Re: [Discuss-gnuradio] GRCon15 Talk recordings

[Andy Walls]

> Message: 2
> Date: Sat, 29 Aug 2015 17:30:29 -0400
> From: Sylvain Munaut
> To: GNURadio Discussion List <address@hidden>
> Subject: [Discuss-gnuradio] GRCon15 Talk recordings

> Hi,
> 
> 
> For anyone interested into digging into this, here are some recordings
> of the mic RF signal during talks at GRCon 15
> 
> http://people.osmocom.org/~tnt/grcon/mic-f5.030000e%2b08-s2.000000e%2b06-t20150826090018.cfile
> http://people.osmocom.org/~tnt/grcon/mic-f5.030000e%2b08-s2.000000e%2b06-t20150826090030.cfile
> 
> (Hehe, I bet you didn't expect that when you clicked on that email title :p)
> 
> 
> Microphone system is from Shure, their ULXD line of products (
> http://www.shure.com/americas/products/wireless-systems/ulxd-systems )
> FCC filing for the microphone that was used: https://fcc.io/DD4ULXD1G50
> 
> Modulation is 8-PSK and symbol rate is 156250 sym/s

This brief may confirm a lot of what you already figured out:
http://cdn.shure.com/uploaded_file/upload/80/Advanced_Wireless_Seminar_July_2013.pdf
http://cdn.shure.com/uploaded_file/upload/118/2013-11-19_Advanced_Wireless_Seminar_Portland_OR.pdf
http://cdn.shure.com/uploaded_file/upload/174/advanced_wireless_seminar_may_2014_english.pdf
Slide 70 or 72, in particular, shows the symbol values for a Gray Coded
8-PSK constellation.


8-PSK at 156,250 symbols/sec = 1.25 Mbps.
At 44.1 ksps audio, one could still fit 24-bit uncompressed audio
samples in 1.25 Mbps and still keep up with some room still for
overhead.  I'm going to speculate whatever audio coding Shure is using,
it is compressing audio to use 16 bits per sample for over the air.

Also, Shure mentions in a few places that their audio codec has 2.9 ms
latency.  On page 26 of the PDF of this journal publication:
http://www.aes.org/journal/sample_issue/JAES_V50_11_ALL.pdf
it notes that 128 samples / 44.1 ksamples/sec = 2.9025 ms.

So I'm going to speculate that Shure's audio codec operates on blocks of
samples that are N*128 samples long.  That payload block size might be
visible in OTA transmissions.  It may be easiest to spot with audio
near-silence samples that are at the end of a transmission (when the VOX
decides to cut off).

I have not looked at the files BTW.

Regards,
Andy

> 
> Cheers,
> 
>     Sylvain
>