Ideas about crackling noise in LMMS

[AncientMind]

Hi altogether (especially musicbear, but as well someone of the development team),
first of all: This posting is about the value-changes (especially volume) in LMMS which happen to be not continuous (aka sample exactness) thus causing (more or less - for different persons - but always measurable) audible crackling noise.
For musicbear especially: You asked me for an example project producing this noise. I made two small 'projects'. Each one is built using a kicker track as kick, one kicker as HiHat (which is muted) and a bassline, which is made in one project using Monstro, in the other with TripleOscillator.
I sent to you, musicbear, these projects by e-mail (to your hotmail account), as well as 4 exported sounds coming from both projects (perhaps you want to hear exactly what I hear), which I transcoded to FLAC.
Perhaps you find a way to make the noise vanishing...?!?

Well, to the others: Why does LMMS not - as a first workaround - simply wait for a zero crossing of a signal for changing a volume (or better: the multiplication factor of a signal that passes a slider or - as well - an automation envelope) ? Even though there still remains a more or less intensive 'click', but it will for sure be much more quite than changing it in the maximum of a signals curve (as it happens until recently - and as far as I can see still in LMMS 1.1.3). I can explain you my idea more in detail, if necessary.

What's your opinion? BTW, this really should be only a workaround, not an all-day solution, of course!

[musikbear]

First off, there is indeed a crackle artefact, and it is from the physical movement of the slider-controller. This is easy to see here in the first part of
https://youtu.be/3bs4hQwpTzI
But i changed your sidechain, to not use a physical movement, and only change volume internaly, and in the 2. part that is used. That gives a crackle free sidechain. Ofcause it is then not a peak-controller-sidechain, but you can do a lot with the controllers.
I hope you like this output? I think its quite good. I will upload the file for all, on the LSP :)

[AncientMind]

Indeed, that's a really good result - thanks a lot . Of course there need to be no visible movement at all... why the heck is it not simply an option (which can be switched off!) to display a values' movement? Obviously the graphics engine either kills the whole timing or it simply takes too much processing time in combination with a much too high priority... hmmm, perhaps I can try to change the thread's priority (using Process Hacker that's no problem - presumed I know which thread's priorities I have to change).
This also should be a very useful approach to kill these damn noises completely... I'll come back later.

[Tycoonie259]

First off, there is indeed a crackle artefact, and it is from the physical movement of the slider-controller. This is easy to see here in the first part of
https://youtu.be/3bs4hQwpTzI
But i changed your sidechain, to not use a physical movement, and only change volume internaly, and in the 2. part that is used. That gives a crackle free sidechain. Ofcause it is then not a peak-controller-sidechain, but you can do a lot with the controllers.
I hope you like this output? I think its quite good. I will upload the file for all, on the LSP

I don't really get the project, how did you do it without using a physical movement? I'm finishing up a project and I really need to fix this!

[musikbear]

I uploaded it on LSP (lmms sharing platform)
https://lmms.io/lsp/?action=show&file=6849
Guess i should have added the link earlier - sorry

[Tycoonie259]

I uploaded it on LSP (lmms sharing platform)
https://lmms.io/lsp/?action=show&file=6849
Guess i should have added the link earlier - sorry

I don't really understand what you did in that project, could you please exactly explain?

Also, when will sample-exactness be implemented? As I can wait for about a month before I have to release this project.

[musikbear]

when will sample-exactness be implemented?

unknown :)

The file is the one in the orr. posters AncientMind problem-project (demo)
(He wanted the hats do duck the bass - i bet thats what confused you :)
Dont look at the kick, look at the hats :)

[AncientMind]

The file is the one in the orr. posters AncientMind problem-project (demo)
(He wanted the hats do duck the bass - i bet thats what confused you
Dont look at the kick, look at the hats

That's not correct, musikbear . I wanted to duck the bassline by using the kick as side chain signal. You knew this exactly and I ask myself: Is it really necessary to tell Tycoonie259 fairy tales (means in clear words: LIES) simply because you still want to avoid someone could suspect LMMS to be the error prone and only occasionally intensive developed application it indeed is since many years now?
The problem was and still is, that nearly every kind of automation in LMMS means to move a knob, a slider,... whatever. As you found out by yourself every such movement causes crackling noise and, to be honest, meanwhile I seriously doubt there will ever be a working solution. Long ago I wrote a review of LMMS in which I named LMMS a hobby project... you know what? It indeed is one and will always keep on being this hobby.
I go on working with Reaper (Cockos.com) and ... even in Linux there seems to be at least one very well equipped and made (BTW very professional looking) program called 'Ardour'. So in my eyes there is absolutely no need to wait any longer for a work to be done... what if it never is done...? Wait until the day of judgement? No, thank's! I won't live forever and want to create music r i g h t n o w. Perhaps somebody fumble around LMMS, perhaps not; perhaps he/they'll find solutions, perhaps not... who cares?

For all you who still search for excuses for LMMS... have a good time.

[Gps]

As Linux user you do have more choice then just lmms, I will agree on that.
I do not agree that every automation causes crackling, this does not happen for me on openSUSE nor on Windows 7 pro.

LMMS might not be perfect yet, but for a free program its pretty close to perfect, and has improved a lot over the years.

You want LMMS to progress faster ? Find a few programmers and a sponsor, and or find a way so the current programmers can work on this 24/7.
(minus time to sleep)

You’re entitled of course to you’re own opinion, but such a rant about a free program, made by people in their free time, is to me not done.

About Adour, have you noticed how that program is not free anymore ?

http://ardour.org/download.html

[musikbear]

The file is the one in the orr. posters AncientMind problem-project (demo)
(He wanted the hats do duck the bass - i bet thats what confused you :)
Dont look at the kick, look at the hats :)

That's not correct, musikbear :x . I wanted to duck the bassline by using the kick as side chain signal. You knew this exactly and I ask myself: Is it really necessary to tell Tycoonie259 fairy tales (means in clear words: LIES) simply because you still want to avoid someone could suspect LMMS to be the error prone and only occasionally intensive developed application it indeed is since many years now? :/ :/ :/
The problem was and still is, that nearly every kind of automation in LMMS means to move a knob, a slider,... whatever. As you found out by yourself every such movement causes crackling noise and, to be honest, meanwhile I seriously doubt there will ever be a working solution. Long ago I wrote a review of LMMS in which I named LMMS a hobby project... you know what? It indeed is one and will always keep on being this hobby.
I go on working with Reaper (Cockos.com) and ... even in Linux there seems to be at least one very well equipped and made (BTW very professional looking) program called 'Ardour'. So in my eyes there is absolutely no need to wait any longer for a work to be done... what if it never is done...? Wait until the day of judgement? No, thank's! I won't live forever and want to create music r i g h t n o w. Perhaps somebody fumble around LMMS, perhaps not; perhaps he/they'll find solutions, perhaps not... who cares?

For all you who still search for excuses for LMMS... have a good time.

wooo wait a secd - i certently dont 'try to find excuses for lmms'
I said

First off, there is indeed a crackle artefact, and it is from the physical movement of the slider-controller. This is easy to see here in the first part of

the file i made and link to duck the bass on hats, and with compression, not peak-control, and i gave it to you as a 'hack' to get the effect you wanted, without physical movement of any slider, and it 'works' - and you said

8-) Indeed, that's a really good result - thanks a lot :)

I kind of thought you were satisfied :p
I uploaded the file, but it was not clear for Tycoonie259 how i had done it, so i asked him to look at the hat-mixer-challel, where the controlls are in the file I uploaded
I must tell you that i am one of the harshest critics of lmms, and has made so many change- and bug-reports, that most 'hates' me :p
I definitely do not try to find 'excuses' for lmms' inferiority where there are some -And there ARE some! -especially in respect to samples, but those will be solved, and lmms will be even better. I wish you the best with your music made on Reaper. It is a great program, albeit technical and complicated, as are many high-line professional tools.
B.r mB