Date: Wed, 7 Jun 2017 14:56:31 -0400
From: Markus Vester <markusvester@aol.com>
To: rsgb_lf_group@blacksheep.org
Message-Id: <15c83ea53fb-5674-3a681@webprd-a27.mail.aol.com>
In-Reply-To: <59382E2E.5000306@posteo.de>
MIME-Version: 1.0
Subject: Re: LF: Bugfix for ebnaut_ifft utility
Content-Type: multipart/alternative; 
	boundary="----=_Part_292095_1461638698.1496861791224"
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Status: RO

------=_Part_292095_1461638698.1496861791224
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

Hi Stefan,

this is not easy to say as it depends on a number of circumstances. In your=
 local test, the difference was quite large (EbN0 =3D 40.25 dB vs 41.57 dB)=
, but this is because the lack of external noise emphasized the relatively =
small artifact on the signal. In a more realistic scenario the individual s=
ymbols are buried in the noise. Considering an EbNaut sequence with e.g. N=
=3D1000 symbols, the signal energy is quasi-randomly spread over 1000 frequ=
ency bins. Only one of these bins would be affected or "lost", suggesting a=
n SNR loss on the order of 0.1% power or  0.004 dB.

Best 73,
Markus

-----Urspr=C3=BCngliche Mitteilung-----=20
Von: DK7FC <selberdenken@posteo.de>
An: rsgb_lf_group <rsgb_lf_group@blacksheep.org>
Verschickt: Mi, 7. Jun 2017 18:49
Betreff: Re: LF: Bugfix for ebnaut_ifft utility


Thanks Markus,=20

 What does minor SNR degradation mean, in dB, and on what does this value d=
epend? (e.g. transmission time, symbol length, etc).

 73, Stefan

 Am 07.06.2017 17:59, schrieb Markus Vester:=20


Dear EbNauteers,=20

I recently stumbled upon a bug in my ebnaut_ifft3a utility, which is used t=
o convert SpecLab FFT export files to IQ wav files. The first step of the i=
nverse FFT is to take the complex conjugate of the data array. Due to the b=
ug, the respective loop was executed only N-1 times, leaving a sign error i=
n the imaginary part of a single bin next to the center frequency. When fed=
 with a high-SNR signal close to SpecLab's center frequency, this can produ=
ce strange undulation patterns in the phase display and a minor SNR degrada=
tion for the decoder.

The bug has been fixed in version 3b, available on
  http://df6nm.bplaced.net/VLF/fec_tests/ebnaut_ifft3b.zip

 Best regards,
 Markus (DF6NM)
------=_Part_292095_1461638698.1496861791224
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable

<div style=3D"font-size:10pt;font-family:arial,helvetica,sans-serif;color:b=
lack;"><div style=3D"color: black; font-family: arial,helvetica,sans-serif;=
 font-size: 10pt;">Hi Stefan,<br><br>this is not easy to say as it depends =
on a number of circumstances. In your local test, the difference was quite =
large (EbN0 =3D 40.25 dB vs 41.57 dB), but&nbsp;this is because&nbsp;the la=
ck of external noise emphasized the&nbsp;relatively small artifact on the s=
ignal.&nbsp;In a more realistic scenario the individual symbols are buried =
in the noise.&nbsp;Considering an EbNaut sequence with e.g. N=3D1000 symbol=
s, the signal energy is quasi-randomly spread over&nbsp;1000 frequency bins=
. Only&nbsp;one of these bins would be affected or "lost", suggesting an SN=
R loss on the order of 0.1% power or&nbsp; 0.004 dB.<br><br>Best 73,<br>Mar=
kus<br><br>-----Urspr=C3=BCngliche Mitteilung----- <br>Von: DK7FC &lt;selbe=
rdenken@posteo.de&gt;<br>An: rsgb_lf_group &lt;rsgb_lf_group@blacksheep.org=
&gt;<br>Verschickt: Mi, 7. Jun 2017 18:49<br>Betreff: Re: LF: Bugfix for eb=
naut_ifft utility<br><br></div>
<div id=3D"AOLMsgPart_2_3c6aca1a-65f5-47cc-8497-454dba059609">
<div class=3D"aolReplacedBody" text=3D"#000000" bgcolor=3D"#ffffff">
Thanks Markus, <br>
<br>
What does minor SNR degradation mean, in dB, and on what does this
value depend? (e.g. transmission time, symbol length, etc).<br>
<br>
73, Stefan<br>
<br>
Am 07.06.2017 17:59, schrieb Markus Vester:
<blockquote cite=3D"about:blank">
  <div style=3D"color: black; font-family: arial,helvetica,sans-serif; font=
-size: 10pt;">
  <p>Dear EbNauteers, </p>
  <p>I recently stumbled upon a bug in my ebnaut_ifft3a utility, which
is used to convert SpecLab FFT export files to IQ wav files. The first
step of the inverse FFT is to take the complex conjugate of the data
array. Due to the bug, the respective loop was executed only N-1 times,
leaving a sign error in the imaginary part of a single bin next to the
center frequency. When fed with a high-SNR signal close to SpecLab's
center frequency, this can produce strange&nbsp;undulation patterns in the
phase display and a minor SNR degradation for the decoder.</p>
  <p>The bug has been fixed in version 3b, available on<br>
&nbsp;<a class=3D"aolmail_moz-txt-link-freetext" href=3D"http://df6nm.bplac=
ed.net/VLF/fec_tests/ebnaut_ifft3b.zip" target=3D"_blank" rel=3D"noopener n=
oreferrer">http://df6nm.bplaced.net/VLF/fec_tests/ebnaut_ifft3b.zip</a><br>
  <br>
Best regards,<br>
Markus (DF6NM)<br>
  </p>
  </div>
</blockquote>


</div>
</div></div>
------=_Part_292095_1461638698.1496861791224--