Default length of AFP sample is 0.5? #929

Closed
Sti2nd opened this Issue Jul 4, 2014 · 17 comments

Projects

None yet

6 participants

@Sti2nd
Contributor
Sti2nd commented Jul 4, 2014

Sometimes when I open samples with AFP, the length is not set to 1, it is set to 0.5 - the half length of the sample.
This happened right now after installing 1.0.3, and have happened before when I moved projects to a new location and had to relocate all drum samples.
So this is actually two bugs now:

  1. Default length of AFP endpoint knob is 0.5, it should be 1.0. But are these percentage at all?
  2. When the location of a sample changes, AFP don't remember the length, but resets to default.
@Sti2nd Sti2nd changed the title from BUG: Default lenght of AFP sample is 0.5 to BUG: Default lenght of AFP sample is 0.5? Jul 4, 2014
@musikBear
  1. Default length of AFP endpoint knob is 0.5

confirm for xp 32 1.0.3

And it should be added that if a longer sample (>1,5 secd) is loaded, the END knob wont cover whole sample-envelope. A part cant be reached. And if a short sample is replaced by a longer, the last part of this wont be in scope, either
Further more did 103 crash, as i moved the sample inside the eronious envelope
(Dont know if its feasable, but 0.9.1 AFP 'works*' (in respect to envelopes), so perhaps that code could be used easy-peasy?)
*Do take a look on #927 -something is odd

@musikBear

Pictures of this bug
Here first a long sample (13 secd, not shown) was loaded. It loaded fine and played correctly
Then a sample of 0.92 secd was loaded
http://snag.gy/RKJBH.jpg
The long sample is still shown as loaded, but the short sample played
StartPoint is shown as tool-tip <= 0
After pressing 'R' The short sample apeared
http://snag.gy/oBWDb.jpg
since the start-point is asumed to be '0', the start diel is locked, and it is not possible to get more of the sound inside the start-end area.
Turning 'LoopPoint' shows values from 0 to .999, again lmms asumes that the whole span of the sound is inside the start-end area.
But it is not the start-end area that lmms plays if a key is pressed. It sounds af most of the sound is played. If looped, the part sounds to the the correct relative part, controlled with the loop-diel (only moves inside start-end area.)
The bug seams to show a different span, that the one that is played.

@diizy
Contributor
diizy commented Jul 13, 2014

Does this happen on the 1.1 branch?

@musikBear

10.92 sorry should have said that
I actually think that all calculation of how the UI is displayed, is done from the very first sample loaded, and all loaded after the first are ignored in calculation.

@diizy diizy added this to the 1.1.0 milestone Jul 13, 2014
@Sti2nd
Contributor
Sti2nd commented Jul 17, 2014

@diizy Probably, cause it has happened from I started to use LMMS, so unless the AFP code has been tempered with since 1.0.0 (I can't remember seing any ticket on that), it should be reproducable. Try to install LMMS on a different location, and then open a project with samples. Naturally the samples will not play until you reconnect them, and then the end knob "resets" to 0.5

@diizy
Contributor
diizy commented Jul 18, 2014

Ok I can't reproduce the 0.5 length knob.

I also can't reproduce any issue with first loading a short sample and then loading a long sample - seems to work just fine here.

I can reproduce an issue when first loading a longer sample and then a shorter one: the end part of the old sample stays visible in the wavegraph, but if you zoom in it'll go back to normal and won't zoom out again to show the old sample. This seems purely a visual bug and doesn't affect playback.

If there's any issue in the 1.1 branch other than this, please list proper steps to reproduce. And remember that bugs in 1.0.3 are useless to report if those bugs do not happen in 1.1, because the 1.0 branch is EOL and 1.1 is the next release.

@Sti2nd
Contributor
Sti2nd commented Jul 21, 2014

I just reproduced the 0.5 length knob with Tres' build for LMMS 1.0.3

@diizy
Contributor
diizy commented Aug 6, 2014

Ok I still can't figure out this "0.5" thing.

Can someone post exact steps to replicate in 1.1?

@musikBear

any 1.1 win32 exe? -then i will

@Sti2nd
Contributor
Sti2nd commented Aug 7, 2014

I just reproduced the problem "first loading a short sample and then loading a long sample" on 1.1. Guess you did that yourself, too?

I couldn't reproduce the AFP 0.5 by just moving the sample, guess I will have to install LMMS on another location, as that is what have triggered it before. I am moving tomorrow, so we'll se if I ever get to it.

Tres has latest builds, @musikBear . Don't remember where I found them, but they are somewhere

@musikBear

@Sti2nd -oki Thanks! I will ask him.

@Lukas-W Lukas-W added the bug label Aug 23, 2014
@Lukas-W Lukas-W changed the title from BUG: Default lenght of AFP sample is 0.5? to Default length of AFP sample is 0.5? Aug 23, 2014
@musikBear

1.093 win x32 xp (1.1 PR 04)
I have tried to replicate the bug, but it gave problems.
It is not consistent 0.5, but it is still restricted to a limited span
Method:
from folder 'effects' load explode01
make a selection - I used 0.14 .. 0.27 ~ 458 msek

Drag scratch01 (same folder) in
Startpoint can now go from 0..
Endpoint can go to '1' , but that is in a faulty pos:
http://snag.gy/BCnDB.jpg
Can you replicate?

(btw - if you tackle this diiz, how about 'parking' the loop pos default between START and END -I think it will be more intuitive, and not 'block' scrolling END ?)

@diizy
Contributor
diizy commented Aug 23, 2014

On 08/23/2014 08:20 PM, musikBear wrote:

1.093 win x32 xp (1.1 PR 04)
I have tried to replicate the bug, but it gave problems.
It is not consistent 0.5, but it is still restricted to a limited span
Method:
from folder 'effects' load explode01
make a selection - I used 0.14 .. 0.27 ~ 458 msek

Drag scratch01 (same folder) in
Startpoint can now go from 0..
Endpoint can go to '1' , but that is in a faulty pos:
http://snag.gy/BCnDB.jpg
Can you replicate?

Yeah, like I said: that's just a visual bug, it shows the tail side of
the earlier sample in the graph. It doesn't actually affect playback,
and if you zoom the graphview in, the visual bug disappears too.

@Sti2nd
Contributor
Sti2nd commented Aug 26, 2014

We can skip the 0.5 bug for now, it only happens when you install LMMS on a different location or with a different name in the folder, and then have to reopen the AFP samples. Annoying as hell, but most people want to install on the same location so they don't have to relocate samples...

@tresf
Member
tresf commented Dec 23, 2014

@curlymorphic I spent quite a bit of time trying to fix this one but just couldn't wrap my head around it... basically if you load a shorter sample in AFP, it messes up the GUI. Closing and reopening the instrument window resolved the problem, but I couldn't find a way to reproduce the steps of reloading of the window.

P.S. Bumping to 1.2 unless we determine it's a small fix that can make the cut for 1.1.

@tresf tresf modified the milestone: 1.2.0, 1.1.0 Dec 23, 2014
@curlymorphic
Contributor

Am looking at this now :)

@tresf
Member
tresf commented Jan 15, 2015

Closed via #1491.

@tresf tresf closed this Jan 15, 2015
@tresf tresf modified the milestone: 1.1.0, 1.2.0 Jan 15, 2015
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment