I wonder if linux audiophiles are a thing.
"plain ALSA sounds warmer than pipewire"
"linux kernel compiled with ALSR makes for a more chaotic high end"
"ext2 sounds better than ext4"
I wonder if linux audiophiles are a thing. "plain ALSA sounds warmer than pipewire" "linux kernel compiled with ALSR makes for a more chaotic high end" "ext2 sounds better than ext4" 148 comments
12
@vidister Well at least files becoming lossy on ext4 would be true :P
0
0
16 September at 10:13 | Open on queer.hacktivis.me
@vidister there was a reddit post or something a whole back where someone was unironically arguing about how the audio driver using C sounded better or some such @arichtman I suspect all that ferrous oxide that people like to introduce in modern software development might make a difference. (As for audiophiles: I seriously had someone tell me that I'd get better audio if I used a high-end gold-plated €150 USB cable instead of a cheaper one. I really wasn't in the mood at the time to explain time-/frequency-domain equivalence and differential signalling... at the time I basically just shrugged and for once thought "well, you do you". 🙂) @pdh_sedlr @f4grx @noodle @vidister warm up those sweet sound waves for max audiophile listening pleasure of course. @datacop @vidister ironically one could argue, that JACK cannot be audiophile, since it uses single precision floating point (IEEE754 32 bit) for samples, and will irrecoverably loose precision if somewhere along the signal chain the values exceed full scale. Ages ago I one argued with the developer of JACK about the pros/cons of using float samples. @vidister@chaos.social "I use a realtime kernel because otherwise stuff can in the way of my music player" @vidister I mean... the colors are different if you ask me. ALSA is a fierce red compared to the calmer black and blue of pipewire. Might be my #synaesthesia speaking more so than acoustics, though 😬 @vidister How much, if any, is down to resampling at either the software or DAC level making the difference? None of that stuff should happen with straight up alsa. I don't consider myself quite an #audiophile saying that: - FIFO: First In First Out #linux #audio #mao #computermusic #recording #homestudio (I think I read recently that switching PREEMPT_RT on the fly was in the work for a future kernel release, but I just skimmed the article) ;-) that's the default Linux settings NOT to have PREEMPT_RT stuffs enable. But that constrains heavily on Real-Time audio computing. You have to recompile the kernel otherwise, or use a distro which is based on such a kernel (like Ubuntu Studio) For playing Shirley Manson, you have to switch the Pink Hair flag in the BIOS and that's it ;-) @vidister @lunareclipse
alsa and pipewire introduce too many delays. I use jack for real time audio on my zoom calls act 1: shitposting about linux audiophiles act 2: post is boosted by gargron, immediately blows up act 3: people start arguing about audio stuff in my mentions sometimes the mere act of joking about a hypothetical thing brings it into existence @CIMB4 ah, yeah, accepted. @vidister I wanted to respond to this with a snarky message intended for humor, but even my best attempts at it just read like a real post I've undoubtedly read on a forum somewhere else. @vidister Careful, Icasus. you’ve learned the basic nature of the tohu wa bohu. But you’re not a wild card yet … @vidister all the time @vidister Linux audiophiles? You mean there are people who actually got their audio working on Linux? Sorcery. @aruban35mm @vidister i mean, there's at least something to it: PA might do automatic resampling while ALSA when exclusively using the device does not. :D @vidister@chaos.social @q66 @vidister My favorite is still the gold-plated Toslink cable. Because it's become so mainstream that it's the only type of toslink cable available in some stores (I mean, stores that don't even target a particularly audiophile audience). Gold plating is already rarely useful (if ever) for a digital connector, but Toslink is an optic fiber, often made out of plastic (and sometimes glass)! 🤣 @vidister "MP3s sound so much warmer than the clinical FLACs, especially with VBR" "I have a collection of original IUMA recordings I like to play on my emulated WinAMP 0.20a" "For extra cleanliness I like to use the original Frauenhofer decoder and pipe it to /dev/audio on a classic kernel" @cdamian @vidister @Gargron to the hardware guys: has anyone built a mostly-analog MP3 decoder already? Surely there is a bit of digital frontend, but do we really have to all computations in cold digital realm and then put it through a digital-sounding DAC? it'd be much warmer to the ear if most of the decoding was already pure and fat analog!! "pulseaudio doesn't crash bluetooth as much anymore" "yes, I can use high fidelity sink when using my microphone. Don't know what y'all issues are." @vidister I remember following tutorials to achieve "bit perfect" audio on Linux many years ago 😋 @vidister Definitely not an audiophile (those are delusional people), but old fashioned 4front oss sounded better than ALSA. Perhaps, all these centuries later it might not be true, but it was then. @vidister it surely exists, but there isn't a huge industry propping it up, making tons of money off gullible people. @vidister I mean, the first thing that PulseAudio did was to add at least half a second latency. (Low latency sound server my a**.) Watching videos or playing games was impossible. But back then simply uninstalling PulseAudio was easy. Later it got too deeply integrated into the distributions that I used so when you uninstalled PulseAudio you had no sounds at all. Haven't tried PipeWire yet. I belief so, after I read a report about the sound quality of different Ethernet cables in a German audiophile magazine. And it was NOT the April edition.
I mean, I *did* once read (well, skimmed) an article comparing different brands of SSD for sound quality when playing lossless audio... (Nobody--not me, not the link aggregator I found it on--could tell if it was satire or not.) @suetanvil @vidister oh good, I get to link to the old but still excellent Wat HiFi? — https://wathifi.com/ It's not quite new enough to feature the audiophile SSDs with the system crystal replaced by a bulky Oven Controlled Crystal Oscillator, but still brings a tear to the eye @vidister Anything which reduces jitter will likely make audio sound better, but special hardware like streamer or DAC hats are the way to go. @vidister ZFS is definitely required for maximum sound reproduction. The very low latency of an array of disks and fast seek allow for the DAC to never be starved. The parity bits going back to the CPU with the data provide a gentle filter to the data, shielding it from the electrostatic noise of your other components. The recommended configuration is 5 disks raidz2, mirrored. The mirroring allows to filter off machine noise from signal, again improving clarity. @vidister Oh 100%. My ears are shit but I swear plain ALSA just sounds fresher than pulseaudio @vidister music from non-advanced format drives sounds better than from advanced format and shingled sounds worse than conventional. I can hear parts of the other music tracks on those (/j) @vidister You guys got it all wrong. It's the old story of electrolytic vs. solid capacitors. The electrolytic sound considerably warmer with a 6.11 Kernel, at least as long as they haven't dried out, exploded or whatnot. If - and I repeat, this is very undesirable in terms of sound quality - if you absolutely have to use solid capacitors, go for tantalum polymer ones paired with a 4.x Kernel. Everything else is inferior. I can clearly hear the difference on my rig. @vidister there is a certain je ne sais quoi to the staccato click of pipewire xruns that you don't get with regular alsa @vidister this is what JACK is for Linux audiophiles are obsessed with low latency rather than warmth or audio quality. The patches are dead now but some of them still compile their own custom kernels, eg. https://wiki.linuxaudio.org/wiki/system_configuration @vidister well, good music takes you on a journey, so ext4 has some clear advantages for that. @vidister Linux audiophile be like “ooh I think I hear a sound” after seven hours of tinkering with alsaconf @vidister M-Audio Audiophile 2496 Sound Card @madonius@chaos.social @vidister@chaos.social "btrfs makes the sound more ...buttery" @vidister yep, some of us out here 😂 Music does sound better on my Linux running empeg https://www.empeg.com @vidister I've lived in that household and just the thought of it has triggered the twitching. But if I remember correctly: have a dedicated machine with no internet access, use a low-latency kernel, don't install X, don't use ALSA, pre-load your FLACs into memory to avoid jitter from the fs, buy a custom DAC, spend a fortune on cables, give up and develop a vinyl fetish instead. @vidister searches thread to see if anyone else has had this take ALSA? God that's far too harsh, I maintain my own patch set of OSS/Commercial that I rebase against each new kernel as it comes out. I could go with a newer official release of OSS, but they don't have as wide a sound as the older ones (...TIL OSS for Linux is still a thing that had an update as recently as 2019) @vidister Not *exactly* audiophile, but... https://mastodon.gamedev.place/@nh/112097524494239379 @vidister reminds me of the time I was doing a search on I can't remember what about PCIe and one of the first results was a thread on an audiophile forum where people were debating what was better between a SATA or NVMe ssd for audio quality. @vidister Linus must block merging Rust sound drivers into the kernel! Dynamic range is utter trash with drivers that use substructural type systems. @vidister@chaos.social i've never thought or knew about this intersection of topics before until now. does audio sound better open source? /j @vidister I'm a linux audiophile, i spend most of my life telling other audiophiles that digital transmission of files is lossless. Also is the best way to listen to Dubstep... because even when it (alsa) works correctly... You still get a slightly different remix of your music every single time you play it. 😉 @vidister probably, we definitely have people who think vectorised memcpy ruins the sound and they need to use a classic scalar memcpy for better soundstage. this is not a joke @vidister I have long argued that we should use the term "audiophile" in place of the (racist, ESR-promoted) term "cargo cult". The so-called "cargo cults" are full religions practised by informed 21st century members, and portraying them as mere ignorance and blind adherence to routine is offensive. @vidister My rationale for this is that the stuff we think of as audiophile quackery now looks not far off from the sort of thing that was necessary to get decent sound out of 1950s and 1960s stereo equipment. It's absurd these days because technology has rendered it all obsolete. It's as if electric car owners were still insisting on installing carburettors, and tuning them obsessively. @vidister I’ve seen it over the years with our Logic Pro releases: “the new version sounds warmer, better, worse, quieter, etc.” despite that we tested that the digital output is identical for an identical song. @vidister the conversations below remind me that someone did make an ffmpegFS module. https://github.com/nschlia/ffmpegfs read-only FS in tandem with ffmoeg to dynamically transcode media. |