Marvin Adeff wrote: > Sorry it took so long to respond Marco, it took a while before i had > time to try your suggestions. Thank you for your help. > > Interestingly enough, none of steps a - c solved the hiccup. But when i > lowered the upsample rate, the problem disappeared. On playing with the > settings more, i discovered that when i had 2 rates selected (352800 and > 384000 -- my original setting) the hiccup occurred almost every time. > But if i deselected one or the other, the problem disappeared again. So > i left only 384000 selected. No hiccup. And in fact i noticed an > immediate improvement in fidelity. The fidelity is quite excellent! > > My configuration has squeezelite running on an RPi3 with piCorePlayer > 3.5.0Audio. Squeezelite is compiled per Soundcheck's advices ( > https://soundcheck-audio.blogspot.de/2018/01/raspberry-pi-audio-engine-part-5.html > ) using: > export CFLAGS="-O3 -mcpu=cortex-a53 -mfpu=neon-fp-armv8 > -fno-delayed-branch -fno-selective-scheduling2 -fno-whole-program > -mfloat-abi=hard -fno-fast-math -pipe -fstack-protector > --param=ssp-buffer-size=4 -DVISEXPORT -DLINKALL" > > I don't care about DSD. I'm only upsampling flac to pcm at the highest > rate supported by the Piano/Kali DAC hat -- 384000 (at that rate the > Piano/Kali bypasses any filtering). The idea being to minimize cpu load > on the RPi3 and let the LMS server do all the heavy lifting. All my > settings are toward minimum phase response, both in 3-CPO and > Piano/Kali. > > I run LMS 7.9.1 on a seperate Windows 10 box. I replaced the LMS > included flac.exe, sox.exe and faad.exe with the latest versions > available at their respective websites. They are connected over WiFi > 108.11n 5MHz. I turned off the RPi3's onboard WiFi and have an external > WiFi dongle over the main USB port (also following soundcheck's advices > in his blog) with a .5 meter extension cable to get it away from the > RPi. > > Is this behavior with two sample rates selected to be expected? Any > suggestions?
Please try to use squeezelite on the win machine to see if the problem is the transcoding process or the RPI reproduction. If the problem is the transcoding: What's the setting for "target sample rate" (closest supported, Closed synchronous,...)? Selecting one of the latest, C-3PO investigate the actual sample rate of the file, calculate the target sample rate accordingly, then LMS start an addictional, separated, process to transcode, in windows, involving some particular code (socketwrapper), known to be not so efficient, that's the only difference. If that is the problem, the only chance you have is to fix the output sample rate or select the first option, but I developed C-3PO in windows and I never had this problem, so it looks strange to me. ________________________________________________________________________ Author of C-3PO plugin, Squeezelite-R2, Falcon Web interface - See www.marcoc1712.it ------------------------------------------------------------------------ marcoc1712's Profile: http://forums.slimdevices.com/member.php?userid=34842 View this thread: http://forums.slimdevices.com/showthread.php?t=105309 _______________________________________________ Squeezecenter mailing list [email protected] http://lists.slimdevices.com/mailman/listinfo/squeezecenter
