No announcement yet.

Firefox 99 Available With Strengthened Linux Sandbox, Web MIDI

  • Filter
  • Time
  • Show
Clear All
new posts

  • #61
    Originally posted by DrYak View Post
    That's why some crazy japanese people will make a Sound Canvas sing the lyrics of Bad Apple through composite sine
    Pfft! snd-pcsp (Linux kernel module) has done PCM output using a PWM-ed square wave!


    • #62
      Originally posted by Sonadow View Post
      The only advantages Firefox has over Chrome(ium) at this point of time are[LIST=1][*]Works almost perfectly under Wayland with proper hidpi scaling. Chrome(ium) completely fucks up its scaling on hidpi displays and I'm always forced to fucking down-res the display from 4K to 1080 instead as a workaround.
      I have been using both Firefox and Chromium on Wayland eversince they started advertising native Wayland support. I've never had any serious issues with scaling. Even right now I am using a 27" 4k display for desktop use and a 13" 4k display on my laptop. Scaling settings are different, obviously, for these two displays but even hotplugging the laptop into the external display works fine and both browsers change their scaling accordingly.

      I remember having some issues with scaling when initially I had to input the settings manually into a config file, and later on the browsers gained the ability to automatically detect optimal scaling settings, which then went nuts when combined with my on-disk configs. Everything settled down when I understood to delete the old config files.


      • #63
        Originally posted by Ironmask View Post
        You can ignore all those things and I myself ignore them. I use Brave search now as a replacement for DuckDuckGo since DDG is stagnating IMO (and now openly supports censorship), but I ignore the crypto stuff and I don't even really pay attention to Brave's adblocker, I think uBlock Origin is superior so I still use that. What I certainly make use of though is the built-in Tor mode.
        So the browser doesn't consume tons of resources with those "extra" features?
        Yes a substitute for DuckDuckGo is needed ASAP, agree entirely on that!

        Originally posted by Ironmask View Post
        I really don't understand this argument. "I don't know how to feel about Brave, because of all those optional opt-in features! I don't know what to do with them all!" it sounds sarcastic, how is this an actual thing people respond with when asked about Brave? What is this clown logic? Sorry, I'm assuming you use Linux since you post here? Do you know what options Linux has? It's a lot more than the couple opt-in things Brave has, I can tell you that.
        Ironmask, I am not against Brave

        I use unix systems at least 20-25 years now, so yes Linux also.
        I simply would love to see a browser that does what is supposed too, and no more strange things.

        But its good to know that you can use brave without it impacting your system( with all those p*rn features.. )


        • #64
          Originally posted by uxmkt View Post
          Pfft! snd-pcsp (Linux kernel module) has done PCM output using a PWM-ed square wave!
          Had been doing that in DOS in assembler when I was a kid (combined sound playing + joystick polling routine. Though I also did some of the sound mixing in the interrupt handler which wasn't a very bright idea in retrospect).

          And later tricking my AdLib through volume modulation, too. (Didn't have the necessary pieces to solder my own Covox Speechthing-like parallel port DAC like the cool kids. Eventually I got a Sound Blaster at Christmas later on)

          But all in all, playing PCM is relatively straight forward, and PWM is too. (Specially on a device where the speaker is more or less directly connector to the output of a timer. It's pretty low-level doing crazy things with it did entirely make sense).
          The magic in Linux comes from managing the precise timing to achieve it.
          (No DMA at all + Single pulse trigger on the PC Speaker = so you must manually trigger a pulse for every sample. It's easy to achieve in a single task environment like DOS (use a timer IRQ on AT+ class computers, or count your cycles on older PC). It a much more involved feat in a multi-tasking multi-user environment, where you also need to keep all the other tasks into consideration.

          Composite sine synthesis of MIDI is an entirely different can of worm. You aren't playing straight PCM, but you need to correctly extract your formant from the sound (Fourrier transform + lots of clean-ups upstream of the playing) and you're playing it over a channel that is very high-level: you aren't directly driving anything, you're passing very abstract messages (to the instrument(s) listenning on channel CC: please select instrument number XX from your bank, and then please play note YY, with velocity ZZ).

          The PC Speaker equivalent would be trying to play PCM samples... while doing it using exclusively Music Macros Language (e.g.: MS-BASIC's PLAY command).

          (Which BTW could perhaps be achieved but not on stock music macro. You'd need a fair amount of PEEKing and POKEing (and IN/OUTing) to change the frequency of the interrupt handler used by the BASIC's MML player, remove sanity checks in the MML interpreter so you could send arbitrarily short durations, probably tweak both timers even further, and use a very beefy computer so the whole MML macro routine can run at the high frequency required by PWM).

          (And if you think about it, those PEEK/POKE/IN/OUT commands/functions in the above speculation are BASIC's equivalent to MIDI's SYSEX).