LRPT Laboratory Bandwidth testing


Patrick Lindecker
 

Hello Fred,

 

>Question, In the I/Q mode is the Global Level relevant?   It did not go above 2% for this pass.  I did not correct
>the RSP gain at all.

In principle, weak signal can pose a problem if the digital noise is relatively high. For example, I remember that for old 16 bits sound cards, only 13 or 14 bits could really be used. So in that case very weak signals were mixed with a drop of noise.

 

I have not tested my RSP1A but Multipsk only uses the first 8 bits of the 10 or 12 bits possible, so I suppose that these 8 bits are surely “clean”. 2% of level means that you work with about 2 bits. It’s not ideal but it is not so bad, seeing your nice images.

 

Under DOS, there were decoders working with one bit and they worked well…

 

73

Patrick

 

De : multipsk@groups.io [mailto:multipsk@groups.io] De la part de Fred Albertson
Envoyé : mardi 22 décembre 2020 14:22
À : multipsk@groups.io
Objet : [multipsk] LRPT Laboratory Bandwidth testing

 

Hi Patrick,

This morning, on a medium quality pass, I used the RSP with no filters or amps. Again I used the 120kHz filter which
does not seem to affect the picture quality at all.  One thing I note is that the picture lines are more consistent
between the frames.

Question, In the I/Q mode is the Global Level relevant?   It did not go above 2% for this pass.  I did not correct
the RSP gain at all.

64                                                                               65                                                                             68


Great day,

Fred


Patrick Lindecker
 

Hello David and Fred,

Yes, internally small Jpg images (8x8) are transmitted (it's perhaps why it's a Low rate Picture Transmission and not a High rate).
I was wondering if it would not be possible to fill these lines of 8 pixels width with a set of "most probable pixels", taken into account the pixels at the frontier of these lines and the different derivative in the border zones of these lines. It would be an interesting mathematical game.

73
Patrick

-----Message d'origine-----
De : multipsk@groups.io [mailto:multipsk@groups.io] De la part de David J Taylor via groups.io
Envoyé : mardi 22 décembre 2020 15:21
À : multipsk@groups.io
Objet : Re: [multipsk] LRPT Laboratory Bandwidth testing

On 22/12/2020 13:25, Fred Albertson wrote:
Hi David,

I don't think so, because they decode ad BMP's from AMIGOS. I convert
them to JPG's for the e-mail. I have noticed no difference.

Thanks and great day,

Fred
Thanks, Fred.

At least one references suggests it's JPEG data:

http://jmfriedt.free.fr/glmf_meteor_eng.pdf
https://www.rtl-sdr.com/forum/viewtopic.php?t=343

If so, it might be better to work in the JPEG space rather than go JPEG=>BMP=>JPEG. BMP is much more convenient for subsequent processing, of course.

I'm uncertain whether or not JPEG=>BMP decoding always produces the same results or whether some assumptions are made (beyond the tables included with the JPEG data) which affect the results.

73,
David GM8ARV
--
SatSignal Software - Quality software for you
Web: https://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv


David J Taylor
 

On 22/12/2020 13:25, Fred Albertson wrote:
Hi David,
I don't think so, because they decode ad BMP's from AMIGOS.  I convert
them to JPG's for the e-mail.  I have noticed no difference.
Thanks and great day,
Fred
Thanks, Fred.

At least one references suggests it's JPEG data:

http://jmfriedt.free.fr/glmf_meteor_eng.pdf
https://www.rtl-sdr.com/forum/viewtopic.php?t=343

If so, it might be better to work in the JPEG space rather than go JPEG=>BMP=>JPEG. BMP is much more convenient for subsequent processing, of course.

I'm uncertain whether or not JPEG=>BMP decoding always produces the same results or whether some assumptions are made (beyond the tables included with the JPEG data) which affect the results.

73,
David GM8ARV
--
SatSignal Software - Quality software for you
Web: https://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv


Fred Albertson
 

Hi David,

I don't think so, because they decode ad BMP's from AMIGOS.  I convert
them to JPG's for the e-mail.  I have noticed no difference.

Thanks and great day,

Fred


Fred Albertson
 

Hi Patrick,

This morning, on a medium quality pass, I used the RSP with no filters or amps. Again I used the 120kHz filter which
does not seem to affect the picture quality at all.  One thing I note is that the picture lines are more consistent
between the frames.

Question, In the I/Q mode is the Global Level relevant?   It did not go above 2% for this pass.  I did not correct
the RSP gain at all.

64                                                                               65                                                                             68


Great day,

Fred


David J Taylor
 

On 21/12/2020 22:14, Patrick Lindecker wrote:
Hello Fred,
[]
I looked at these lines. In fact, they are 8 pixels width. A line must be composed of elementary small pictures of 8x8 pixels, transported by one or several bad received frames.
73
Patrick
JPEG chunks, perhaps?

73,
David GM8ARV
--
SatSignal Software - Quality software for you
Web: https://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv


Patrick Lindecker
 

Hello Fred,

 

TKS for the LRPT pictures.

 

>I was using the Sawbird but kept the signal levels low (which may help, or hurt, I don't know yet). 

If you can avoid it, better is to avoid. In my opinion, even if of good quality, it brings a small signal degradation.


>Anyway, beautiful snow pictures, but plenty of lines. 

I looked at these lines. In fact, they are 8 pixels width. A line must be composed of elementary small pictures of 8x8 pixels, transported by one or several bad received frames.

 

73

Patrick

 

De : multipsk@groups.io [mailto:multipsk@groups.io] De la part de Fred Albertson
Envoyé : lundi 21 décembre 2020 15:06
À : multipsk@groups.io
Objet : [multipsk] LRPT Laboratory Bandwidth testing

 

Hi Patrick, Andrea, Josef, et al,

Using the I/Q output from my RSPduo into MultiPSK (still refining the procedure), I decoded an almost perfect
pass this morning at 1318Z.  M2 was 88 degrees overhead and 826kM away.  The NORAD altitude information shows:

Perigee: 825.3 km
Apogee: 835.1 km


So it was about as close as it is going to get until it falls out of the sky!

The only problem was that NOAA19 came wandering in at about 1328Z, which halted the decode.  By that time
I had about 3700 synchros, and some pretty good pictures:

64                                                                              65                                                                             68                                                                             R-G-IR


I was using the Sawbird but kept the signal levels low (which may help, or hurt, I don't know yet). 

My theory was that narrowing the bandwidth, by using the 120kHz filter in the RSP, vice the full 192kHz,
would keep the VDL2 signals (136.975kHz), which I receive loud and clear, out of the signal that AMIGOS,
was seeing.  I believe that I did that effectively. I can not tell for sure that other VHF signals are not in
the bandwidth, but continuous manual monitoring with no satellite in view does not show interference.

Anyway, beautiful snow pictures, but plenty of lines. 

This is an Alpha level report, so I am not drawing any conclusions.

Thoughts?

Great Day,

Fred



Fred Albertson
 

Hi Patrick, Andrea, Josef, et al,

Using the I/Q output from my RSPduo into MultiPSK (still refining the procedure), I decoded an almost perfect
pass this morning at 1318Z.  M2 was 88 degrees overhead and 826kM away.  The NORAD altitude information shows:

Perigee: 825.3 km
Apogee: 835.1 km

So it was about as close as it is going to get until it falls out of the sky!

The only problem was that NOAA19 came wandering in at about 1328Z, which halted the decode.  By that time
I had about 3700 synchros, and some pretty good pictures:

64                                                                              65                                                                             68                                                                             R-G-IR


I was using the Sawbird but kept the signal levels low (which may help, or hurt, I don't know yet). 

My theory was that narrowing the bandwidth, by using the 120kHz filter in the RSP, vice the full 192kHz,
would keep the VDL2 signals (136.975kHz), which I receive loud and clear, out of the signal that AMIGOS,
was seeing.  I believe that I did that effectively. I can not tell for sure that other VHF signals are not in
the bandwidth, but continuous manual monitoring with no satellite in view does not show interference.

Anyway, beautiful snow pictures, but plenty of lines. 

This is an Alpha level report, so I am not drawing any conclusions.

Thoughts?

Great Day,

Fred