Hi Albrecht,
comments in line below.
On 05/10/2020 12:55, Albrecht Mugler wrote:
Dear OMs!
The following questions / suggestions about:
WSJT-X software version 2.3.0-rc1
Windows 10 operating system
FST4W mode
"TX NEXT"
If the TX NEXT button is pressed after restarting the program, CAT
(and also the other options) does not always switch to TX. This also
occurred in older versions. With the FST4W 1800, however, this is very
annoying because you have to wait a long time for the next transmission.
In this version "Tx Next" has been given higher priority but Next means
Next, it does not mean start transmission now, but at the start of the
next period.
"Audio Overload" error message
This message is likely caused by excessive latency in the audio
channel. This leads to the transmission being aborted, which is
disadvantageous in the case of long, repetitive transmissions. Perhaps
an automatic restart of the audio interface and the information for
the user would be a good solution here?
I will investigate, but you should try and reduce latency variation
since timing is critical to successful decoding of teh block modes in
WSJT-X.
"Upload to WSPRNET.ORG"
As I found out, messages are only forwarded to WSPRNET.ORG if they
were received in the audio ranges between 1400 Hz and 1600 Hz. Perhaps
this can be extended to the entire reception area?
This is fixed for the next release.
"OOB message"
If a frequency of, for example, 8270 Hz is set in the VLF area (Dial +
Audio), the message "OOB - out of band" appears. The frequency range
from 0 Hz to 8300 Hz is probably not regulated in most countries
around the world and can therefore be used for VLF applications. It
would be nice if there wasn't a "red" error message.
We use the ADIF specified bands to determine Amateur Band ranges. The
red highlighting and OOB band name is only a warning, is benign, and can
be ignored.
"VLF-LF-MF direct sampling"
The sampling rate is currently set to 48 kS / s. Numerous sound cards
now allow sampling rates of up to 784 kS / s and more. I am using a
384 kS / s Brooklyn equipement and can receive and transmit VLF and LF
frequencies directly. I wouldn't need any additional mixers or other
software (like SpectrumLab) if WSJT-X allows it. The reception of VLF,
LF and MF would therefore be very easy. An older WSPRX version could
receive and transmit 8270 Hz directly.
So it would be great if there was an "advanced option" for an extended
audio frequency range and higher sample rates!
The first step in decoding in WSJT-X is a downsample to 12000 Hz along
with a LPF with ~6 KHz cutoff. Feeding higher sample rates is of no
advantage. The core bandwidth of WSJT-X is <6 kHz. All remaining DSP is
based on this first "IF", so it would be a major change to increase the
bandwidth. WSJT-X is an add-on to an existing SSB transceiver and does
not attempt to replace RF and IF stages in such a rig. You might
consider a basic downsampling and filtering stage to select any 6 kHz
section of your wide-band ADC output shifted down to base-band. The
operating system will resample that to 48 kHz to satisfy the WSJT-X
request for a 48000 Hz sample stream. Neither step should be a source of
resampling noise as the ratios are integral. Although the upsample to
48000 Hz and then down again to 12000 Hz are redundant, I doubt they add
any significant overhead.
Also of note for transmission, currently we synthesize whole Tx
waveforms as a single PCM buffer, this is convenient with the GFSK
modulation scheme. Because this is done at 48 kHz he buffer is large for
up to 1800 s transmissions. Doing the same at higher sample rates would
need a lot of memory resource.
73
Albrecht
DL7NN
73
Bill
G4WJS.
_______________________________________________
wsjt-devel mailing list
wsjt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/wsjt-devel
