The log file produce by my Taranis is already a .csv file but I still want to be able to produce a .gpx file with the tool for displaying gps data on google earth.

But I can't figure how to make it work as the tool requires a .txt file ...

Nick,
Is there any tools available to convert the current blackbox log to something Dashware can read? I couldn't get the old tools to open.

This is from the FAQ on dashware.net
I want to make my own data file. How can I make it work with DashWare?
Here are some notes on making a DashWare-friendly CSV data file:

-It’s good to have a static line of text at the top declaring that it is your data format. Something like “Some Cool Logger Data File”. The DashWare Data Profile for your format can then be easily set up to automatically detect one of your data files.
-We require a CSV header line: A line of text containing the column headers/names.
-Then the raw data lines would follow after that.
-There should be a Time column (full timestamps work great, e.g. “2011/08/07 20:10:02.200”, or the total running time in seconds e.g. “1.1”)
-Latitude and Longitude should be in +/- decimal degrees
-Be consistent with your data units, or at least change the column name when you use different units. DashWare can be used to create calculated data columns for other units.

You will then need to make a Data Profile that understands your format. Bring up the Data Profile Editor (File Menu > Edit Data Profiles…) and press F1 for detailed help. Tips:
-The column mappings (bottom right area of dialog) are used to map your column names to known DashWare data types. Set up the mapping for your time column first, so you can use the “Test File / View Data…” button. It is very handy. “ / Data Running Time, Seconds” is the mapping you need.
-To get a map to show up, you need to use our “Degrees to Cartesian Converter” (calculators/converters are in the bottom left area of dialog) to create X and Y units in feet. You must map these new X and Y output columns to our “Positional \ Longitude / X Position (Feet)” and “Positional \ Latitude / Y Position (Feet)” data types.

Have a look at an existing data profile, such as the GPX one, as an example.

if one does call blacckbox_render with for example --fps=49.3 or --fps49,3 the result is the same as with --fps=49

could this be fixed? without this I need to make my video editing software create additional frames (or drop some if I instead would generate the blackbox pngs with 50 fps) which reduces the smoothness of the flow of the frames..

When decoding log files from blackbox_decode it would be great if the PID tuning values used during that log were output to a text file (rather than having to keep a manual log for optimising tuning). I think they're already decoded as part of blackbox_render but that makes it difficult to batch-analyse files

@thenickdude : What do you think about adding event header to blackbox protocol? It should be possible to reuse code for I-Frame, maybe without predictor

H Event 20 ID:Gtune result
H Event 20 name:axis,gyroP,newP
H Event 20 signed:0,1,1
H Event 20 predictor:0,0,0
H Event 20 encoding:0,0,0

And either add new encoding or use VB for log events ...

blackbox_decode does not compute RC rate and PID error (for anyone interested in doing offline analyses on these). My current workaround is to load the log in explorer and save an csv file from there which contains these variables, so presumably they are computed in explorer. I am trying to do offline Matlab analyses on PID error, and it'd be less cumbersome, and easier to implement within Matlab scripts, if blackbox_decode did this.

Add support for decoding current sensor data in blackbox_render and blackbox_decode

Change icon to different aircraft (at least the most common for inav: wing, plane) to match controls (1/2 motors and 2 to 4 servos)
It would be nice if naming will be changed on graph to, at least from motors to servos, even for motor channel

Wing icon as example with 1 motor and 2 servos. Servos should be not circles, but deflection up/down

I have attempted here:
jelinj8@f2e005d
but that just prints the date to stderr.

suggestions:

• Naming output files
• fixed GPX output timestamps
• optional timestamp column in output (e.g. for DashWare telemetry rendering)

My C++ is too rusty for that.

Header lines look like this:
H Firmware revision:INAV 1.9.1 (03a5c1922) MATEKF405
H Firmware date:Apr 21 2018 13:25:29
H Log start datetime:2018-05-04T18:17:08.376+00:00

Trouble is that the filenames (and files) are created before parsing the log. Probably renaming (replacing a placeholder) the files before memfreeing the filenames would be the simplest solution?

Hi,

I'm wondering how the raw data from the gyro's is converted to rad/s and then deg/s

I use 'blackbox_decode' for retrieving GPS tracks from iNav logs.
Sometimes it happens that a single file contains more than 31 logs.
For this reason I increased constant FLIGHT_LOG_MAX_LOGS_IN_FILE.
Is there any reason why is it limited now? Could it be increased in next release?

Thank you!

In gpx decoding the file only contains the relative timestamp from the beginning of logging. When there was a GPS Fix on Arm there is an absolute timestamp in the log in the ASCII Part (At least with inav)

Use that timestamp to correct gpx timestamps to be able to use the gpx to correlate with camera images with "gpxcorrelate".

Header of blackbox log file:

[...]
H Field S encoding:1,1,7,7,7,1
H Firmware type:Cleanflight
H Firmware revision:INAV 1.9.1 (03a5c1922) OMNIBUSF4PRO_LEDSTRIPM5
H Firmware date:Apr 21 2018 13:15:03
H Log start datetime:2018-06-16T13:10:49.264+00:00
H Craft name:
H P interval:1/1
[ ... ]

Only relative timestamps in gpx file:

<trk><name>Blackbox flight log</name><trkseg>
  <trkpt lat="52.3008762" lon="8.5960688"><ele>57</ele><time>2000-01-01T00:02:02.389080Z</time>

i got this error on 2 different logs, the decode had no errors.
other index on the same log worked.
why do i get it and how i prevent it in the future?

(i use a mac-mini with osx 10.10.4 and the tools 0.4.1)

Select graphs to draw by name. Group it. ((Name, data_name, data_name), (Name, data_name, data_name), (Name, data_name, data_name))

Something like (("Attitude", "atttitude[0], "attitude[1]"), ("RC Input", "RC Command [roll]", "RC Command [pitch]"))
In this case we will have two graphs (Attitude and RC Input). On first one w will have two lines ('atttitude[0]' and 'attitude[1]'), on second one two lines too (RC Command [roll], RC Command [pitch])

With such option we can draw any kind of graph we need for any king of aircraft

Hi,

I have just upgraded my setup to betaflight 3.2.2, and now the blackbox-tools all seem to deadlock on my blackbox logs.
These logs were created on the SD card in an SPRacing F3 NEO board, used to work nicely with cleanflight in the past.

Even with the raw and debug-parameters the software just deadlocks and uses 100% CPU.
blackbox_render, blackbox_decode and the Chromium-webapp all show this behaviour.

https://screenshot.tbspace.de/xawdoshiczy.png
https://tbspace.de/content/downloads/LOG00008.BFL

Any ideas on what could be wrong with the new files? (i have 8 of these here, happens with all of them).

Thanks,
Tobias

I'm having trouble syncing HD Video to the BlackBox File. Is there any way to simplify this? I wish there was a way to movie the video separately and the BlackBox Log File separately so I can just sync them that way.

In blackbox_decode the acceleration out in done in controller coordinate system. It would be useful to have possibility to output accelerations in the Earth coordinate system, with gravity removed (az-1g).
It is done in what way in blackbox_render, so it should be quite simply.
All functions already exists in the code blackbox_render.
It is just a matter of output these values to .csv file by blackbox_decode.

Proposed command line switch for blackbox_decode:
--acceleration-csys acceleration output coordinate system (controller|Earth), default is controller

When i click on the application, the window pops up but then goes away, not sure what going on here

While converting *.BFL file to *.csv in the headers there are units of barometer (cm), rotational velocity in ex. (deg/s) and acceleration (m/s/s) if the units were selected by command line switches. But there in no command line switch to select units of attitude provided when --simulate-imu is activated.

The orientation angles are not raw values,, are in (deg). To have consistent file, the units of angle should be added to roll, pitch and yaw in headers

It is now:
accSmooth[0] (m/s/s) accSmooth[1] (m/s/s) accSmooth[2] (m/s/s) roll pitch heading

It should be:
accSmooth[0] (m/s/s) accSmooth[1] (m/s/s) accSmooth[2] (m/s/s) roll(deg) pitch(deg) heading(deg)

Hi all, I am trying convert the decoded blackbox CSV data into actual conventional sensor readings that we are viewing on Blackbox explorer for my own analysis. Would like to ask if there are any guidelines on how I can do this?

Thanks in advance!

I just tried the luxfloat PID controller and the blackbox-log of that flight can be decoded just fine, but the rendering shows a unchanging PID table, unchanging command inputs and missing all the lines of motor outputs and gyro.

blackbox_decode works just fine and produces a csv document that looks all right.

kaefert@mint ~/Downloads/quadcopter_stuff/blackbox-log/2015-08-09_1110 try luxfloat PID $ ~/Downloads/quadcopter_stuff/blackbox-log/blackbox-tools/obj/black
box_decode LOG00138.TXT 
Decoding log 'LOG00138.TXT' to 'LOG00138.01.csv'...

Log 1 of 1, start 43:28.366, end 45:53.195, duration 02:24.828

Statistics
Looptime           3504 avg           28.3 std dev (0.8%)
I frames    1292   51.2 bytes avg    66199 bytes total
P frames   40029   29.4 bytes avg  1175762 bytes total
H frames      12   11.0 bytes avg      132 bytes total
G frames     915   11.5 bytes avg    10523 bytes total
E frames       2    9.5 bytes avg       19 bytes total
Frames     41321   30.1 bytes avg  1241961 bytes total
Data rate  285Hz   8659 bytes/s      86600 baud

1 frames failed to decode, rendering 4 loop iterations unreadable. 6 iterations are missing in total (21ms, 0.01%)

kaefert@mint ~/Downloads/quadcopter_stuff/blackbox-log/2015-08-09_1110 try luxfloat PID $ ~/Downloads/quadcopter_stuff/blackbox-log/blackbox-tools/obj/black
box_render LOG00138.TXT 
4355 frames to be rendered at 30 FPS [2:26]

Rendered 500 frames (11.5%)...
Rendered 1000 frames (23.0%)...
Rendered 1500 frames (34.4%)...
Rendered 2000 frames (45.9%)...
Rendered 2500 frames (57.4%)...
Rendered 3000 frames (68.9%)...
Rendered 3500 frames (80.4%)...
Rendered 4000 frames (91.8%)...
Rendered 4355 frames (100.0%).

would be very useful to have the current PID settings as a renderable item for the overlay. will be very helpful for those that use in flight adjustments to tune.

Hi all, I am running a Lumenier F7 dual gyro f7 controller with the Matek F422SE firmware on INAV. I recorded a flight using the blackbox logger with a Beitian 880 GPS and compass module and tried to view the log. The explorer is unable to parse the data and returns the following error:
Missing field encoding header for field #0 'GPS_fixType'

WIth Blackbox decode tools I was able to get three files out, but they are very short and do not contain the logs.

Hi,

I'm developing a bot that uses blackbox decode and the performance is really poor under this system. I understand that the HW used is not the best for this task, and using best hw will improve the way it works.

As i see on htop, the software is just using one core to perform the actions, I don't know if it will be possible to do something to improve runtime:

here is a capture of the htop command showing the cpu usage:

And the task itself:

as you can see, PID task 10052 is been running now for almost half an hour, when on a windows 10 system (it takes about 1 or 2 minutes top).

Is anything I can do? I've downloaded and compiled the code on the raspberry pi 3. Log file:

This is de log that i'm using to test it:

btfl_004.zip

Thanks.

A few more options would be nice:

• Hide "Batt. cell" (bottom left corner)
• Hide "Acceleration" (bottom left corner)
• Change colors

It would be neat to have a 3 axis G meter. A peak hold indicator would be nice also. The most I have hit was 2.8 G's with KK2.0, F330. Thanks guys, BB is great!