Re: [Discuss-gnuradio] QT Freqency Sink vs. Log Power FFT values

[John Ackermann N8UR]

Thanks!

Interesting that the default value shown in the box is "2" not "2.0" -- 
that would have given the clue.


I tried setting the scaling value to 4.0 and that seems to align very 
nicely with the Frequency Sink results.


John

On 08/06/2017 07:21 PM, Kevin Reid wrote:
On Sun, Aug 6, 2017 at 1:57 PM, John Ackermann N8UR > wrote:


I also have a QT Frequency Sink as eye candy, and notice that its
display is about 6dB different than the Log Power FFT output.  I can
think of a number of reasons why that could be, but my simple
question is... is there an appropriate way to adjust one or the
other to align their readings?  (Later I'll send the log power
results into a calibration routine to convert to dBm based on
measured results.)

Also while I'm at it, there's a "Reference Scale" parameter in the
Log Power FFT, set by default to "2".  Is there a way to use this
parameter to align the two readings?  I tried changing it to "1" but
that threw an error at runtime.


I took a look at logpwrfft's code and the first thing that is done to 
that number is ref_scale/2, so you need to pass in a float (1.0) rather 
than an integer to not get it truncated to zero. Sounds like something 
to fix.


Also, there doesn't appear to be any window applied (or at least
apparent) in the Log Power FFT block.  Is there a way to window the
output?


The GRC definition for the block doesn't let you set it but there 
actually is a parameter win which defaults to Blackman-Harris.


___
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio

Re: [Discuss-gnuradio] QT Freqency Sink vs. Log Power FFT values

[Marcus Müller]

Hi Kevin, Hi John,


On 08/07/2017 01:21 AM, Kevin Reid wrote:
On Sun, Aug 6, 2017 at 1:57 PM, John Ackermann N8UR > wrote:


I took a look at logpwrfft's code and the first thing that is done to 
that number is ref_scale/2, so you need to pass in a float (1.0) 
rather than an integer to not get it truncated to zero. Sounds like 
something to fix.



agreed!

https://github.com/gnuradio/gnuradio/pull/1415

Only made this against master, not maint, as it *does* change behaviour. 
Not sure. If wanted, I can do against maint, too. Maintainers?


Cheers,
Marcus
___
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio

Re: [Discuss-gnuradio] QT Freqency Sink vs. Log Power FFT values

[Kevin Reid]

On Sun, Aug 6, 2017 at 1:57 PM, John Ackermann N8UR  wrote:

> I also have a QT Frequency Sink as eye candy, and notice that its display
> is about 6dB different than the Log Power FFT output.  I can think of a
> number of reasons why that could be, but my simple question is... is there
> an appropriate way to adjust one or the other to align their readings?
> (Later I'll send the log power results into a calibration routine to
> convert to dBm based on measured results.)
>
> Also while I'm at it, there's a "Reference Scale" parameter in the Log
> Power FFT, set by default to "2".  Is there a way to use this parameter to
> align the two readings?  I tried changing it to "1" but that threw an error
> at runtime.
>

I took a look at logpwrfft's code and the first thing that is done to that
number is ref_scale/2, so you need to pass in a float (1.0) rather than an
integer to not get it truncated to zero. Sounds like something to fix.


> Also, there doesn't appear to be any window applied (or at least apparent)
> in the Log Power FFT block.  Is there a way to window the output?
>

The GRC definition for the block doesn't let you set it but there actually
is a parameter win which defaults to Blackman-Harris.
___
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio

[Discuss-gnuradio] QT Freqency Sink vs. Log Power FFT values

[John Ackermann N8UR]

I'm doing a simple flowgraph to determine the peak value (dBFS) within 
an FFT.  The Log Power FFT sends vectors to a Max block which outputs 
the result -- it seems to be working reasonably well.


I also have a QT Frequency Sink as eye candy, and notice that its 
display is about 6dB different than the Log Power FFT output.  I can 
think of a number of reasons why that could be, but my simple question 
is... is there an appropriate way to adjust one or the other to align 
their readings?  (Later I'll send the log power results into a 
calibration routine to convert to dBm based on measured results.)


Also while I'm at it, there's a "Reference Scale" parameter in the Log 
Power FFT, set by default to "2".  Is there a way to use this parameter 
to align the two readings?  I tried changing it to "1" but that threw an 
error at runtime.


Also, there doesn't appear to be any window applied (or at least 
apparent) in the Log Power FFT block.  Is there a way to window the output?


Thanks,
John

___
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio