m0xEE @m0xee@librem.one

@amerika
I've only had proper turntable for a pretty short period of time and I've given all the vinyl records I had to my friends so it's probably too late to start collecting them again.
I do love CDs though! But since shipping and handling started costing more than the CDs themselves (and now it became at all impossible), I stopped buying them.
@mischievoustomato @newt

@newt @mischievoustomato
That probably is true! I'm kinda happy that tinkering with this stuff makes me happy and brings me calm now — it wasn't always like that 😂

@newt
In this particular case avoiding pulse at least allowed me to get rid of separate pulseaudio process that had to be running all the time — less RAM on an already resource-constrained machine, and a couple percent less CPU usage as no unnecessary routing is done. It had its merits.
As I didn't have to spend a week on it, like I said — I just came up with an idea, while messing with alsa on another machine, and it worked, why the hell not?
@mischievoustomato

@newt
I'm not attempting to sell you ALSA and again, I don't think there is need for one-size-fits-all solutions. I just don't need anything pulseaudio/pipewire provides — no transmission over the network, no advanced routing, nothing of it. I just want my output as dumb as it can possibly get.
@mischievoustomato

@newt @mischievoustomato
Why is that? 🤔

@newt
Man, do you realize that I'm just teasing you? I don't want to have a serious discussion about it again. I think you got my point — that theoretically DMA might still be advantageous, and I got yours — in most real world usage scenarios it no longer matters. And I've never stated otherwise: use whatever you're comfortable with — always. My rotten museum-class hardware gets the shit done for me, does it so well that I don't see the point in getting anything new and shiny 🤷

@newt @mischievoustomato
As much as it is embarrassing to admit, but until recently it was even using PulseAudio to work around the bug that at certain sample formats and sampling rates its audio interface was producing silence. Luckily I've found a workaround that allows me using bare ALSA last week. Point is, I wasn't even so much worried about about Pulse in this case — I just came up with an idea of fix and it worked, other than that — let corner cases be corner cases.

@newt @mischievoustomato
Making music and video games are relatively niche use cases compared to what most people use wireless audio peripherals for. In both of these cases wired would probably always be better. No need to apply some kind of universal solution for everything: you know me — I *despise* USB audio, but if it's a tiny machine hooked up to my amp that I can listen my digital collection on and that I can control remotely — why would I care?

@newt @mischievoustomato
Now that is a valid point indeed, I don't use wireless headphones at all, but I suspect it to be pretty much unusable.
I just don't see why you attempt to apply the same principle to audio *consumption*, I didn't even get it at first — what you were talking about and why buffering might be bad.

@newt
No need to! Buffering can defeat the deficiencies even of such a sorry ass of interface as USB. But of course that wouldn't even be necessary were superior DMA-capable hardware used.

@newt
Of course it is better! Might even allow you to implement audio decompression in JS — computing power well spent 😜

@newt
Sure, everyone knows that you need modern desktop-class hardware to play audio 😂

@newt
You obviously have no experience using low processing power machines, which use USB (an utter garbage of interface 😏) for audio, networking AND storage.
Dropouts are gay and lame and whatever, buffering is amazing!

@newt
Wat is wrong with buffering? Without it you're fucked in a lot of cases: network delays, physical media spun down on detecting vibrations to prevent damage, physical media having to be re-read due to errors, etc…
And 1.25 seems to be just right to be reading FLACs on average — studio recordings are often heavily compressed/filtered and use even lower bandwidth.

@AndyGER
Don't get me wrong though, I'm not a fan of everything AI in the slightest, but I won't say that it's bad just because it has "AI" in it in this particular case — it should be good enough in a lot of cases, and in those when it's not, it should provide a good starting point. But I'm afraid that some people would start submitting descriptions generated by it as is — adding a line of warning text seems reasonable.
@duco @3kh0 @mozilla

@AndyGER
Speaking of translation, the result would probably be very similar to attempting to translate a joke that is built on wordplay — lacking at best. Proper translation of such a joke might not exist at all, and automated one will only cause confusion.
@duco @3kh0 @mozilla

@AndyGER
Translation and describing the image are two separate tasks. Sure, if it's a photo of nature, this AI thing might do a pretty good job, better than a lot of us would be willing to put time and effort in, but if it's some humorous image, I'm afraid, it would do a very poor job — the description it would generate will have lots of unnecessary details: it simply doesn't know what to focus on.
@duco @3kh0 @mozilla

@j4n3z @mntmn
To a point that I was able to build and use it even on my old PowerMac G5.
I might be wrong about it, but I think Node.js is no longer even supported on Big Endian PowerPC architecture — wouldn't be a problem in this particular case, but this makes openrazer more portable.

@j4n3z
This one? https://github.com/the-via/app
Didn't know about it 😅 Considering it's specifically for QMK keyboards, it definitely makes sense.
To me Node and Chrome/Electron seem too heavy of a requirement though.
openrazer is completely opensource BTW, all parts of it: kernel modules, the daemon and the Python module.
@mntmn

@romin
Of course not! In fact, I'm three bitter old men, who don't love anybody, in a catsuit!