[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Discuss-gnuradio] RA:FFT

From: Marcus D. Leech
Subject: Re: [Discuss-gnuradio] RA:FFT
Date: Tue, 29 Dec 2015 15:34:42 -0500
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.7.0

On 12/29/2015 03:29 PM, Simon Olvhammar wrote:

Ok, thanks.
So it removes fft frames at the output, then frames are lost? What am I missing?
A more concrete question, does e.g. a 10 min measurement contain the same fft data with no respectively with decimation?

Could you explain more in detail how input/output fft frame flow look like, with say decimation at 100?

Input time-domain data are presented to the FFT in FFT-size chunks, the FFT output is integrated with a single-pole-IIR filter, and then
  decimated according to decimation.  So if you have 1000 integrated FFT outputs, and decimation is set to 100, you only see 10 of
  them, but those 10 are post-integrator.

Let's say your input rate is 1.0e6, and your FFT size is 1000.  That means a new FFT every 1msec.  If that is then decimated after
  integration with a decimation rate of 100, then you get one integrated FFT output every 100msec.

Den 29 dec 2015 19:31 skrev "Marcus D. Leech" <address@hidden>:
On 12/29/2015 01:01 PM, Simon Olvhammar wrote:

I have a question regarding the block RA:FFT in the RA Blocks package.
If I set, for example, the decimation at 100 does this mean that only 1 in 100 items will be kept?
In that case it would occur to me if, I do say a measurement of 100 minutes at a signal, I only get 1 minute of data?

Best regards

Discuss-gnuradio mailing list
No, the decimation is applied to FFT *output* frames, after integration.   It is typically the case that in radio astronomy spectrometry
  that you integrate spectra over some time period, since the SNR is typically quite poor.

Discuss-gnuradio mailing list

reply via email to

[Prev in Thread] Current Thread [Next in Thread]