Re: [Discuss-gnuradio] Problems with dvbt 8k 64QAM

[Ron Economos]

A .cfile is a binary file with 32-bit floating point numbers 
representing IQ values. It's pretty much a GNU Radio thing and means 
"complex file". The reason you're seeing zeros in every floating point 
number is that your capturing from a device with only 8 bits of 
precision. But the test files were generated from the GNU Radio DVB-T 
transmitter with full floating point precision.

The 8-bit samples get normalized to +/- 1, so the floating point values 
can only be 1/128, 2/128, etc. The floating point representation of 
1/128 is 00 00 00 3C, 2/128 is 00 00 80 3C, 3/128 is 00 00 C0 3C and so on.

Ron

On 03/19/2018 05:46 AM, Juan Antonio wrote:
> Hi  Ron
>
> The signal I am using has been measured with an electromagnetic field 
> meter. Power, BER, MER and impulse response have values well above the 
> average
>
>
>
> I have been analyzing the files with baudline and there is a curious 
> thing in the  bitmaps
>
> While adv16.cfile and adv64.cfile have the format:
>
> xxxx xxxx xxxx xxxx xxxx xxxx xxxx xxxx
>
> My files have the format:
>
> 0000 xxxx 0000 xxxx 0000 xxxx 0000 xxxx
>
> The x, of course, is a hexadecimal value
>
> Maybe that's the problem, I've been looking for information about 
> cfile files but I have not found almost anything. If you want I can 
> send you some file recorded by me with gqrx or with osmosdr and filesink
>
>
>
> Best regards