I see similar to this… Running my test currently and will post results a bit later… Seems alleyoopie has less traffic? perhaps that’s a factor?
It’s quite the opposite. He has very strong signals and lots of traffic. v2.2-RC1 is coming.
Ah apologies I merged allyoopie’s graph with jlb56’s AC count.
I ran this file with the new fix at -e 50.
DF 0 : 3869
DF 4 : 2078
DF 5 : 409
DF11 : 11701
DF16 : 211
DF17 : 3378 <=
DF18 : 9
DF19 : 7
DF20 : 3169
DF21 : 1993
Total : 26824
Max AC Count: 258
CPU time used: 3.875 s
I wish I had discovered the “sample approach” before. I finally have the feeling I do see real impacts, instead of just making (often false) assumptions. There is perhaps room for kind of a “benchmark script/option” here … But perhaps there are too many things to take into account.
1 Like
I’ve graphed my DF11s as they are the most plentiful in my sample. TOTAL seems to follow the same pattern though. Blue is NARROW and Orange is WIDE. Sample taken around 8am when traffic was medium to high.
Pi4/Buster with 1GB sample from a mini.
Currently the same settings shown at -m 20 -e 26, as that gives me around 80% CPU
(from the graph there seems to be a sweetspot around -e 32? But not sure I’ll get there with CPU 
airspy_adsb -v -F /media/airspy_sample/sample.bin -f 1 -w 5 -e 26 -m 20 -B narrow
airspy_adsb v2.2-RC0
DF 0 : 3212
DF 4 : 7612
DF 5 : 1886
DF11 : 18001
DF16 : 236
DF17 : 6376
DF18 : 79
DF19 : 0
DF20 : 7385
DF21 : 2563
Total : 47350
Max AC Count: 125
CPU time used: 20.188 s
1 Like
Yeah centered between London, Paris, Brussels, Amsterdam and Frankfurt.
We see more airplanes than birds over here 
2 Likes
Will be a step function soon, preamble_filter needs to get less granular.
But we’ll keep old input values for compatibility.
1 Like
Thank you for this valuable input. I think these curves will be even smoother after the fix.
1 Like
I am currently not testing any longer.
Piaware 6.0 is causing MLAT trouble at least for me and a few others
I’ve connected my Jetvision device again, Will return to the Airspy-Raspberry once the issues are fixed.
We are aware of the MLAT UDP issues and are working on a fix.
1 Like
2.2-RC1 available via the scripts
- fix overflow, high gain high traffic samples should now see a linear increase with the preamble setting … where as the previous behaviour was a bit erratic
- preamble settings are different now, new 38 corresponds to old 98, you’ll figure it out
- slight changes to RSSI might happen, please post graphs
1 Like
You had to post 3 pages? 
Use pastebin for that
I’ll adjust for it to be less aggressive.
2 Likes
Not sure I get the point about the new 38 being the old 98. My system is now at e 101 with the same C limit… 
You loaded before that was done.
Load again.
I had pushed to github when prog and me agreed to change the input scale as the internal thing is now a whole number, so it makes sense for the command line option as well.
Hadn’t announced yet, so why you take the bread out of the oven when it’s not ready
Just checking, the CPU utilization diff between -B narrow and -B wide used to be pretty small. Now wide uses about 2x more CPU than narrow. Working as intended? This is CPU util with -e 30 wide then narrow then wide on the just released binary.
Not much time yet, but signal change isn’t crazy
Traffic fluctuating so this looks normal for msgs.
Oops, I did that too maybe. I noticed the binary in github was 2 mins old when I started my lunch break at 19:15 UTC.
EDIT: looks like my post above was with the latest 19:13 posted binary. 
Reloaded v2.2-RC1. Now the preamble values are back to normal. But I am still seeing the auto preamble adjustment bouncing from 12 to 13 to 12 from 2 to 4 times a minute. See the log below.
OPTIONS= -v -f 1 -C 85 -e 11 -w 5 -b -t 60 -B narrow
SAMPLE_RATE= 20
Sep 03 15:38:29 piaware airspy_adsb[21011]: airspy_adsb v2.2-RC1
Sep 03 15:38:29 piaware airspy_adsb[21011]: Listening for beast clients on port 47787
Sep 03 15:38:29 piaware airspy_adsb[21011]: Acquired Airspy device with serial C2D067DC34944997
Sep 03 15:38:29 piaware airspy_adsb[21011]: Decoding started at 20 MSPS
Sep 03 15:38:30 piaware airspy_adsb[21011]: Push client connected to localhost:30004 (beast)
Sep 03 15:38:39 piaware airspy_adsb[21011]: CPU 75.9 %, target 85.0 %, adjusting preamble filter: 11 → 13
Sep 03 15:39:29 piaware airspy_adsb[21011]: Client connected from 127.0.0.1:33560 (beast)
Sep 03 15:39:39 piaware airspy_adsb[21011]: CPU 88.8 %, target 85.0 %, adjusting preamble filter: 13 → 12
Sep 03 15:40:08 piaware airspy_adsb[21011]: CPU 82.0 %, target 85.0 %, adjusting preamble filter: 12 → 13
Sep 03 15:40:18 piaware airspy_adsb[21011]: CPU 88.3 %, target 85.0 %, adjusting preamble filter: 13 → 12
Sep 03 15:40:28 piaware airspy_adsb[21011]: CPU 82.1 %, target 85.0 %, adjusting preamble filter: 12 → 13
Sep 03 15:40:48 piaware airspy_adsb[21011]: CPU 87.8 %, target 85.0 %, adjusting preamble filter: 13 → 12
Sep 03 15:42:18 piaware airspy_adsb[21011]: CPU 82.1 %, target 85.0 %, adjusting preamble filter: 12 → 13
Sep 03 15:42:28 piaware airspy_adsb[21011]: CPU 87.8 %, target 85.0 %, adjusting preamble filter: 13 → 12
Sep 03 15:42:38 piaware airspy_adsb[21011]: CPU 81.6 %, target 85.0 %, adjusting preamble filter: 12 → 13
Sep 03 15:42:48 piaware airspy_adsb[21011]: CPU 88.6 %, target 85.0 %, adjusting preamble filter: 13 → 12
1 Like
I think I’m seeing the same, even after running the update script again.
I haven’t changed my -E (42) from yesterday and now it’s reached it with the CPU 20% lower.
Sep 03 20:43:33 pi4-tracker airspy_adsb[14082]: airspy_adsb v2.2-RC1
Sep 03 20:43:33 pi4-tracker airspy_adsb[14082]: Listening for beast clients on port 47787
Sep 03 20:43:33 pi4-tracker airspy_adsb[14082]: Acquired Airspy device with serial 26A464DC28751693
Sep 03 20:43:33 pi4-tracker airspy_adsb[14082]: Decoding started at 20 MSPS
Sep 03 20:43:34 pi4-tracker airspy_adsb[14082]: Push client connected to localhost:30004 (beast)
Sep 03 20:43:43 pi4-tracker airspy_adsb[14082]: CPU 29.2 %, target 88.0 %, adjusting preamble filter: 2 -> 14
Sep 03 20:43:53 pi4-tracker airspy_adsb[14082]: CPU 44.1 %, target 88.0 %, adjusting preamble filter: 14 -> 23
Sep 03 20:44:03 pi4-tracker airspy_adsb[14082]: CPU 48.6 %, target 88.0 %, adjusting preamble filter: 23 -> 31
Sep 03 20:44:13 pi4-tracker airspy_adsb[14082]: CPU 52.0 %, target 88.0 %, adjusting preamble filter: 31 -> 38
Sep 03 20:44:23 pi4-tracker airspy_adsb[14082]: CPU 53.9 %, target 88.0 %, adjusting preamble filter: 38 -> 42
All day it’s been like this
Sep 03 09:34:05 pi4-tracker airspy_adsb[27774]: airspy_adsb v2.2-RC0
Sep 03 09:34:05 pi4-tracker airspy_adsb[27774]: Listening for beast clients on port 47787
Sep 03 09:34:05 pi4-tracker airspy_adsb[27774]: Acquired Airspy device with serial 26A464DC28751693
Sep 03 09:34:05 pi4-tracker airspy_adsb[27774]: Decoding started at 20 MSPS
Sep 03 09:34:06 pi4-tracker airspy_adsb[27774]: Push client connected to localhost:30004 (beast)
Sep 03 09:34:15 pi4-tracker airspy_adsb[27774]: CPU 26.5 %, target 88.0 %, adjusting preamble filter: 2.00 -> 13.06
Sep 03 09:34:25 pi4-tracker airspy_adsb[27774]: CPU 36.7 %, target 88.0 %, adjusting preamble filter: 13.06 -> 22.31
Sep 03 09:34:35 pi4-tracker airspy_adsb[27774]: CPU 50.2 %, target 88.0 %, adjusting preamble filter: 22.31 -> 29.12
Sep 03 09:34:45 pi4-tracker airspy_adsb[27774]: CPU 60.3 %, target 88.0 %, adjusting preamble filter: 29.12 -> 34.11
Sep 03 09:34:54 pi4-tracker airspy_adsb[27774]: CPU 67.9 %, target 88.0 %, adjusting preamble filter: 34.11 -> 37.72
Sep 03 09:35:04 pi4-tracker airspy_adsb[27774]: CPU 70.2 %, target 88.0 %, adjusting preamble filter: 37.72 -> 40.93
Sep 03 09:35:14 pi4-tracker airspy_adsb[27774]: CPU 73.7 %, target 88.0 %, adjusting preamble filter: 40.93 -> 42.00
Quite something. CPU utilization has fallen through floor without dropping message rate or signal strenghts. The -45.7 (the lowest value I see) has lot of AC now. At Gain 19.