like this supports a different file format and should live in another decoder plugin. Is there an advantage in implementing both formats in one plugin? I get build errors: src/decoder/dsdiff_decoder_plugin.c: In function 'dsdiff_read_metadata': src/decoder/dsdiff_decoder_plugin.c:376:2: error: format '%u' expects argument of type 'unsigned int', but argument 4 has type 'goffset' [-Werror=format] src/decoder/dsdiff_decoder_plugin.c:357:11: error: unused variable 'samplecnt' [-Werror=unused-variable] src/decoder/dsdiff_decoder_plugin.c:330:11: error: unused variable 'metadata_offset' [-Werror=unused-variable] src/decoder/dsdiff_decoder_plugin.c:327:11: error: unused variable 'file_size' [-Werror=unused-variable] -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] [PATCH]
On 2012/04/14 17:07, Jurgen Kramer gtmkra...@xs4all.nl wrote: Patch to update doc/user.xml to reflect added support for DSF files to the dsdiff decoder. Not merging because the DSF patch hasn't been merged yet (see my question in that thread). Please fold into the DSF patch. -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] curl streams EOF after a long pause
On 2012/04/14 15:29, Iskren Chernev iskren.cher...@gmail.com wrote: I think a better approach would be to try and reconnect to the stream (especially if its non seekable -- I'm not 100% if that happens only for radios, or some stupid https serving mp3 may also be), all the time, or maybe after a long pause. I just got a rough understanding of how the input and decoder layer work, but I'm not sure if the change should be introduced inside input_curl or on a higher level. We have a bug report for this, but I havn't gotten around to fix it yet. What we need is two things: - the input plugin API should get a new method, maybe called pause(), which suspends buffering for relevant plugins. - for radio streams (remote, not seekable), the pause() method could tell its caller don't bother keeping me open, and some code a few layers up will close it and reopen on resume Max -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team