00:18
<
pie_ >
sooo...nix-on-droid lets me use taskwarrior on my phone
00:18
<
pie_ >
not sure where that gets me and typing or horrible
00:24
drakonis has joined #nixos-chat
00:25
waleee-cl has quit [Quit: Connection closed for inactivity]
00:30
drakonis has quit [Ping timeout: 246 seconds]
00:39
drakonis has joined #nixos-chat
01:11
nh2 has quit [Ping timeout: 250 seconds]
01:11
sdier has quit [Ping timeout: 252 seconds]
01:13
sdier has joined #nixos-chat
01:13
ctp has joined #nixos-chat
01:13
aria has joined #nixos-chat
01:13
nh2 has joined #nixos-chat
01:13
elvishjerricco has joined #nixos-chat
01:32
drakonis has quit [Ping timeout: 252 seconds]
01:36
drakonis has joined #nixos-chat
01:39
<
{^_^} >
#73192 (by pasqui23, 4 days ago, open): nixos-shell:init at 2015-05-23
02:06
pie_ has quit [Ping timeout: 240 seconds]
02:54
red[evilred] has quit [Quit: Idle timeout reached: 10800s]
03:33
worldofpeace has quit [Quit: worldofpeace]
03:35
drakonis_ has joined #nixos-chat
04:42
drakonis1 has joined #nixos-chat
04:46
drakonis has quit [Ping timeout: 252 seconds]
05:07
drakonis_ has quit [Read error: Connection reset by peer]
05:28
evanjs has joined #nixos-chat
05:29
evanjs has quit [Client Quit]
05:32
evanjs has joined #nixos-chat
05:49
endformationage has quit [Ping timeout: 240 seconds]
06:32
monsieur1 has joined #nixos-chat
06:32
chreekat has joined #nixos-chat
06:37
cransom has quit [*.net *.split]
06:40
xd1le has joined #nixos-chat
06:40
chreekat has quit [Ping timeout: 252 seconds]
06:42
kgz has quit [Ping timeout: 240 seconds]
06:42
samueldr has joined #nixos-chat
06:42
chreekat has joined #nixos-chat
06:43
edef has joined #nixos-chat
06:46
kgz has joined #nixos-chat
06:47
chreekat has quit [Ping timeout: 252 seconds]
06:48
kraem has joined #nixos-chat
06:48
cransom has joined #nixos-chat
06:48
adisbladis has joined #nixos-chat
06:50
chreekat has joined #nixos-chat
07:15
adisbladis has quit [*.net *.split]
07:15
kraem has quit [*.net *.split]
07:15
cransom has quit [*.net *.split]
07:15
monsieur1 is now known as monsieurp
07:16
cransom has joined #nixos-chat
07:16
kraem has joined #nixos-chat
07:16
adisbladis has joined #nixos-chat
07:17
Synthetica has joined #nixos-chat
08:21
Jackneill has joined #nixos-chat
08:22
Jackneill has quit [Remote host closed the connection]
08:23
Jackneill has joined #nixos-chat
08:28
Jackneill has quit [Read error: Connection reset by peer]
08:51
__monty__ has joined #nixos-chat
09:18
Jackneill has joined #nixos-chat
09:47
<
tilpner >
__red__: I think there might be a few people interested in seeing your module for that
10:16
psyanticy has joined #nixos-chat
11:16
<
Twey >
clever: Are you working on IOHK?
11:16
<
clever >
Twey: yeah, ive been with iohk for over 2 years
11:17
<
Twey >
clever: May I ask what you're on/what it's like? It looks like a pretty interesting group
11:17
das_j has joined #nixos-chat
11:17
<
clever >
Twey: i'm on the devops team, dealing with CI and maintaining the servers in aws, and nixops is a must
11:24
<
Twey >
clever: Are you UK-based? Or remote?
11:25
<
clever >
Twey: atlantic canada, and remote
11:26
<
__monty__ >
IOHK's 100% remote, no?
11:27
<
clever >
__monty__: yeah, pretty much everybody is remote
11:30
<
__monty__ >
Let me know when IOHK starts looking for juniors.
11:30
evanjs has joined #nixos-chat
11:31
xd1le has quit [Quit: leaving]
12:08
avn has joined #nixos-chat
13:24
tilpner has joined #nixos-chat
14:07
endformationage has joined #nixos-chat
14:20
worldofpeace has joined #nixos-chat
14:25
aanderse has joined #nixos-chat
14:27
avn has quit [Read error: Connection reset by peer]
14:28
avn has joined #nixos-chat
14:36
rardiol has joined #nixos-chat
14:43
atopuzov[m] has joined #nixos-chat
14:43
colemickens has joined #nixos-chat
14:43
thefloweringash has joined #nixos-chat
14:43
dtz has joined #nixos-chat
14:43
jtojnar has joined #nixos-chat
14:43
tokudan[m] has joined #nixos-chat
14:43
leons has joined #nixos-chat
14:43
arcnmx has joined #nixos-chat
14:47
waleee-cl has joined #nixos-chat
14:56
pie_ has joined #nixos-chat
15:01
tilpner has quit [Quit: tilpner]
15:09
worldofpeace has quit [Ping timeout: 252 seconds]
15:30
worldofpeace has joined #nixos-chat
15:46
tilpner has joined #nixos-chat
16:23
jD91mZM2 has joined #nixos-chat
16:36
tilpner has quit [Quit: tilpner]
16:45
tilpner has joined #nixos-chat
16:49
{^_^} has joined #nixos-chat
16:56
tilpner has quit [Read error: Connection reset by peer]
16:57
tilpner has joined #nixos-chat
16:58
drakonis1 has quit [Ping timeout: 245 seconds]
16:59
<
aleph- >
clever: How is nixops in production?
17:00
drakonis1 has joined #nixos-chat
17:03
drakonis has joined #nixos-chat
17:05
drakonis_ has joined #nixos-chat
17:06
drakonis1 has quit [Read error: Connection reset by peer]
17:09
drakonis has quit [Ping timeout: 276 seconds]
17:20
<
ashkitten >
is nixops particularly helpful for single servers? i just mostly want a tool to let me test config changes on my server without having to edit it on the server end
17:29
<
cransom >
imo, no. `nixos-rebuild --target-host <somewhere else>` is plenty.
17:30
<
ashkitten >
oh i didn't realize that was a thing either
17:30
<
cransom >
--build-host and --target-host if you wanted to build somewhere else as well
18:13
drakonis_ has quit [Read error: Connection reset by peer]
18:13
drakonis_ has joined #nixos-chat
18:22
drakonis_ has quit [Ping timeout: 250 seconds]
18:23
drakonis_ has joined #nixos-chat
18:34
drakonis has joined #nixos-chat
18:41
<
infinisil >
Well damn
18:41
<
__monty__ >
Hot damn! : )
18:41
<
infinisil >
I wanted to figure out how many kbps I need for opus so I couldn't distinguish it from original flac quality
18:42
<
__monty__ >
It's super small, right?
18:42
<
infinisil >
So I wrote an AB-test script, did a bunch of measurements
18:42
<
__monty__ >
64 or less?
18:42
<
infinisil >
I can distinguish 160kbps from flac..
18:42
<
infinisil >
Oh wait!
18:42
<
infinisil >
Hold up
18:43
<
adisbladis >
It's pretty good
18:43
<
__monty__ >
SWEET, if you decline their cookies you get pure html : D
18:43
<
infinisil >
I made a mistake in my script, I did `--gain 15` for opusdec (to convert it to wav), but that clips loud noises!
18:44
<
__monty__ >
Why would you convert it
*to* wav?
18:44
<
infinisil >
But if I don't do that, my flac file is so much louder than opus, which is obviously detectable
18:44
<
__monty__ >
Gain just obviously distorts sound though.
18:44
drakonis__ has joined #nixos-chat
18:44
<
__monty__ >
Maybe try ffmpeg for the encoding?
18:44
<
infinisil >
__monty__: Convert flac -> opus -> wac/flac, then I use the same player to play the files
18:44
<
infinisil >
More uniform
18:45
<
infinisil >
adisbladis: I'll check it out
18:45
<
__monty__ >
Is that a fair comparison though? The resulting flac could be way bigger.
18:46
<
adisbladis >
(it's just a python script that shells out to ffmpeg)
18:46
<
infinisil >
__monty__: They're all about 15MB
18:47
drakonis has quit [Ping timeout: 265 seconds]
18:48
<
__monty__ >
Ah, the plain html version doesn't have the quiz : /
18:48
worldofpeace has quit [Ping timeout: 276 seconds]
18:49
<
infinisil >
I still like my script more, can choose my own songs :)
18:50
<
infinisil >
adisbladis: Oh but that has only uncompressed, mp3 320 and 128
18:50
<
infinisil >
I'm trying to figure it out for opus :)
18:50
<
__monty__ >
I barely hear a difference between 64Kbps opus and flac. Still store flac though.
18:50
<
adisbladis >
infinisil: Ahh :)
18:50
<
adisbladis >
__monty__: Regardless flac is a better archive format
18:50
<
adisbladis >
Transcode to anything \o/
18:50
<
infinisil >
Do I need that though?
18:50
<
infinisil >
Takes up a lot of space
18:51
<
adisbladis >
infinisil: I store in flac on my file server and vorbis on all playback devices
18:51
<
adisbladis >
It's not that much space, even a 1T drive can fit almost all my music
18:51
<
infinisil >
Yeah I'm also choosing flac as storage whenever available, opus playback through http stream
18:51
<
__monty__ >
Why vorbis? Compatibility? Most things play opus now, no?
18:52
<
adisbladis >
__monty__: Because for music fidelity opus may or may not be better than vorbis
18:52
<
adisbladis >
And vorbis "just works"
18:52
<
adisbladis >
Also I've been using vorbis for close to 15 years now :)
18:52
<
adisbladis >
Force of habit
18:53
<
__monty__ >
Yeah but opus is newer, so obviously you have to shun vorbis.
18:54
<
infinisil >
ALright I fixed my script
18:54
<
infinisil >
Had replaygain tags on the input file
18:55
<
infinisil >
Alright, gonna test 16, 32, 64 and 128 kbps opus vs flac now
18:55
<
__monty__ >
Pastebin it so we can critique the approach if the results aren't what we want them to be >: )
18:55
<
adisbladis >
infinisil: I'm curious what gear you have :)
18:56
<
adisbladis >
Fellow nixos audiophiles unite!
18:56
<
__monty__ >
Also wonder why you don't throw in aac for example?
18:56
<
infinisil >
Hehe my setup is horrible, I'm just using my gaming headset to listen
18:56
<
infinisil >
__monty__: I don't really care about aac :P, I think opus is the future
18:56
<
__monty__ >
And you can
*still* hear a difference? Are you randomizing?
18:57
<
adisbladis >
infinisil: At higher bitrates it's highly gear dependent if I can hear any difference
18:57
<
adisbladis >
infinisil: The quiz I sent you I can only get full score on my IEMs & one of my open backs
18:57
<
adisbladis >
A lesser pair of open backs are not revealing enough
18:57
* samueldr
encodes at 56kbps to steam from a boopy-doop-beep modem
18:57
<
infinisil >
Call it with a .flac file as an argument
18:57
<
samueldr >
to stream*
18:58
<
gchristensen >
how about those audiophile-grade SATA cables
18:58
<
adisbladis >
gchristensen: Aww yiss
18:59
<
adisbladis >
gchristensen: Hmm, there must be an untapped market for audiophile grade nvme
18:59
<
__monty__ >
samueldr: Would that work with VBR codecs though?
18:59
<
samueldr >
no idea, haven't researched the subject matter enough for a quick quip on IRC
18:59
<
infinisil >
Wait remove the --pure from the script, it can't connect to pulseaudio if you don't
19:00
<
drakonis__ >
is this satire
19:00
<
samueldr >
no, it's SATA-re
19:01
<
adisbladis >
Drakonis__: Welcome to the world of high end audio
19:01
<
adisbladis >
Where nothing makes sense
19:01
<
__monty__ >
When are we getting ECC cables though?
19:01
<
drakonis__ >
250 euros
19:01
<
samueldr >
ethernet cables with an arrow for the right data flow direction
19:01
<
adisbladis >
samueldr: You must turn it off before unplugging so the bits dont fall out
19:01
<
samueldr >
don't forget the split USB cables, so data and power do not travel parallel
19:01
<
adisbladis >
samueldr: Is that a thing?
19:02
<
gchristensen >
it is not satire, but it is a joke
19:02
<
drakonis__ >
there are people that actually bought this
19:02
<
samueldr >
audiophile gear is...
19:03
<
samueldr >
it just is
19:03
<
__monty__ >
Airgap your USB cable : )
19:03
<
gchristensen >
USB data-diode maybe
19:03
<
adisbladis >
samueldr: Suddenly I feel like a bad audiophile :)
19:03
<
samueldr >
oh oh oh
19:03
<
tilpner >
gchristensen: Did you not see my pings, or did you see them exactly when I was offline (I was offline lots today)?
19:03
<
samueldr >
you know
19:03
<
samueldr >
gold plated
19:03
<
gchristensen >
tilpner: I saw your pings and asked {^_^} to ping you back :P
19:03
<
adisbladis >
USB chain is: Computer -> monitor (hub) -> soundcard
19:04
<
adisbladis >
How can I live like this?!
19:04
<
samueldr >
adisbladis: you're gonna get images bleeding all over your sound
19:04
<
adisbladis >
How to throw more money at the problem
19:04
<
samueldr >
(though those gold plated cables are not audiophile cables... since they're cheap)
19:08
<
__monty__ >
infinisil: Rather than picking two random elements and comparing why not files[$i] and files[1-$i]?
19:08
<
infinisil >
__monty__: Yeah that's an idea
19:09
<
infinisil >
Though it's also interesting to know whether you can detect bigger differences
19:10
<
__monty__ >
Handcrafted using the mix of purest silver & gold alloy with silver copper conductors in special geometry
19:10
<
__monty__ >
Ah, yes, that ArCu metal. They used that in Avengers, right? Sounds legit.
19:10
<
drakonis__ >
audiophile woo
19:11
<
__monty__ >
infinisil: Ah, thought you were doing just two at a time.
19:12
<
__monty__ >
What have your results been? Is it just low - flac?
19:12
<
__monty__ >
Or are there surprises like 16 is better than 90?
19:15
worldofpeace has joined #nixos-chat
19:16
<
infinisil >
__monty__: I did once have a 64kpgs > flac
19:16
<
infinisil >
But often I can also detect flac > 128kpgs
19:17
<
infinisil >
Gotta do some more tistng
19:17
<
infinisil >
testing
19:17
<
__monty__ >
I wonder whether enc/dec is lossy twice?
19:18
<
infinisil >
decoding opus to wav shouldn't be lossy
19:18
<
__monty__ >
Sure, *shouldn't*.
19:19
<
drakonis__ >
decoding flac shouldnt be lossy either
19:22
<
infinisil >
This is hard..
19:22
<
infinisil >
Trying with bitrates 64 96 128 160 now
19:28
<
infinisil >
Two times in a row it showed me 96 vs 128. One time I said the former is better, another time the latter -.-
19:28
<
infinisil >
(well that's the point of ab testing I guess, but still!)
19:28
worldofpeace has quit [Quit: worldofpeace]
19:29
<
__monty__ >
You should add a third option "idk" and test AA so you can see whether you're guessing a lot.
19:30
<
infinisil >
I do have that option, but I just never feel the need to use it
19:30
<
infinisil >
I can always find a tiny thing that seems better in one of them
19:30
<
infinisil >
So I think I'll just to a whole bunch of tests, then look at the probabilities
19:30
<
__monty__ >
Then do the AA tests so you can see whether that's really hearing a quality diffence or just bias.
19:31
<
infinisil >
How would that work?
19:32
<
infinisil >
I mean with an AA test I know I'll choose either of them 50% of the time
19:32
<
__monty__ >
That's why you need that third option.
19:32
<
__monty__ >
If you don't use it you're unreliable.
19:32
<
__monty__ >
If you use it a
*lot* you don't hear a significant difference.
19:32
<
infinisil >
We're humans that's what we do!
19:33
<
infinisil >
I can just look at the probabilities of the AB tests, if I have 50% for two samples, I can't distinguish them
19:34
<
infinisil >
The further away from 50%, the more distinguishable
19:34
<
__monty__ >
Sure but you need a lot more samples.
19:34
<
__monty__ >
Maybe you could come up with a score, where you take into account the probabilities and the size of the files.
19:35
<
__monty__ >
Definitely needs some hypothesis tests : )
19:48
<
infinisil >
__monty__: Gotta randomize the bitrates themselves!
19:49
<
infinisil >
Or not randomize
19:49
<
infinisil >
But like, smartly select bitrates to find exactly the point at which it becomes indistinguishable
19:49
<
infinisil >
Something that would be called probabilistic binary search
19:50
<
__monty__ >
We need game theory up in here!
19:51
<
__monty__ >
Have the computer trie to "shortchange" you. It wants to give you the lowest bitrate it can get away with : )
19:51
<
infinisil >
Hehe yes
19:51
<
infinisil >
Every day it reduces bitrate by 1kbps until you notice
19:53
<
infinisil >
But maybe for deciding which is better, it should be like a slider, where you can say how much better one is than the other
19:53
<
infinisil >
I guess this also solves my "but I feel like this one is a tiny bit better"
19:54
<
infinisil >
Or maybe implement a timeout too: If you haven't decided within 30 seconds, they feel the same
19:54
<
infinisil >
That would be an interesting project tbh
19:55
<
__monty__ >
Oo, you should get it to switch back and forth within a song.
19:56
<
__monty__ >
If you don't notice it's switching you can't tell the difference.
19:57
drakonis_ has quit [Read error: Connection reset by peer]
19:59
drakonis_ has joined #nixos-chat
20:02
<
infinisil >
__monty__: Yeahh
20:02
<
infinisil >
Didn't know how to do that with mpv unfortunately
20:03
<
aanderse >
i joined office hours meeting but it hasn't started... just me?
20:07
<
drakonis__ >
no meeting apparently
20:19
psyanticy has quit [Quit: Connection closed for inactivity]
20:30
drakonis has joined #nixos-chat
20:33
drakonis_ has quit [Ping timeout: 240 seconds]
20:35
<
gchristensen >
ouch
20:36
<
infinisil >
LnL: How about using the patched nixpkgs directly/
20:37
<
infinisil >
(Can't do that in configuration.nix though)
20:39
<
LnL >
that's what I meant, I think it would be really nice if nixpkgs could be customized in some way
20:39
<
LnL >
not necessarily from configuration.nix but rather nix-channel (ie. flakes?)
20:41
<
infinisil >
Oh yeah I think this was one of flakes goals actually
20:42
<
LnL >
I've not been following it closely enough to know if this would fit in
21:14
MichaelRaskin has joined #nixos-chat
21:39
<
eyJhb >
Can anybody remember the word for when you combine everything at once, and pray that it works, instead of testing each individual module ? (software development)
21:39
<
eyJhb >
I remember there is SOME word for it, but I am not sure what it is
21:43
<
samueldr >
holistic?
21:44
<
samueldr >
>> characterized by comprehension of the parts of something as intimately interconnected and explicable only by reference to the whole.
21:49
<
MichaelRaskin >
monolith?
21:50
<
eyJhb >
Might just be my professor at uni that called it something like, chaotic testing, etc.
22:11
obadz has quit [Ping timeout: 245 seconds]
22:22
<
__monty__ >
I'd call it the Yenga model of construction. Build something, poke at it, layer things on top for as long as you can balance it.
22:26
waleee-cl has quit [Quit: Connection closed for inactivity]
22:27
MichaelRaskin has quit [Ping timeout: 252 seconds]
22:31
__monty__ has quit [Quit: leaving]
23:11
<
clever >
aleph-: main nixops issue is trying to deploy over 100 machines at once
23:18
<
aleph- >
Nod, over SSH right? So slow?
23:29
<
gchristensen >
it tries to evaluate nixos 100 times, requiring gobs of ram
23:55
Synthetica has quit [Quit: Connection closed for inactivity]
23:55
<
tazjin >
it's part of my monorepo-with-nix tooling investigation
23:56
<
gchristensen >
oh cool, tazjin
23:56
<
{^_^} >
tazjin/depot#9 (by tazjin, 21 minutes ago, merged): Configure automatic package layouts via repository structure