[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [linrad] linrad operation



I've noted that some of the problems with linrad may be due to the video 
drivers. For instant, the SiS driver can only operate with 1024k of memory in 
SVGAlib. The driver that SIS allowed to be used with linux is more supported 
in Xwindows. I'm having a few problems with a Rage 3d pro AGP board also.
On my system, linrad does not survive console switching either.
John.

On Wednesday 02 October 2002 03:45 pm, Leif Åsbrink wrote:
> Hello Arie,
>
> > Thank you for your comments on the waterfall. But what is the N I have to
> > adapt?
>
> N="Waterfall avg" one of the parameters you are supposed to want to
> change now and then in real time.
>
> > But I think it is usefull to inform you about what happens here (
> > version 44 and higher).
> >
> > I will tell you what happens in the various steps.
> >
> > I start with the function U to set up the a/d convertor.
> >
> > Linrad comes back with nrs 32 and 33 which are dsp and the linked dsp0.
> > It tells that the max sampling rate is 192000, 2 ch., 16 bit and all the
> > three RW modes.
>
> OK. The fact that max rate is 192000 most probably indicates that
> your sound device does not respond properly to the ioctl call that sets
> the speed. If a too high speed is attempted, the device should respond
> with the highest speed it actually supports. 192000 is the speed Linrad
> tries to set and it is NOT the value expected in return.
>
> > I choose 32 , 1 channel. A warning comes that device does not respond
> > correctly to speed setting command. But it finishes with
> > configured without errors.
>
> Yes. Could work ok - but I assume you have asked for something
> other than 192000 Hz. Many device drivers work properly if you set
> a speed that the hardware actually supports.
>
> > pressing A then results in
> >
> > 200 Hz, sin(3), FFT version 0, storage 2s, amplitude 100, second
> > FFT 0 which
>
> "results in" is not adequate. You decided to select those values which
> will not produce very useful results for weak signal operation.
> A very large bandwidth when second fft is disabled. It would be
> more reasonable to set the bandwidth to somewhere between 2 and 20 Hz.
>
> > then results in :
> >
> > DMA15 mode 0 16 bit MMX present, 1st fft size 128, increment 64, SIMD
> > present, 1st fft sin3, BW 62,5 Hz, 1st fft buffer 131 kB, timing FFFT
> > version 0 : ...radix 2 DIF C
> > 0,53% N189, buffer 0/128.
>
> This is fine - if it is what you want. (for very fast response on
> relatively strong signals)
>
> > I press Y
> >
> > I chhose 1 channel first mixer.
> > Then the picture comes up.
> >
> > Without input signal between 500>3000 Hz about 0 dB on the graph and
> > pressing A gives:
> > A/D 57,24, 2,7%, 0,0007, floor 0.0
>
> So your input is 57.24dB below saturation of the A/D.
>
> > After inserting a noise signal I see the graph between 500 and
> > 200 Hz around
> > 60 dB with ( the input is only noise) a weak signal around 1000 Hz.
>
> So your signal is most probably not noise only, there is some 1kHz
> tone present as well.
>
> > A gives : A/D 14,64, 4,0 % 0,0005 floor 0.0.
>
> You are only 14.64 dB below saturation for the A/D input.
> There is too much signal. There is no need to put the noise floor more than
> 20 dB above the noise floor you have without any signal so you should
> reduce by about 40dB.
>
> > When putting the green line on the graph a spurious signal is visible
> > dependant on the output tone frequency ( some image I suppose)
> > and the ouput level.
>
> No, most soundcards have a switch that connects the output to the
> A/D input. If you use oss the mixer settings rlinesw and llinesw should
> both be OFF.
>
> > So far so good.
> >
> > But then : I switch console for changing the mixer controls. After having
> > done this I have to restart linrad ( it does not survive console
> > switching)
>
> Yes. Linrad does NOT survive console switching. I do not know why
> but I am pretty sure it is related to svgalib.
>
> > and then pressing A the graphs appear but the mouse is very slow.
> > pressing the A then gives A/D 14,01 , 89,2% !!!!, 0,0003, floor 0,0.
> > If I then let it go linrad locks and I have to go to another
> > console to kill it.
> > If I press X-P fast enough I do the setup again by confirming the already
> > present parameters and the linrad runs with about the same values as with
> > the first setup.
> >
> > So something is happening when restarting Linrad.
>
> I think you have to stop the sound system and restart it if
> Linrad crashed due to a console switching. No one else has reported
> this problem. Are you by any chance using ALSA?
>
> > This is the reason I give you all those details, for it might give you an
> > indication what I am doing wrong.
>
> You are not really doing anything "wrong" here. I think you observe
> errors within svgalib and the sound system;-)
>
> > The program runs on a Thinkpad600E, 350 MHz pentium, 128 MB RAM ,
> > soundcart
> > set up by ALSA ( which is quite integrated in SuSE 7.3) for
> > general CS4236+.
> > This card  in this thinkpad anyhow is a strange device for alsa
> > sees another
> > card but mentions that for thinkpad this (PCI) card shall not be
> > chosen but
> > the cs4236 emulation in stead. In windows both interfaces of this card
> > are seen/setup.
>
> As far as I know nobody else has been able to run Linrad on a laptop.
> It seems to me you have been able to set it up sucessfully and you only
> need some trivial parameter changes to run it properly.
>
> I would suggest something like this for microwave use:
>  5 Hz, sin(2), FFT version 2, storage 20s, amplitude 1000, second
>  FFT 0
>
> For fast response set "Waterfall avg" to 1 (you have to click on
> the little box with a number in it in the lower right corner of the
> main spectrum for "1" to show up in the box.)
>
> Use the F1 help key!!!!
>
>
> 73
>
> Leif  /  SM5BSZ

LINRADDARNIL
&