Sat Jan 25, 2020 9:21 pm
Hi there,
first of all, since this is my very first message on the forum, I want to say a big, huge THANK YOU to all guys who made LMMS. This software is simply great, it can easily compete with expensive piece of software (I plan to join the fun of participating to it sooner or later, as I am a developer myself, but too busy with the Zynthian project for now
I am facing two big issues with version 1.2.1
- first, the WDM-KS seems totally unstable on my machine. I have three different audio interfaces on my machine to do the tests, and they all fail in a way or another
Here is what I see :
- in 50% of cases, LMMS refuse to start if it has been configured to use WDM-KS. When the software starts, it displays the configuration panel, where I see the WDM-KS devices, but when I choose one interface and restarts LMMS, software restarts by displaying the configuration panel as if it detects something wrong like a missing interface
- when the software accepts to start, the CPU load rises immediately at alarming levels (60 to 70%), even with the default project. If I load even a simple project, I immediately reach 100% on the bargraph. If I look to system stats, I see that "system interrupts" is taking 100% of the CPU
- if I click on PLAY button, sequence does not even start
- I did the test with a "Realtek HD", M-Audio and a XMOS USB based interface : same result, as soon as I choose WDM-KS
Second issue :
there is a terrible noticeable delay between audio and MIDI output, even if I use WASAPI (I don't even want to think about MME or Direct Sound, which are s**t for serious real time applications). The strange thing is that I get almost acceptable results if I reduce the buffer size to the minimum (64 samples / 1.5ms), which is a totally unrealistic value even on a fast PC (most audio cards are unable to go down to 64 samples without clicks on Windows)
The symptom is the following one : if I create a track with SF2 player for example and send the same track to MIDI OUT, I clearly hear the same track being played twice (not like a chorus effect which would be normal, but like a delayed track)
On a kick drum for example, the audio ouput comes around 20 to 30ms later (and clearly not only 1.5ms as defined in the audio configuration) than the hardware synth played by MIDI, and I can hear something like "bom bom" where there is only one note on the track
Any idea of what can explain these behaviours (and how I can avoid them) ?
Thanks
Benoit
first of all, since this is my very first message on the forum, I want to say a big, huge THANK YOU to all guys who made LMMS. This software is simply great, it can easily compete with expensive piece of software (I plan to join the fun of participating to it sooner or later, as I am a developer myself, but too busy with the Zynthian project for now
I am facing two big issues with version 1.2.1
- first, the WDM-KS seems totally unstable on my machine. I have three different audio interfaces on my machine to do the tests, and they all fail in a way or another
Here is what I see :
- in 50% of cases, LMMS refuse to start if it has been configured to use WDM-KS. When the software starts, it displays the configuration panel, where I see the WDM-KS devices, but when I choose one interface and restarts LMMS, software restarts by displaying the configuration panel as if it detects something wrong like a missing interface
- when the software accepts to start, the CPU load rises immediately at alarming levels (60 to 70%), even with the default project. If I load even a simple project, I immediately reach 100% on the bargraph. If I look to system stats, I see that "system interrupts" is taking 100% of the CPU
- if I click on PLAY button, sequence does not even start
- I did the test with a "Realtek HD", M-Audio and a XMOS USB based interface : same result, as soon as I choose WDM-KS
Second issue :
there is a terrible noticeable delay between audio and MIDI output, even if I use WASAPI (I don't even want to think about MME or Direct Sound, which are s**t for serious real time applications). The strange thing is that I get almost acceptable results if I reduce the buffer size to the minimum (64 samples / 1.5ms), which is a totally unrealistic value even on a fast PC (most audio cards are unable to go down to 64 samples without clicks on Windows)
The symptom is the following one : if I create a track with SF2 player for example and send the same track to MIDI OUT, I clearly hear the same track being played twice (not like a chorus effect which would be normal, but like a delayed track)
On a kick drum for example, the audio ouput comes around 20 to 30ms later (and clearly not only 1.5ms as defined in the audio configuration) than the hardware synth played by MIDI, and I can hear something like "bom bom" where there is only one note on the track
Any idea of what can explain these behaviours (and how I can avoid them) ?
Thanks
Benoit
) and ask first to the community if someboady was already aware of this and knew the solution