Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2023-08-14

00:06:37 Quit CH23_M (Read error: Connection reset by peer)
00:07:16 Join CH23_M [0] (~CH23@revspace/participant/ch23)
00:27:51Nyaanice, it worked
00:28:19Nyaai made the piezo beeps match headphone beeps lol
00:29:24Nyaaexcept the maths in the existing function for calculating the beep frequency on ipodvideo seem wrong and they're slightly higher pitched than they should be
00:33:41*Nyaa changes the 91225 to a 125000 just out of a hunch and sees if that fixes it
00:37:16 Quit m01 (Quit: Konversation terminated.)
00:39:35Nyaacloser but still off lol, so not 1mhz/8 for piezo clock
00:39:57 Join m01 [0] (
00:59:08Nyaatime to just set it to /10 and measure the frequency lol
01:04:46 Quit braewoods__ (Ping timeout: 256 seconds)
01:24:23 Join braewoods [0] (~braewoods@user/braewoods)
01:29:39Nyaaah ha, piezo appears to be running at 1.536mhz
01:35:47Nyaa[or maybe 49.152mhz idk i can't be bothered to maths the source clock, it's a multiple of 192khz]
01:43:11Nyaaer wait 192khz is the half-period time at 50% duty cycle so it's a multiple of 96khz oops lol forgot i was measuring a square wave
01:47:21*Nyaa is confused because it sounds too high pitched again set to 96khz and sets it back to 192khz
01:52:29***Saving seen data "./dancer.seen"
02:00:02 Quit tchan (Read error: Connection reset by peer)
02:27:51*Nyaa going through other targets to adapt new piezo code
02:34:15Nyaadoes anyone here know what kind of hardware generates the "click" on the philips targets
02:38:25sporka tiny man with a hammer
02:38:53Nyaai'm asking so i don't end up writing code that could damage the hardware
02:42:57 Quit _bilgus_ (Remote host closed the connection)
02:44:34*Nyaa shrugs, decides she should just adjust the code to handle new changes as-is with no functionality change on that target and moves onto the others for now
02:45:12 Join _bilgus_ [0] (~bilgus@
02:54:53sporki wonder if there are many of those philips targets still actively used
02:55:24Nyaano idea, if anyone around here has one and wants to to try out code that might kill the click/feedback hardware let me know lol
02:55:34sporknice incentive
02:55:56Nyaaif it works, the upside is that it will beep instead of just clicking :p
03:09:51Nyaaanyway does anyone here have other ipod models they can test my changes on? works well on the ipodvideo i have but idk if i got the frequency/durations right for the 6g and nano2g [plus it might be the case that the 4g, colour, mini1g, and mini2g use a different piezo clock]
03:28:10 Join PheralSparky [0] (~Shawn@user/shawn/x-4432647)
03:42:26*Nyaa notices a slight unintentional behaviour change with her code but leaves it for the time being until getting feedback on if it's even worth putting the extra effort into it
03:52:30***Saving seen data "./dancer.seen"
04:02:11Nyaahere's my modified version if anyone wants to try it,
04:10:29Nyaai suspect beeps/clicks on ipodnano2g and ipod6g are going to be an octave higher than they should be and ~4x shorter than they should be
04:12:08 Quit PheralSparky (Read error: Connection reset by peer)
04:19:31Nyaahmm oh
04:19:39Nyaai get how that timer works now
04:19:41*Nyaa fixes it
05:24:16Nyaa[if anyone wanted to test it, now it has the fixed code lol]
05:24:41Nyaastill might be an octave higher than it should be but duration should be correct
05:30:35Nyaawait i see a bug in my maths lol
05:38:58 Quit CH23_M (Read error: Connection reset by peer)
05:43:15 Join CH23_M [0] (~CH23@revspace/participant/ch23)
05:51:46Nyaaok _now_ it should be testable lol
05:52:34***Saving seen data "./dancer.seen"
05:57:42 Quit CH23_M (Read error: Connection reset by peer)
05:58:05 Join CH23_M [0] (~CH23@revspace/participant/ch23)
07:52:36***Saving seen data "./dancer.seen"
09:04:04 Join munkis [0] (~munkis@
09:04:27 Nick munkis is now known as Guest699 (~munkis@
09:05:11 Nick Guest699 is now known as munkis (~munkis@
09:05:33munkisDo we host simulator builds?
09:33:33 Join tchan [0] (
09:52:40***No seen item changed, no save performed.
10:12:36 Quit berber_l5 (Ping timeout: 246 seconds)
10:49:52 Quit munkis (Quit: Client closed)
11:33:40 Join munkis [0] (~munkis@
11:34:03 Nick munkis is now known as Guest1400 (~munkis@
11:34:57 Nick Guest1400 is now known as munkis (~munkis@
11:34:58munkisNVM found rashers sims from 2018.
11:52:44***Saving seen data "./dancer.seen"
12:12:49 Quit munkis (Quit: Client closed)
12:16:18 Join lebellium [0] (
13:52:46***Saving seen data "./dancer.seen"
14:51:20Nyaanice, made a build for myself with the older tinysid engine lol
14:52:09Nyaamaybe i could make a patch that builds it instead of crsid for targets without enough hardware specs for crsid
15:52:50***No seen item changed, no save performed.
17:52:51***No seen item changed, no save performed.
17:58:48 Quit CH23_M (Read error: Connection reset by peer)
18:00:11 Join CH23_M [0] (~CH23@revspace/participant/ch23)
18:17:19 Quit lebellium (Quit: Leaving)
19:06:09 Join massiveH [0] (~massiveH@2600:4040:a982:c800:b8a8:aa7c:24d0:3961)
19:32:28 Quit kugel___ (Ping timeout: 246 seconds)
19:34:12 Join kugel_ [0] (
19:52:52***Saving seen data "./dancer.seen"
20:13:06 Join speachy [0] (
20:13:07 Quit speachy (Changing host)
20:13:07 Join speachy [0] (~speachy@rockbox/developer/speachy)
20:13:07Mode"#rockbox +v speachy" by ChanServ (
20:14:37speachyNyaa: the PP-based ipods and various m68k targets are probably the only ones we have that are too slow to handle crSID.
20:15:08Nyaathere's the issue with ram too
20:15:25Nyaai don't think there would be a build flag for not building it on small ram targets otherwise
20:15:49speachythe tinyRAM targets still have a couple hundred mhz arm9s to play with IIRC
20:16:11speachyit's the sub-100MHz ARM7TDMI stuff I'd be concerned about
20:17:20Nyaatbf that multimedia chip runs at 150mhz apparently :p
20:17:38Nyaai still haven't gotten around to emailing broadcom yet though lol
20:17:40speachymight be able to just say if (! (ARCH == m68k || (ARCH == arm && ARCH_VERSION < 5)))
20:18:20Nyaaeh, i'm looking at potentially making it faster
20:18:47Nyaaare there tools for doing code profiling on rockbox
20:19:18Nyaaoh it's in the advance configure options
20:22:05speachyhmm, you might want to try the codec_benchmark tool on those SID files, to see just how underpowered the ipod is
20:22:57Nyaai mean, lower cpu usage would improve all targets by lowering power usage so optimising it if possible would be good either way
20:23:29speachyof course
20:23:55speachythat would give you some hard numbers to compare against regardless
20:24:25Nyaaalso "those SID files" is the entire HVSC so lol
20:24:39speachyhmm, more arm7tdmi targets than I remeembered but they seem to max out at 75MHz
20:24:41Nyaai haven't tried 2sid or 3sid ones yet but i figure they'll be even worse
20:26:17Nyaafails to build with profiling on
20:26:18speachyoptimizing the crSID emulator is likely to be a bit of open-ended non-trivial work; a more short term solution IMO would be to make tinysid vs crsid a build-time option
20:26:40Nyaalemme make sure it's not my toolchain breaking it lol
20:27:20Nyaanope not related to that
20:28:06Nyaaeh i will do the benchmark instead for now and figure out compiling profiling in later
20:29:09speachyI think the codec benchmark plugin is gated behind the debug plugins configure option
20:29:22Nyaaplus with benchmark i can compare gcc 4.9 and 13.2 and see if that makes any difference too
20:30:42speachymunkis: And yes, there's no automatic repository of sim builds. That could probably be changed but I don't know how generally useful they'd be to make the bandwidth/disk usage worthwhile.
20:31:39Nyaawhy do builds get done every day even if no code changes happened
20:31:43speachy(IMO only the windows builds are semi-useful but last time I tried to cross-compile one it ended very bnadly)
20:31:46Nyaathat alone seems like a waste
20:32:09speachythey don't. the nightly build script just collects the most recent per-commit builds.
20:32:27speachyit does, however, regenerate the voices and manuals, which is a bit of a waste.
20:32:43Nyaaoh, well i did open the website once while it was doing that and got confused for a second at like 80% of the links being missing lol
20:33:42Nyaaalright got the benchmark plugin compiled, though i do have changes to crsid i was making so i'll have to run it again after with those changes undone lol
20:33:53speachythis is the largest devleopment lull in a couple of years FWIW.
20:34:54Nyaaoh also idk if anyone has any of the other ipod targets and wants to test my [mostly proof of concept] piezo beep changes
20:35:12speachyI thought only the "fat" ipods had piezos?
20:35:48Nyaanano 1g/2g, mini 1g/2g, standard 4g/5g/6g
20:35:58Nyaaat least according to rockbox code
20:36:15Nyaai think 1g/2g/3g also have something but maybe not supported in rockbox idk
20:36:52speachyI have a 5.5g and a mini2g in my rockboxbox
20:37:04Nyaathe 5/5.5g is what i'm using now lol
20:37:12speachy(^2 ?)
20:37:32Nyaait seems to work fine here but i'm not 100% sure i measured the frequency right, still sounds like funny harmonics but that might just be square wave shenanigans
20:38:01Nyaachanges i made cause the piezo to make all the interface beep sounds instead of just keyclicks
20:38:48speachyheh, fwiw the fat 1g-3g are going to be _way_ too slow for SID if the 5g barfs on 'em, as they have the older PP that effectively runs at half the clock freq
20:39:10Nyaai think mini 2g uses the same code that the 5.5g uses so it will probably work the same
20:39:27speachymini2g is the same as the 4g grey IIRC
20:39:32Nyaait's the 6g and nano 2g that use different piezo code and i'm very unsure about that
20:40:08Nyaaall i know is that the maths seemed to check out and that it compiled lol
20:40:26speachynano1g is PP, nano2g is samsung (along with 6g)
20:41:08Nyaatargets i have i can test against are the uh, cowon d2 and ipod 5/5.5g
20:41:18Nyaai might have a sansa clip motherboard in a box somewhere but no other parts of it lol
20:41:59speachyI don't recall if anyone with a Cowon ever confirmed that the newer-toolchain builds even worked.
20:43:51Nyaatime to run codec text
20:43:54speachyheh, the D2 is 48MHz.
20:44:06Nyaais that base or after boost
20:45:09Nyaaspeed test running
20:45:32speachyah, 192 max. that's respectable.
20:46:34Nyaaoh i also have an ipod shuffle 2g that i've been meaning to try to do shenanigans with
20:46:54Nyaawould need to find a way to execute arbitrary code on it first
20:47:32Nyaaand i doubt the specs are anywhere good enough to run rockbox on it even if i figure that out lol
20:49:49speachya voice-only UI would be an interesting endeavour
20:51:34Nyaahmm it has a STMP3550
20:52:35Nyaai guess it depends on if it has enough SDRAM attached or not lol
20:57:17Nyaatbh the most interesting thing would be to find a way to inject an opus decoder into the shuffle 2g to be able to fit more/higher quality music on it lol
20:57:34speachymight be a combined SDRAM+flash chip
20:57:48speachybut there's not a whole lot of flash on there regardless
20:58:34Nyaayeah i think the one i have is a 1GB model
20:59:41Nyaahonestly the only reason i have it was because it's purple lol
21:00:02Nyaait was going to get thrown out [it had never even been opened lmao] but i thought it looked pretty
21:00:57speachyyay, another iFashionAccessory
21:01:59Nyaawell i also don't like stuff becoming e-waste, especially if it's completely functional
21:02:19Nyaathis ipod 5g only had like 40 hours of use before i got it
21:02:40Nyaai think the battery degradation it has was from being stored at 0% charge
21:03:08Nyaa[well, 0% to the firmware, it cuts off at like 10% actual charge]
21:03:18speachythe hdd-based (or external card-based) players are the ones holding up
21:03:36Nyaathe hdd has a couple bad sectors but i'm hoping they're just factory
21:04:05Nyaaoh also i also have a half-written change to add SMART status reporting to the ipod 5/5.5g lol
21:04:21Nyaathe diagnostics menu shows the SMART status, not 100% sure how to fetch it otherwise though
21:05:12Nyaaalso i don't think the benchmark tool works with SID playback
21:05:21Nyaabecause i think it will play back the SID indefinitely
21:05:32Nyaait just says "0 of 0"
21:06:34Nyaalet me test with an aac file
21:07:02Nyaajust to confirm the benchmark tool works lol
21:07:38Nyaa240% realtime on the aac file i tested
21:08:34Nyaalemme see if it works with a gbs file
21:10:06Nyaahmmm, seems to but i think the results are gonna be wrong because of the way the multitrack files report position
21:15:23Nyaayeah, the benchmark tool does not work on multitrack files because each track is reported as 1 second
21:20:58Nyaamaybe they could be modified to report each track as if it was cuesheet info
21:50:52 Quit massiveH (Quit: Leaving)
21:52:54***Saving seen data "./dancer.seen"
23:05:00 Quit speachy (Quit: WeeChat 3.6)
23:52:56***Saving seen data "./dancer.seen"

Previous day | Next day