| Time |
Nickname |
Message |
|
01:26
🔗
|
|
davidar has joined #jsmess |
|
03:31
🔗
|
SketchCow |
I'm excited we're so along on c64 |
|
03:31
🔗
|
SketchCow |
We could punt and do the 24 but I feel like the 32 is just missing something obvious |
|
03:50
🔗
|
|
zino has quit IRC (Excess Flood) |
|
03:52
🔗
|
|
zino has joined #jsmess |
|
03:55
🔗
|
|
balrog_ has joined #jsmess |
|
04:08
🔗
|
|
balrog has quit IRC (Ping timeout: 1212 seconds) |
|
04:08
🔗
|
|
balrog_ is now known as balrog |
|
04:32
🔗
|
Sgeo_ |
It would probably be better if I had an easy way to iterate on it. Suppose I could try builds with 10 years hvsc built in |
|
05:07
🔗
|
|
Lord_Nigh has quit IRC (Ping timeout: 268 seconds) |
|
05:07
🔗
|
|
logchfoo3 has quit IRC (Ping timeout: 268 seconds) |
|
05:09
🔗
|
|
logchfoo0 starts logging #jsmess at Mon Sep 10 05:09:06 2018 |
|
05:09
🔗
|
|
logchfoo0 has joined #jsmess |
|
05:09
🔗
|
|
Lord_Nigh has quit IRC (se.hub efnet.portlane.se) |
|
05:09
🔗
|
|
logchfoo3 has quit IRC (se.hub efnet.portlane.se) |
|
05:09
🔗
|
|
SketchCow has quit IRC (se.hub efnet.portlane.se) |
|
05:09
🔗
|
|
bai has quit IRC (se.hub efnet.portlane.se) |
|
05:13
🔗
|
|
SketchCo1 has joined #jsmess |
|
05:25
🔗
|
|
LordNigh2 is now known as Lord_Nigh |
|
06:17
🔗
|
Sgeo_ |
https://github.com/wothke/websid/blob/master/makeEmscripten.sh |
|
06:17
🔗
|
Sgeo_ |
Build file of an Emscripten-based SID (C64 music) player. |
|
06:18
🔗
|
Sgeo_ |
Obvious differences: -Os -O3 (optimization), closure (Google's JS optimizer thing), Binaryen (I need to figure out what this is) |
|
06:32
🔗
|
db48x |
Sgeo_: have you checked that vice and the browser agree on the sample rate? |
|
06:33
🔗
|
db48x |
we had a similar problem with PCE where the emulated system could only produce sound samples at a fixed rate, and it wasn't one that was supported by SDL |
|
06:33
🔗
|
db48x |
so we have to resample it ourselves before handing it off to SDL |
|
06:36
🔗
|
Sgeo_ |
...my audio knowledge is limited, I'm not sure what a sample rate is, exactly |
|
06:36
🔗
|
db48x |
ok |
|
06:36
🔗
|
db48x |
a sound wave is an analog signal; a wavy line that goes up and down |
|
06:37
🔗
|
db48x |
sampling is just measuring the height of the wave repeatedly to get a list of digital numbers |
|
06:37
🔗
|
db48x |
lists of digital numbers can be easily manipulated by computers to do mixing, effects, etc |
|
06:38
🔗
|
db48x |
then you send the numbers to your sound card which runs them through a DAC to turn them back into an analog waveform |
|
06:38
🔗
|
db48x |
but everyone has to aggree on the sample rate, which is the number of samples per second |
|
06:42
🔗
|
db48x |
if, for example, vice is producing samples at 44100 samples per second, but SDL (and therefore the browser, the OS, and the soundcard) thinks that the proper rate is 48000 samples per second, then every little bit of sound that vice produces will be played back too fast |
|
06:42
🔗
|
db48x |
leaving regular gaps in the playback (and shifting the frequencies up a notch) |
|
06:47
🔗
|
Sgeo_ |
Since VICE has an SDL implementation, and sound sounds good in that, it can't be a mismatch between VICE and SDL, can it? |
|
06:48
🔗
|
|
Ipggi has joined #jsmess |
|
06:49
🔗
|
Sgeo_ |
"ntsc is important because browser requestAnimationFame is going to deliver 60 fps which means less cpu time is wasted during vsync delay" |
|
06:50
🔗
|
Sgeo_ |
That's what vice.js docs say for sync mode. I think my build is considered async mode? |
|
06:50
🔗
|
db48x |
yes |
|
06:51
🔗
|
db48x |
if you're building with SDL2, then I would expect the sample rates to be matched already |
|
06:51
🔗
|
Sgeo_ |
SDL1 |
|
06:51
🔗
|
db48x |
if you're building with SDL1, and vice outputs anything other than 48k, then probably not |
|
06:53
🔗
|
Sgeo_ |
VICE's sample rate is configurable, but vice.js insists it should be 22050 for some reason. Also, hmm, taking a look at the configure change, it looks like maybe... soundsdl.o isn't being included? I'll need to check that |
|
06:54
🔗
|
Sgeo_ |
SketchCo1, check what happens and if anything breaks if you change the 22050 argument to 48000. |
|
07:03
🔗
|
Sgeo_ |
Argument change does not seem to be working |
|
07:19
🔗
|
Sgeo_ |
Increasing the soundbufsize makes it sound worse. Decreasing it throws an error |
|
08:15
🔗
|
|
arkiver has quit IRC (irc.Prison.NET west.us.hub) |
|
08:15
🔗
|
|
SketchCo1 has quit IRC (irc.Prison.NET west.us.hub) |
|
08:16
🔗
|
|
logchfoo1 starts logging #jsmess at Mon Sep 10 08:16:39 2018 |
|
08:16
🔗
|
|
logchfoo1 has joined #jsmess |
|
08:17
🔗
|
|
zino_ has joined #jsmess |
|
08:22
🔗
|
|
azakai_ has joined #jsmess |
|
08:26
🔗
|
|
Sgeo_ has joined #jsmess |
|
08:49
🔗
|
|
Sgeo_ has quit IRC (west.us.hub irc.Prison.NET) |
|
08:55
🔗
|
|
Sgeo_ has joined #jsmess |
|
15:21
🔗
|
|
bai_ is now known as bai |
|
15:29
🔗
|
|
db48x has joined #jsmess |
|
16:02
🔗
|
SketchCo1 |
can help once I labd |
|
16:03
🔗
|
SketchCo1 |
land |
|
16:43
🔗
|
|
azakai_ has quit IRC (Quit: Ex-Chat) |
|
22:45
🔗
|
|
SketchCo1 is now known as SketchCow |
|
22:46
🔗
|
SketchCow |
Sgeo_: Still need those tests? |
|
23:38
🔗
|
Sgeo_ |
SketchCow, no. Was able to try it myself. I found a setting that made it worse, but the opposite of that setting caused it to crash |