No audio output (SOLVED)
-
- KVRian
- Topic Starter
- 687 posts since 27 Dec, 2003
I'm pretty new to TX16Wx.
Version 2.3.1c in REAPER 4.62 (64bit)
-> load TX16Wx and put a sample on C3
-> insert MIDI item with some notes
=> playback = fine
-> duplicate the tack
-> playback
Result: no audio output on the duplicated track unless you open the GUI once
SOLVED
Version 2.3.1c in REAPER 4.62 (64bit)
-> load TX16Wx and put a sample on C3
-> insert MIDI item with some notes
=> playback = fine
-> duplicate the tack
-> playback
Result: no audio output on the duplicated track unless you open the GUI once
SOLVED
Last edited by Dandruff on Thu Jun 05, 2014 4:02 pm, edited 1 time in total.
-
- KVRAF
- 2398 posts since 27 May, 2005 from Stockholm
This seems to work fine for me. Are you sure this is not a reaper setup issue?
TX16Wx Software Sampler:
http://www.tx16wx.com/
http://www.tx16wx.com/
-
- KVRian
- Topic Starter
- 687 posts since 27 Dec, 2003
Don't know. Opening a saved projects also give me no output on all TX16Wx tracks. I have to open to GUI to get sound!
SOLVED
SOLVED
Last edited by Dandruff on Thu Jun 05, 2014 4:02 pm, edited 1 time in total.
-
- KVRian
- Topic Starter
- 687 posts since 27 Dec, 2003
With the same project (bassdrum sample) I also have overload problems. TX16Wx outputs high levels so that REAPER's track muting kicks in very often (set to >24dBFS getting muted). This Bassdrum (mono) sample peaks at -12.7dBFS so it's shouldn't overload on TX16Wx.
Weird.
Sample = 48kHz
Project (ASIO)= 48kHz
SOLVED
Weird.
Sample = 48kHz
Project (ASIO)= 48kHz
SOLVED
Last edited by Dandruff on Thu Jun 05, 2014 4:03 pm, edited 1 time in total.
-
- KVRian
- Topic Starter
- 687 posts since 27 Dec, 2003
Found the culprit for the overload problem: It's the Sinc128 resample mode! Setting it to Sinc64 fixes the problem. Obviously the Sinc128 does add some weird spike sometimes.
But why does the resample mode is active in this case at all?
48kHz sample + 48kHz output at root key shouldn't need resampling. Or I'm wrong?
But why does the resample mode is active in this case at all?
48kHz sample + 48kHz output at root key shouldn't need resampling. Or I'm wrong?
-
- KVRAF
- 2398 posts since 27 May, 2005 from Stockholm
Please send me the offending project + samples. That sounds weird.
TX16Wx Software Sampler:
http://www.tx16wx.com/
http://www.tx16wx.com/
-
- KVRAF
- 2398 posts since 27 May, 2005 from Stockholm
Also, interpolation is not set automatically. If its set to something, its because you changed it.
I've verified that Sinc128 works as expected on some test material, including 1:1 sample rate, non pitch-shifting playback. It is not just simply the fact that Sinc128 is _very_ CPU intensive? Since the render callback to the plug-in might take longer than the time slice rendered, you could get choppy audio / spikes just because of CPU load (do not look at full CPU load, you most likely need to look per core, since a single track will usually be rendered on a single core).
Sinc interpolation >32 should really only be used for offline rendering.
I've verified that Sinc128 works as expected on some test material, including 1:1 sample rate, non pitch-shifting playback. It is not just simply the fact that Sinc128 is _very_ CPU intensive? Since the render callback to the plug-in might take longer than the time slice rendered, you could get choppy audio / spikes just because of CPU load (do not look at full CPU load, you most likely need to look per core, since a single track will usually be rendered on a single core).
Sinc interpolation >32 should really only be used for offline rendering.
TX16Wx Software Sampler:
http://www.tx16wx.com/
http://www.tx16wx.com/
-
- KVRian
- Topic Starter
- 687 posts since 27 Dec, 2003
Ok, sent to info@cwitec.com.
-
- KVRian
- Topic Starter
- 687 posts since 27 Dec, 2003
elcallio wrote:Also, interpolation is not set automatically. If its set to something, its because you changed it.
I've verified that Sinc128 works as expected on some test material, including 1:1 sample rate, non pitch-shifting playback. It is not just simply the fact that Sinc128 is _very_ CPU intensive? Since the render callback to the plug-in might take longer than the time slice rendered, you could get choppy audio / spikes just because of CPU load (do not look at full CPU load, you most likely need to look per core, since a single track will usually be rendered on a single core).
Sinc interpolation >32 should really only be used for offline rendering.
Ok, I change the playback mode to Sinc32.
Any idea with the non-opened GUI = no playback on track duplicate and project load issue?
SOLVED
Last edited by Dandruff on Thu Jun 05, 2014 4:03 pm, edited 1 time in total.
-
- KVRAF
- 2398 posts since 27 May, 2005 from Stockholm
I checked the file you sent. You are running reaper with Settings->Plugins->Compatibility->Disable saving full plug-in state On. This will cause reaper to only save a "preset" (VST FXP) when storing the project. TX16Wx handles this less than perfect, and while (most) of the relevant data is stored, some is not, including performance assignment. Thus when loading, no performance is selected -> no sound. Opening the UI will force select the first (must show something), thus you get sound then.
While I should fix so that FXP saving is a bit more consistent, the fact is still that this is _NOT_ the option you want turned on in this case.
While I should fix so that FXP saving is a bit more consistent, the fact is still that this is _NOT_ the option you want turned on in this case.
TX16Wx Software Sampler:
http://www.tx16wx.com/
http://www.tx16wx.com/
-
- KVRian
- Topic Starter
- 687 posts since 27 Dec, 2003
You're right. Seems to be the first plugin I own that requires VST bank saving.
Working fine now. Also no overloads anymore with the Sinc32 setting. Cool, thanks!
Working fine now. Also no overloads anymore with the Sinc32 setting. Cool, thanks!
-
- KVRAF
- 2398 posts since 27 May, 2005 from Stockholm
While I can (and probably should) fix so that FXP:s make more sense, they (i.e. "minimal" plugin state) is not a good option for a sampler. A TX bank consists of N performances, each using M programs and K waves/matrices. Logically, a FXP would be a performance, the referenced program(s) and waves. But suppose you are working on a song using more than one perf (program change?) or you simply want to preserve the state of your work, i.e. including for example loaded but unused waves (wip)? If we use FXP saving and the previous logic, data would be lost on project saving. Less than ideal (imho).
Sure, I could _ignore_ the difference and save the same data, but then the whole "export FXP" becomes meaningless. Also bad.
Maybe this sheds some light on why things are as they are.
/C
Sure, I could _ignore_ the difference and save the same data, but then the whole "export FXP" becomes meaningless. Also bad.
Maybe this sheds some light on why things are as they are.
/C
TX16Wx Software Sampler:
http://www.tx16wx.com/
http://www.tx16wx.com/
-
- KVRian
- Topic Starter
- 687 posts since 27 Dec, 2003
I'm still getting overshoots (> +32dB) with the Sinc128 setting during offline rendering (kick drum track). Sinc64 doesn't do this.elcallio wrote:Sinc interpolation >32 should really only be used for offline rendering.
Is this expected behaviour or is this a bug in the Sinc128 algorithm?
-
- KVRAF
- 2398 posts since 27 May, 2005 from Stockholm
Can you give me an example project (with audio) so I can see the problem? I will try to pin point if there is something wrong.
TX16Wx Software Sampler:
http://www.tx16wx.com/
http://www.tx16wx.com/