Question: what other modes are there in MMSSTV that will work in
the narrow 500 Hz bandwidth?
I think its limited to
mp73-n, mp-110-n, mp140-n and the
mc110-n, mc140-n, mc180-n modes.
Unluckily i have never found a station qrv in those modes...
Any big fans of MULTIPSK that might like offer why they use it? I'm
guessing you're using it for the ALE applications?
I have made the test of running multipsk in parallel with winpsk and
other psk31 software. The programs were set up to decode the same
signals on 40m. Then i did compare the
I wonder how long we will have to read these sales pitches from the
support group ?
How is what we tested somehow flawed in your mind?
Compare it to testing psk31 against jt65 at snr's of -20db.
psk31 wont work, jt65 will work. Flawed test.
Your report of easypal only getting 1/6 of transmission is a clear
indication that conditions or setup were not adequate. I could not
I think i have expressed my point of view.
Further discussions only per PM please.
Andy does have my real email.
Unluckily i have to say that this comparison is quite flawed...
using easypal which needs minimum 6db SNR in the lowest setting in
conditions of less than 4db snr (1/6 throghput you say) is not a good
idea. Comparing that to a mode which can adapt to lower snr's is BS at
best.
Your invited to
Both soundcards should also be qite precise at 48k sampling rate.
If one or both cards are not able to precisely do 48k samples you will get
bad audio quality, periodic interruptions, or bad SNR values.
For vista users it's important to set the default audio rate to 48k, not
44100. I do not
I noticed that there are images that are sent with an RS 1
through 4 and then I think it converts them to jpg?
The RSx files are additionally encoded. I dont remember the exact numbers
but i think it allows 10% errors on rs1, 50% errors on rs4. You can
actually decode an rs4 file when 50%
If you spend a lot of money on an interface,
it's because you wanted to, not because you need to.
Right !
The main points of such an interface are to provide isolation between PC
and Rig. This is normally done by audio-transformers for the mic and
speaker lines, optoisolators for ptt line.
a very powerful compression scheme, it's perhaps possible.
The compression used is .jpg or .jp2 (jpeg-2000)
For reasonybly sized pictures 10kb is minimum, 20kb is average size.
The quality does deteriorate rapidly below 20kb/picture.
The drm-modem (qam-16 mode b) has a raw troughput of
Hi,
Is there a bozo SVN guide for windows?
The task is not to retreive anything, but to put a project into
sourceforge or similar. The project is about 50 VC++ files and some libs.
Any instructions would be appreciated.
73, Cesco, HB9TLK
I also have a utility called Tortoise SVN
that integrates with Windows Explorer.
I have tried this and got extremely confused. I prefer the command line
tool.
I have not seen a bozos guide but can assist you. Offline from the group
might be better...
TNX.
I will mail you directly, maybe
I will think about it.
It would be nice to have the pskmail functionality in a re-usable form.
The actual code is not easy to run in windows, i did try and failed.
The actual file-io message transfer system has the advantage to be
universally usable. It's not linked to a specific psk31
So,
perhaps we can make more of an effort to use this band, Has anyone
tried it for Digital Voice. ?
Region 1 bandplan 30 m :
10100 - 10140 200 hz BW CW, QRP
10140 - 10150 500 hz BW digimodes
No DV until someone makes a 500hz wide dv-mode for region 1.
Currently there are 25 MP3's on the BARTG website and they include ALE,
DominoEX, Olivia and MT63.
The DRM sound is messed-up.
They got the feldhell-type text messages instead of a drm signal.
Hi Roland,
O Arnaldo PY4BL, fez um ótimo contato hoje
em FDMDV com IZ2GAF Doriano, com boa qualidade
Wow. PY to IZ could be a record distance.
você acrescentar em
Settings, Sound, About a opção View, onde seriam
colocadas as opções de Display (Waterfall, etc...)
You are right. Tnx for
How far off are we
from a day when we can do digital video and voice, video with
perhaps a 5 frames a second rate ? Decades ?
A highly jpg compressed 160*80 picture is about 1.2k At 5 frames/sec
this is 6kb/sec. By using the P and I frames concept this could maybe
be lowered to 3kb/sec.
oops .. those numbers are wrong. I did mix up kbyte and kbit.
The video may be limited to less than 1 frame / sec.
Hi Patrick,
Signal to Noise ratio of -12 dB
about 1.5 sec
about 3 Hz
WOW.
This sounds excellent.
The 1.5 sec is too long to use it before every transmission, but good
for the initial transmission as you suggest. This would allow standby
operation. Nice.
I would be very intrested !
73,
Hi Roland,
Parabéns pelo FDMDV !
Obrigado. Muito feliz de ver seu mail !
Tenho uma sugestão:
Seria melhor se a frequencia de transmissão
fosse FIXA,
A nova versao vai ter uma opcao a de voltar ao centro da banda.
Tambem concertei o auto-tune.
73, Cesco, HB9TLK
Hi Patrick,
It will be perhaps interesting to have a specific RS ID
which would permit to auto-tune automatically
This would allow to drop the central carrier and get a 75hz narrower
signal. It would loose the ability to transfer text, but gain 10% in
robustness. Good idea.
Questions:
How
Here is my XYL after encoding with MELP and FDMDV
Are you trying to discredit the program by posting worst-case examples ?
Pese al ajuste del reloj, las señales empìezan a aparecer en la
cascada en cualquier momento y no en el punto del cambio de minuto
como debería ser.
He did adjust the clock, but signals are starting ramdomly, not at minute
boundaries.
I guess the clock adjustment went wrong.
Hi Steinar,
Anyone for digital voice qso 14.236?
I have uploaded a recording of your CQ call audio in the files-
digitalvoice section.
A 50 baud signal should have either a 50 Hz or a 100 Hz separation by my
understanding.
fdmdv modem is NOT OFDM. (would be called OFDMDV if it was)
And even with ofdm, 50baud and 50hz will not work in real world. Try to
googe ofdm and guard interval.
What spacing would an ofdm systen with
Modem description from Peter G3PLX (the modem author):
---
The modem is based on a raised cosine shaped tone filter response which
has the property of zero ISI and zero adjacent tone interference with no
sidelobes. Half the channel filter is in the tx and the other half in the
rx, so it's
Hi Andy,
FDMDV is the latest DRM digital voice mode on HF
FDMDV does not use the DRM modem, but the FDM modem developed by a
very well known digital mode expert. The FDM modem beats DRM in low-
snr performance and qrm-robustness.
73, Cesco
It's my understanding that 19.2k just isn't possible with a soundcard.
Wrong.
The main question is how much BW your radio provides.
Having 10khz of BW 19.2k should be doable. Have a look at the DREAM
software http://drm.sourceforge.net/ . It should get pretty close to 19.2
on 10khz BW, and
I would be plased to have a complete list of the phonemes and corresponding
audio files from different speakers. I fear 44 phonemes will not be enough
to do a context-free analisis.
The data rate will be closer to 200pbs i think, since you will have to
transfer a magnitude component along with
Very low bitrate algorithms exist now. There are a few that operate from
200 bps to 600 bps. The Navy has software called IVOX that gets in this
range.
Can you somehow lay hands on such a 200 to 600 bps codec?
Im VERY intrested.
The IVOX thing is based on 2400 bps lpc. With silence
I did send you a PM.
TNX for the info.
I think you should post the message again when the baloon is on it's
way. Or just a reminder so i wont forget to listen-in.
I think it's a very intresting experiment lasting only a few days.
So i personally dont understand the attitude of interpreting the
national us band
Could a
higher higher success rate without the fill requests be achieved with
some combination of better FEC, slower data rate, better spreading of
the data over frequency and/or time?
Try it. It's all there.
FEC, data spreading and data rate can be selected in DRM, read the doc.
Easypal
Various sources I have read put typical maximum multipath at 10ms. If
the symbol period is 25ms then there would not be a need for a guard
interval since the critical part of the symbol is undistorted.
Well, we want all of the symbol undistorted .. not only a part.
So the 25ms symbol and the
I think the DV part of dstar is somehow a dead-end...
The intresting ability of d-star is the high-speed mode in the D1 units.
Seems it's capable of something like 80kb/sec, but there is very few info
on it. It could be 140khz wide .. does someone have documentation about ?
This data rate
Since RDFT was released under the GPL license, failure to release the
source code for SCAMP may be violation of the GPL license.
The rdft routines are not integrated into scamp, they are external
exe's. Same trick is used by mixw, gpl-code is moved to an external
library.
While this seems to
Are any minimum SNR numbers of those 2 modes known (375/300 bps)?
The SNR for 8FSK can be about -5dB.
The SNR for 8PSK at 75baud is about -10dB.
Sorry, I don't remember the typical SNR for 300 or 600 baud 8PSK.
I think we must start to specify SNR at Bandwidth, otherwise the numbers
dont
The 8FSK raw speed of ALE-mode provides 375baud,
Try the 8PSK mode in PCALE, under the
The increments are: 75, 150, 300, 600, 1200, 2400 baud.
Thanks.
ARQ makes no sense in DV, and FEC is easily adaptable to meet those 375
or 300 data rates of those 2 modes.
Are any minimum SNR numbers of
don't think you can do much better than WinDRM type OFDM modes
since you need the bps throughput to make it work adequately.
you can.
windrm uses 2400 bps at 8db snr min.
drmdv uses 1200 bps at 5 db snr min.
the new codec would use around 250 bps, wich givs an improvement of
6db in minimum
Question to the experts:
What is the most robust digi-mode with a 400 to 600 bps raw capacity (or
200 to 300 bps fec capacity) ?
Is a boosted psk variant the best coice or is mfsk capable of such rates ?
Other options ?
Background:
Seems someone has developed a 200bps voice codec. It's a
Its now clear why the two-year-old SCAMP busy frequency
detector has
never been incorporated into WinLink servers.
How should they do it?
One is a soundcard app, the other a proprietary microcontroller box.
in the old days, when packet qrm was sometimes a problem we had a
robot silencer
This should do it, but untested with wsjt deep search yet.
http://www.qslnet.de/member/hb9tlk/deepsearch.html
Double-click the trace you want decoded.
Then double click another one. Works fine.
How would one format an HF call3.txt
example from file:
ZS6GPM,KG33XU,,10/02
i think format is
callsign,locator,,date
additional entry would be:
HB9TLK,JN47HJ,,3/07
Guess for HF the EME entries could be deleted to avoid false detects.
The Bozo's Guide is excellent andy !
I would never have used JT65A on HF on my own,
thanks to the group for making some hype around it.
The codebase appears to be GPL
Yes, the source is there, but it's a mix of phyton, fortran and C.
Not easy to see trough, or to build it.
btw the intel fortran compiler is free for 30 days ...
183900 1 6/8KT2Q ZS6WN KG46 1 0
seems he's stuck there. doesent stop calling you.
Try reducing the input level.
I am using the creative mp3 usb for wsjt, and i have to set the input
level slider near the botton for wsjt. Same slider is is in top position
for psk31 or other modes. WSJT is too sensitive for normal position.
Setting the freq. in the SpecJT window does change only rx freq, but
tx stays the same?
I did not know that. Thanks for info.
HI, that one is good.
I guess the Markov Model for DV would be
1 2 3 4 5 test, followed by a software crash.
I could not follow and got messed up. You win Andy !
DV (Digital Voice) using PSK at 93 bps
The new PSKDV Digital Voice soundcard mode allows acceptable voice
quality over quite bad HF channels. It uses a new type of voice
compressor with extremly high compression rates, and a transmission
system very similar to psk31, but 3 times faster.
The
The only other known use for voice-bandwidth data modes is for image
transfers,
which can send an SSTV-size picture, with a very
low error rate, in 30 seconds, using a bandwidth of 2400 Hz.
the same image, at the
same low error rate, can be sent in less than 2 minutes, using a
bandwidth
So using multimulticarrier soundmodem with a
YaeComWood + 1kW PA will
only heat your ham shack without other useful effect.
Negative.
The grounded-grid PA's have no negative effect on SNR if tuned right
(tuned for peak power). Those PA's have better lineariy than the 100W
push-pull
when n becomes big, the ratio tends to 1/square(n)
For example, for MT63 where you have 64 carriers in parallel,
You transmit only 12.5 watts with a 100 watts maximum XCVR.
The same applies to the digi-sstv hamdrm/windrm modes. 30 to 60
carriers, crest factor (average to peak power) is
As I understand it, DigiTRX won't run on less than WIN XP..
No. Digtrx runs happily on win98. But you need 1ghz or better.
In my point of view:
TCP is complicated !
Its easy to build a client (maybe), its more complicated to build a
server, and it's hard to build a server which reliably can accept
and
handle multiple connections on one port !
The solution could be to have a common DLL with an easy interface
for
When I get
my soundcard DLL written for VISTA I'll make the source available.
This would be great !
btw, where did you get the new sound API from ?
So
far there's no IQ software for digital mode reception (please prove
me wrong!)
I hope so:
DREAM is IQ capable.
Start it with dream.exe -c 3 (see dream's help) and feed it with your
IQ signal. More, since dream is open-source, and dreams iq-capability is
a modular phase-shift hilbert
So, , what digital modes exist that are in the 12 to 20 kHz range
could I use if I when and bought a SDR today ?
DRM.
9, 10, 18 and 20khz wide. Really good bitrate.
and it's free. google DREAM drm receiver.
DREAM does transmit too, DV (cd-quality) or data.
If this would be legal we would
I often wonder if there is even one ham working on adapting
the existing
ham DRM type protocol to a pipelined ARQ connected mode that has
adaptability to conditions.
I think no.
The main problem of arq-drm is the very long turnaround time.
It's in the 20sec range. This makes normal arq
How do you determine your specific 20 second turnaround time?
Couldn't it be any reasonable number from say 1 second up to maybe
20
seconds?
The time from start of transmission until receiving the first data
segment is 10s to 15s. That's the sync-zone, the lead-in.
20 sec is not to be
While using a sound card with a wide frequency response, and using high
sampling rates, might intuitively appear to be a good idea, I'd suggest
this is not likely to be the case.
Thoughts? What am I missing in this logic?
my thoughts (not veryfied) :
You should use the max. sample rate the
Near the equator,
there is little frequency spread ( 4 Hz), but it is larger
in near-polar paths and can be very large (up to 40 Hz)
under disturbed conditions.
A question: where does the frequency spread come from ?
Is this a doppler effect of a moving ionosphere, or are
there other
is that dynamic
variable, does it increment amplitude level symbols based upon
updating ARQ feedback,
No
Or fixed symbol based
upon manual operator setting, or a link initialization routine?
Yes.
The transmitter sets the mode, the rx will adapt automatically
Which
modem is it?
The
All LCD or TFT monitors produce severe picture distortion at
any but the default monitor resolution.
I would not care about picture distorsion, but at any but the default
monitor resolution the text becomes blurry (not sharp). The cause is
that any resolution has to be up or downscaled to
OFDM squeezes more carriers into
the same space by making the carriers orthogonal
to their next door neighbors.
Its not only the next door neighbors but ALL the carriers.
If you look at the the ofdm signal with a correctly timed FFT with
exactly a symbol time length, there is NO overlap of
I foresee and recommend this variable OFDM symbol approach as the best
next step in fast HF modems for hams.
Those systems are actually very sucessful in high data rate
applications like digital sstv. Appearing 2 years ago they
have become mainstream in a very short time.
Its not the next
I am QRV on 14.109,5 USB (VFO) and Wait connection...
Calling you, 18:32 utc (i think) 14.109.5 vfo, 14.111 center.
No answer till now.
relevant to its classification to OFDM. Which it is NOT. The
carriers
are on 120 Hz centers and the baud times are 100 Hz. Because the
baud
time is not commensurate with angular frequency of the carriers, the
dot products are not zero and therefore, they are NOT orthogonal in
Is anyone in the grouping using the Rigexpert SD for the WinDRM DV
mode?
The Rigexpert SD is essentially an expensive usb-soundcard with the
correct cables already installed. I dont have one, but there should be
no problem at all. Direct the voice audio inout to your PC's normal
soundcard,
The WinDRM specification is very sketchy ..
Well,... you are invited to provide a better one.
but the FEC is not
described. The interleaving of the pilots and overhead data is
described but not the interleaving of the voice data.
It says This document describes the DIFFERENCE of mode
It was my understanding that the QAM-4 modulation was used for the
text transmission
Negative.
Text and data can be anything from qam-4 to qam-64
FAC data (the callsign) is the only thing which is always qam-4
and you needed to use at least QAM-16 for the
voice.
Since the codecs used
The only solution to do voice exchange with a low S/N would
be to translate all the pronounced words in symbols
Exactly what i am thinking!
We need to establish a phoneme alphabet.
Then, a correlator is needed to extract those phonems from the voice
input. Each phonem should have duration
From my understanding of the documentation for WinDRM, MELP
does work if the speed is at least 1,000 bps.
No. 1000 bp400ms (bit per 400 ms) or 2400 bps
This per second / per 400 millisecond mixture is confusing !
Need a Digital mode QSO? Connect to Telnet://cluster.dynalias.org
Other
75 matches
Mail list logo
