aboutsummaryrefslogtreecommitdiffstats
path: root/README.md
blob: d150658f77852dcecb0affbb8c3333bbb9abac3c (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
CIR scripts
===========

This is a set of script to do correlation measurement measurements
of a DAB signal. These scripts can be used to do channel impulse
response measurements using a RTLSDR receiver.

Right now there are three scripts:

* correlate_with_ref.py: Finds the NULL symbol of an IQ file, and runs
  correlations against the known phase reference to find the components.

* simulate_channel.py: Reads an I/Q file generated by ODR-DabMod and
  adds some additional components, somehow equivalent to signal reflexions.
  Each reflexion has a delay and an amplitude factor.

* cir_measure.py: Runs a small webserver that does the correlation
  all the time and updates a webpage

Example image: ![Example correlation](./example_corr.png)

The image shows the received CIR of a real DAB signal captured with the RTLSDR.

Webserver
---------

To run the webserver on all network interfaces, receiving on 202.928MHz:

    ./cir_measure.py --host 0.0.0.0 --port 8080 --freq 202928000

More options are available (gain, number of samples), see the help with

    ./cir_measure.py -h

Offline processing
------------------

To read in a recording using a RTL-SDR receiver, use:

    rtl_sdr -f 188928000 -s 2048000 -g 20 -S - | \
    dd of=DAB_9A_10_u8_G20.iq bs=2 count=1966080

replace 1966080 (ten Mode 1 transmission frames) by a larger value if you want.

Then feed the data to the script

    ./correlate_with_ref.py u8 DAB_9A_10_u8_G20.iq


Requirements
============
librtlsdr and its tools.

Python with NumPy and matplotlib.

The iq files must be complex float or interleaved unsigned 8-bit.


Licence
=======
MIT. See LICENCE for details.

TODO
====
Improve the sync code, to make sure we align the pictures each time. Align on
the strongest peak instead of the NULL symbol.

Generate long-term visualisations, to see if there is a change over several
days.

Oversample 2x or 4x to get finer time resolution.

Make sure rtl_tcp quits properly on ctrl-c.