[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Discuss-gnuradio] FM pre-emphasis and de-emphasis
From: |
Richard Lamont |
Subject: |
Re: [Discuss-gnuradio] FM pre-emphasis and de-emphasis |
Date: |
Wed, 10 Jan 2018 20:41:06 +0000 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.5.0 |
On 10/01/18 18:53, Andy Walls wrote:
[snip]
>> When working correctly, the amplitude/frequency response should be
>> within, ideally, 0.1 dB of the values shown. I've shown values for
>> both 75 and 50 microsecond time constants.
>>
>> Freq (Hz) Gain dB (75us) Gain dB (50us)
>>
>> 30 0.00 0.00
>> 300 -0.09 -0.04
>> 500 -0.23 -0.11
>> 1000 -0.87 -0.41
>> 3000 -4.77 -2.76
>> 5000 -8.16 -5.40
>> 8000 -11.82 -8.64
>> 10000 -13.65 -10.36
>> 12500 -15.52 -12.15
>> 15000 -17.07 -13.65
>>
>
> FWIW, the attached plots compare the 50 us de-emphasis filter response
> of GNURadio 3.7.10 vs. the numbers you provided.
FWIW they are not my numbers, but appeared to come from a reliable source.
> The filters stay within 0.1 dB of each other up through 5000 Hz, but by
> then your tabular response has noticeably started to bend upward. The
> difference between the two filters is 3.44 dB at 15000 Hz.
It bends up because your plot has got one log axis and one linear one.
If plotted with a log frequency axis rather than a linear one then the
slope at the right should tend to a straight line of 6.02 dB/octave or
20 dB/decade. Indeed the idea of dB/octave or dB/decade only makes sense
if you use a log frequency axis.
Alternatively, you could use a linear voltage axis instead of dB, and a
linear frequency axis. That should also tend towards a straight line on
the right hand side.
--
Richard Lamont