Re: [Musicpd-dev-team] [PATCH] Use g_message and not g_debug when removing song
On 2012/03/23 16:32, Dan McGee d...@archlinux.org wrote: When adding or updating a song, we get a log message even if debug is not enabled. It seems odd that removing a song shouldn't be done at the same log level; otherwise looking at the log leads you to believe songs are never removed from the library on update. Merged to v0.16.x. -- This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] DSD over USB
On 2012/03/22 17:23, Jurgen Kramer gtmkra...@xs4all.nl wrote: I took a quick look at the code, it seems it supports the pre-1.0 version of the standard (0xAA markers). I'll give it a test tomorow. My DAC supports both version 1.0 and pre-1.0. It implements the newest spec I found, dated September 2011, announced on the dCS web site in October 2011. Where can I find a more recent specification? Max -- This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] DSD over USB
On 2012/03/26 19:28, Jurgen Kramer gtmkra...@xs4all.nl wrote: The closest I got is link which a sent to the list a while ago: http://www.audiostream.com/content/usb-link-dsd-audio-pcm-frames-andres-koch It isn't as clear as a 'real' standard but it contains the needed info. Hm, this doesn't state that this document is the successor of the one published by dCS. Not even the revision history mentions it. BTW Max, it seems that the current 24-bit format for DSD-over-USB causes problems with modern audio I/F (e.g USB 2.0 devices using UAC 2.0 and some other FW devices) only supporting 32-bit samples (and sometimes 16-bit as well). So we could either switch to 32-bit sample format for DSD-over-USB or let MPD convert on the fly. Personally I vote for the 32-bit sample format as most, if not all, devices supporting DSD-over-USB will accept that format. How does packing DSD-over-USB into 32 bit samples work? -- This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] DSD over USB
On Mon, 2012-03-26 at 19:36 +0200, Max Kellermann wrote: On 2012/03/26 19:28, Jurgen Kramer gtmkra...@xs4all.nl wrote: The closest I got is link which a sent to the list a while ago: http://www.audiostream.com/content/usb-link-dsd-audio-pcm-frames-andres-koch It isn't as clear as a 'real' standard but it contains the needed info. Hm, this doesn't state that this document is the successor of the one published by dCS. Not even the revision history mentions it. OK, I'll ask around to see if there is something more resembling a standard. BTW Max, it seems that the current 24-bit format for DSD-over-USB causes problems with modern audio I/F (e.g USB 2.0 devices using UAC 2.0 and some other FW devices) only supporting 32-bit samples (and sometimes 16-bit as well). So we could either switch to 32-bit sample format for DSD-over-USB or let MPD convert on the fly. Personally I vote for the 32-bit sample format as most, if not all, devices supporting DSD-over-USB will accept that format. How does packing DSD-over-USB into 32 bit samples work? In code it looks like this :-) The lower 8-bits are not used so basically it's still a 24-bit sample (that sort of details you would want to find in a standard). G_GNUC_CONST static inline uint32_t pcm_two_dsd_to_usb_32b(uint8_t a, uint8_t b) { return 0xaa00 | (a 16) | ( a 8 ); } /* JK: DSD-over-USB v1 */ G_GNUC_CONST static inline uint32_t pcm_two_dsd_to_usb_marker1_32b(uint8_t a, uint8_t b) { return 0x0500 | (a 16) | (b 8); } G_GNUC_CONST static inline uint32_t pcm_two_dsd_to_usb_marker2_32b(uint8_t a, uint8_t b) { return 0xfa00 | (a 16) | (b 8); } Jurgen -- This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] DSD over USB
On 2012/03/26 19:37, Jurgen Kramer gtmkra...@xs4all.nl wrote: In code it looks like this :-) The lower 8-bits are not used so basically it's still a 24-bit sample (that sort of details you would want to find in a standard). So I can just left-shift the 24 bit sample by 8 bit? Neither document says anything about this, so I assumed it was not a legal transformation. (I have no DSD hardware) -- This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] DSD over USB
On Mon, 2012-03-26 at 19:43 +0200, Max Kellermann wrote: On 2012/03/26 19:37, Jurgen Kramer gtmkra...@xs4all.nl wrote: In code it looks like this :-) The lower 8-bits are not used so basically it's still a 24-bit sample (that sort of details you would want to find in a standard). So I can just left-shift the 24 bit sample by 8 bit? Neither document says anything about this, so I assumed it was not a legal transformation. (I have no DSD hardware) I can not vouch for that, I am neither the standard developer nor intimate with the UAC 2.0 spec this standard was originally meant for. All people who tested this are on Linux (AFAIK). My guess is, it will work on all little endian systems. Viewing a binary dump, data comes out as 0x00 L2 L1 MARKER etc. Jurgen -- This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team