Filter cutoff, ASM, compile optimization, velocity curve

You're new to the LinuxSampler world? You don't know where to start and nothing works? Here's the place to ask for help.
Post Reply
sun2z_emily
Newbie
Posts: 2
Joined: Fri Oct 03, 2008 2:29 pm

Filter cutoff, ASM, compile optimization, velocity curve

Post by sun2z_emily » Fri Oct 03, 2008 3:07 pm

Hello,

I have compiled Linuxsampler from cvs (29 sep) and been testing it for a while. But I have a few questions

1 There is a compile time option of max and min filter cutoff frequency. Now, the default is min=100 and max=10000,
is it okay if i change it into something like min=20 max=20000. Im curious, since i thought that usually the parameter for an LPF is between 20-20000.

2 when I run linuxsampler i notice the line "Detected features: disabled at compile time". My proc is an Intel Dual Core (small Core2 Duo), so I figure it should have SSE and SSE2, is there something I should do when compiling? Btw, I compile using dpkg-buildpackage, and modifying the rules like

Code: Select all

CXXFLAGS="-march=nocona -funroll-loops -msse -msse2 -msse3 -mfpmath=sse -fomit-frame-pointer" ./configure --prefix=/usr --mandir=\$${prefix}/share/man --infodir=\$${prefix}/share/info --enable-preload-samples=65536 --enable-max-voices=256 --enable-max-streams=320
Does Linuxsampler affected by denormal problems?

3 Another question is regarding velocity curve. Is there any tips to edit the velocity layer of a gig file. Even after i set the velocity curve on my keyboard, some instruments are not loud enough on higher velocity (im using the maestro grand, and other free gigs). And also, Id like to make my own instruments, what are the most common way to divide velocity layer if Im only going to use two velocity layers.

There is still much to learn regarding linuxsampler, im very thankful for this great software.

Thank you.

Andreas
Developer
Posts: 214
Joined: Sun Feb 03, 2008 8:33 am

Re: Filter cutoff, ASM, compile optimization, velocity curve

Post by Andreas » Fri Oct 03, 2008 5:59 pm

There is a compile time option of max and min filter cutoff frequency. Now, the default is min=100 and max=10000,
is it okay if i change it into something like min=20 max=20000. Im curious, since i thought that usually the parameter for an LPF is between 20-20000.
Those configure parameters are not used anymore, so it doesn't matter what value you set them to. They should probably have been removed from the configure script long ago.

2 when I run linuxsampler i notice the line "Detected features: disabled at compile time". My proc is an Intel Dual Core (small Core2 Duo), so I figure it should have SSE and SSE2, is there something I should do when compiling? Btw, I compile using dpkg-buildpackage, and modifying the rules like

Code: Select all

CXXFLAGS="-march=nocona -funroll-loops -msse -msse2 -msse3 -mfpmath=sse -fomit-frame-pointer" ./configure --prefix=/usr --mandir=\$${prefix}/share/man --infodir=\$${prefix}/share/info --enable-preload-samples=65536 --enable-max-voices=256 --enable-max-streams=320
That's strange. The configure script is supposed to detect if the target cpu is an X86, but in you're case it hasn't. If it had, the printout would have been something like "Detected features: MMX SSE SSE2".
Does Linuxsampler affected by denormal problems?
Yes, it probably depends on the gig file used, but if LS doesn't detect SSE2, I have seen it having denormal problems.
3 Another question is regarding velocity curve. Is there any tips to edit the velocity layer of a gig file.
The velocity curve parameters can be edited in gigedit, on the "Misc" tab. There are three of them: curve, depth, and scaling, and they are not very intuitive. You can see some graphs here that might help a little: http://hem.spray.se/andreas56/curves/

sun2z_emily
Newbie
Posts: 2
Joined: Fri Oct 03, 2008 2:29 pm

Re: Filter cutoff, ASM, compile optimization, velocity curve

Post by sun2z_emily » Sat Oct 04, 2008 3:36 am

That's strange. The configure script is supposed to detect if the target cpu is an X86, but in you're case it hasn't. If it had, the printout would have been something like "Detected features: MMX SSE SSE2".
I found out what's the problem.
The configure script did not detect that my cpu is an x86. I see #define ARCH_X86 0 in my config.log
Im not really sure why though. The config script looks whether target_cpu is one of i386, i486, i586, i686, or i786. I guess mine is not of those

I fixed it by adding ARCH_X86 1 in my config.h

Thank you.

User avatar
dahnielson
Moderator
Posts: 632
Joined: Wed Jan 23, 2008 11:25 pm
Location: Linköping / Tranås, Sweden
Contact:

Re: Filter cutoff, ASM, compile optimization, velocity curve

Post by dahnielson » Sat Oct 04, 2008 12:31 pm

Bug reported and patch submitted:

https://bugs.linuxsampler.org/cgi-bin/s ... cgi?id=107

Thank you for pointing this out!

Update: My patch doesn't solve it completely. You still need to pass CXXFLAGS="-march=nocona" to configure...
Anders Dahnielson

Ardour2, Qtractor, Linuxsampler, M-AUDIO Delta 1010, Axiom 61, Korg D12, AKAI S2000, E-MU Proteus 2k, Roland R-5, Roland HP 1300e, Zoom RFX-1000, 4GB RAM x86_64 Intel Pentium Dual 1.80GHz Gentoo Linux

Post Reply