Lmms: Glitchy Sound in FreeBoy

[popsip]

Something is wrong with FreeBoy's sound when played under Linux version of LMMS. It makes a glitchy/static like sound when played.

I played the same notes with the same FreeBoy settings on both the Linux and Windows version of LMMS.

Linux: http://streams19.tunescoop.com/266702_IncorrectFreeboySound_stream.m4a
Windows: http://streams19.tunescoop.com/266703_CorrectFreeboySound_stream.m4a

• This issue also occurs in Linux version of LMMS 0.4.15

[eagles051387]

I have seen this with just about everything on linux (kubuntu 13.10 64bit).
What i find odd switching to the higher latency pulse audio backend in lmms
and getting off alsa seems to solve the problem.

On Mon, Feb 17, 2014 at 8:20 AM, popsip [email protected] wrote:

Something is wrong with FreeBoy's sound when played under Linux version of
LMMS. It makes a glitchy/static like sound when played.

I played the same notes with the same FreeBoy settings on both the Linux
and Windows version of LMMS.

Linux:
http://streams19.tunescoop.com/266702_IncorrectFreeboySound_stream.m4a
Windows:
http://streams19.tunescoop.com/266703_CorrectFreeboySound_stream.m4a

• This issue also occurs in Linux version of LMMS 0.4.15

Reply to this email directly or view it on GitHubhttps://github.com/LMMS/lmms/issues/326
.

Jonathan Aquilina

[diizy]

Again when speaking of ALSA: were you using actual ALSA or PA masquerading as ALSA?

[eagles051387]

What ever the default is for kubuntu, most likely pulse audio, but im not
sure if they do anything screwey

On Mon, Feb 17, 2014 at 3:32 PM, Vesa V [email protected] wrote:

Again when speaking of ALSA: were you using actual ALSA or PA masquerading
as ALSA?

Reply to this email directly or view it on GitHubhttps://github.com/LMMS/lmms/issues/326#issuecomment-35286357
.

Jonathan Aquilina

[diizy]

You need to set the "device" setting in the LMMS backend settings in order to use actual ALSA.

[eagles051387]

Thats how it is by default and i was getting nothing but static on
everything even the demo songs. would change it to the much higher latency
pulse audio from with in lmms and that solved the problems.

On Mon, Feb 17, 2014 at 5:58 PM, Vesa V [email protected] wrote:

You need to set the "device" setting in the LMMS backend settings in order
to use actual ALSA.

Reply to this email directly or view it on GitHubhttps://github.com/LMMS/lmms/issues/326#issuecomment-35302324
.

Jonathan Aquilina

[diizy]

On 02/17/2014 07:00 PM, eagles051387 wrote:

Thats how it is by default and i was getting nothing but static on
everything even the demo songs. would change it to the much higher latency
pulse audio from with in lmms and that solved the problems.

No, the "device" textbox on the backend selection, it's not set by
default. When you first select ALSA, then you have to enter your device
name into the device textbox. It's empty by default, or it's set to
"default", but this causes the ALSA output to be "intercepted" by PA.
You have to type your actual sound device filename there so that you can
use ALSA directly.

[pgiblock]

If switching to a higher latency setting fixes this, then I believe the problem is a race condition. You reduce the chance of data inconsistencies by having less periods per second. This is more a flaw of the way LMMS works as far as I can see. There should really be separate "models" for the DSP code than the model as seen in the GUI frontend.

[softrabbit]

Looks interesting; sometimes the output goes to 0, sometimes it gets half-wave rectified to one side and sometimes to the other. @popsip, does the noise end if you turn channel 4 (in Freeboy) off?

[popsip]

@softrabbit, no, it does not end the glitchy noise. The examples that I provided had only channel two enabled.
@eagles051387, Switching to PulseAudio does not help.
@all, the only channel that I have enabled in FreeBoy when I provided the examples was channel two (which is supposes to produce a pure square wave when a note is playing). Everything else other than FreeBoy plays completely fine; therefore, the problem does not have anything to do with PulseAudio nor ALSA.

[eagles051387]

I am still finishing up setting things up on my now gentoo system I will
hopefully have better results as I am already seeing overall improvements
with it in terms of system as a whole. I think from what i can gather so
far my gentoo setup is using ALSA, but would need to get lmms built to do
any further testing.

On Thu, Feb 20, 2014 at 11:04 PM, popsip [email protected] wrote:

@softrabbit https://github.com/softrabbit, no, it does not end the
glitchy noise. The examples that I provided had only channel two enabled.
@eagles051387 https://github.com/eagles051387, Switching to PulseAudio
does not help.
@all https://github.com/all, the only channels that I have enabled in
FreeBoy when I was providing the examples is channel two (which is supposes
to produce a pure square wave when a note is playing). Everything else
other than FreeBoy plays completely fine; therefore, the problem does not
have anything to do with PulseAudio nor ALSA.

Reply to this email directly or view it on GitHubhttps://github.com/LMMS/lmms/issues/326#issuecomment-35674843
.

Jonathan Aquilina

[tobydox]

According to Valgrind there are also problems with invalid memory accesses within Gb_Apu - possibly causing the strange noise on channel 4.

[zonkmachine]

There seem to be updates done to this code in the package included in Game_Music_Emu

The version in papu/gb_apu/Gb_Apu.cpp reads:
// Gb_Snd_Emu 0.1.4.
And Gb_Apu.cpp from Game_Music_Emu reads:
// Gb_Snd_Emu 0.1.5.

http://blargg.8bitalley.com/libs/audio.html#Game_Music_Emu
https://groups.google.com/forum/#!forum/blargg-sound-libs

Edit: Fork of Game_Music_Emu with later updates here:
https://github.com/kode54/Game_Music_Emu/tree/master/gme

[karmux]

I tried updating to newer files from https://github.com/kode54/Game_Music_Emu but the noise remained on channel 4.

[zonkmachine]

OK. I had another go at searching and this repo looks more like a proper upstream and with more recent commits.

[karmux]

Sound that I get from channel 4 is actually correct. It's meant to be noise: https://github.com/LMMS/lmms/blob/stable-1.2/plugins/papu/papu_instrument.cpp#L271

Updated Freeboy files in pull request #3618.