| |
VE2HAR > MT63 13.03.05 22:10l 75 Lines 2587 Bytes #-7304 (0) @ WW
BID : 21061SENTTO
Read: GUEST
Subj: Re: [MT63] Re: Crest Factors
Path: DB0FHN<DB0RGB<DB0MRW<OK0PPL<DB0RES<DK0WUE<7M3TJZ<KP4IG<W1NGL<VA2HAR<
VE2HAR
Sent: 050313/2045z @:VE2HAR.#MTL.QC.CAN.NOAM Laval #:42237 $:21061sentto
On Sun, 2005-03-13 at 20:35, Paul L Schmidt wrote:
> $ dd if=/dev/urandom bs=1 count=1024 of=random.dat
> $ mt63tx -srandom.mt63 -trandom.dat
> $ sox -t raw -sw -fs -r8000 random.mt63 -t wav /dev/null stat
Yep, that's basically what I did, except I didn't know about
the stat option to sox (cool!). I used a small piece of self
written code to analyze it.
> The questions remaining is what portion of time does the
> waveform exceed a particular value? -- if these peaks occur
> less than 1% of the time, for example, they'll probably get
> lost when monitoring on a scope. They'll also probably get
> clipped by the output stage of the transmitter and no one
> will notice. Is this important? Given the amount of
> redundance built into the signal, probably not from the
> perspective of data loss -- but how about from the standpoint
> of spectral purity?
I did some experimentation with hard clipping and some other
tricks. (I used octave to clip and analyze the data.)
I don't remember my exact findings any more but off the top
of my head it was around 3...4 dB of hard clipping where the
noise sidebands were ~40 dB below the wanted signal.
If the clipping is done in the finals at RF, the resulting
noise floor output will be very wide as we all know. I was wondering
if it would be better to clip at AF and let the transmitter
filter the resulting broadband noise to a mere SSB bandwidth.
Unfortunately any filtering of the clipped waveform tends to again
make the crest factor larger.
I also read a few IEEE papers to learn how the big boys
in the commercial world tackle this issue, but came to the
conclusion that the modem would probably need to be redesigned
to use the tricks that they use. At least I wasn't able to find
a way to implement those in MT63. Of course that doesn't mean
it could not be done...
--
Tomi Manninen / OH2BNS / KP20JF74
------------------------ Yahoo! Groups Sponsor --------------------~-->
Over 1 billion served! The most music videos on the web.
Click to Watch now!
http://us.click.yahoo.com/xmKGzA/IARHAA/kkyPAA/CPMolB/TM
--------------------------------------------------------------------~->
<< Try MT63 on 80m - great fun!>>
- The MT63 Reflector -
MT63@egroups.com
(To unsubscribe. send email to
MT63-unsubscribe@onelist.com)
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/MT63/
<*> To unsubscribe from this group, send an email to:
MT63-unsubscribe@yahoogroups.com
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
Read previous mail | Read next mail
| |
