[Stk] Tutorial out of date?

[Gary Scavone]

Hi Ross,

One thing that should help is an optimization flag like "-O3" for the  
compiler.  If you're compiling for ALSA, you don't need the "-ljack"  
library.  Let me know if that helps.  Using Jack might give you the  
best performance.

Regards,

--gary

On 18-Jun-08, at 9:55 AM, Ross Clement wrote:

>
>
> On Wed, Jun 18, 2008 at 2:21 PM, Gary Scavone  
> <gary at ccrma.stanford.edu> wrote:
> Hi Ross,
>
> Can you send details regarding your OS, audio card, audio  
> parameters, and compilation flags?
>
> Regards,
>
> --gary
>
> Fedora Core 8.
>
> uname -r
> 2.6.24.3-12.fc8
>
> /sbin/lspci | grep Audio
> 00:1b.0 Audio device: Intel Corporation 82801H (ICH8 Family) HD  
> Audio Controller (rev 03)
>
> g++  -Wall -D__LINUX_ALSA__ -o duplex4 duplex4.cpp -lstk -lasound - 
> ljack
>
> I'm just hacking around at the moment, perhaps not all those  
> libraries are necessary ... checks ...  the -ljack is irrelevant,  
> but removing it doesn't solve the problem.
>
> There is definitely something weird going on. I hacked the program  
> (as you'll see in the source) to dump the first few seconds as input  
> to a file, I've attached that to this email as a .wav file. The  
> "clicks" are there already in the waveform. Finding the glitches  
> visually, it looks as though bits of the waveform are lost. But the  
> glitches are at irregular intervals, rather than the regular  
> intervals I would expect if I was not consuming all of the input  
> waveform. Nor are the glitches what I would expect from an  
> intermittent connection. The waveform is a static waveform produced  
> by a Moog Etherwave Theremin. I've checked the output of the  
> theremin itself and it sounds absolutely fine, even though the same  
> cable as being used.
>
> Note that in the code I've been switching between copying the input  
> to the output, and outputting a simple sine wave (see commented out  
> section). I get glitches in both situations.
>
> I must admit I'm stumped. I would be tempted to blame the hardware,  
> but I get no such problems with portaudio itself. Before I received  
> your email I was wondering how long it would take me to whip up a  
> temporary "RtAudio" class that only does one style of open to work  
> with in the meantime.
>
> The program I attach here is just an incredibly crude hack to prove  
> certain things going. I am concerned about tarnishing my  
> professional reputation by letting anyone see it. But it was meant  
> to be a "can use STK" proof of skills before writing the proper  
> program.
>
> Regards,
>
> Ross Clement <duplex3.cpp><junk.wav>