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

[linrad] RE: Main spectrum zero line

Hi Matthieu,

> What refers the zero line of the main spectrum to exactly?
> When the sound card is not connected, you see the noise floor of the sound
> card, lets say 10dB. So my question is :
> 10dB=10*LOG10(Noise_floor/????)

The zero line is arbitrary. You may shift it up or down with
the parameter "First FFT amplitude". The default value is 1000
and good soundcards (Soundblaster PCI or Delta 44) will have
their noise floors close to 0 dB with the default value.

In case you enable the second fft and select fft implementations
that use MMX instructions, the fft1 floating point output will be
converted to 16 bit integers. Setting the gain correctly then 
becomes critical in case your hardware has very high dynamic range.
This is described in the Jan 04 QEX article.

If you have a flat noise floor on screen you can read it's level
in bits by pressing "A" to get a number on the lowest line
on screen after "Floor". Quantisation noise is at about 0.3 bits
RMS. The flat noise floor can be placed at 3 bits for the 
noise floor inside the digital system to be 20 dB above the 
quantization noise. In case your hardware is really good, this
is necessary. Withsimpler hardware you may place the noise floor
anywhere within a rather wide interval. Other parameters 
determine how wide the interval is in which the noise floor 
can be placed without compromises. The whole thing is described
in QEX Jan 04.

Remember: Linrad is a tool box with which you can design
many different kinds of radio receivers and signal analysers.

Linrad is still in an early stage. I have spent most of my 
time on hardware problems during the latest two years. 
The most important problems to solve are not within the Linrad 
software. The most difficult one right now is to make amateurs 
understand why their transmitters are so bad and emit so 
much splatter and keying clicks. It is the awful mis-use 
of ALC to limit the output power as a "syllabic compressor". 
Writing articles and collecting data as well as making the 
WSE converters available so anyone interested easily can check 
the performance of his own and his friends transmitters takes 
most of my time. That is why Linrad develops so slowly.

I will leave for London in a few hours. Probably I will
loose most of my incoming E-mails thursday to saturday because
all the spam will probably fill my mailbox pretty quickly:(


Leif / SM5BSZ