***********************************************************************
                        WAV FILTER
***********************************************************************

WAV FILTER is set of non recursive filters. Filters are FIR type (Finite 
Impulse Response). This program also includes DFT (Discrete Fourier 
Transform - spectral analyse), but you need fast machine to perform DFT 
of larger data arrays. Generally, DFT is added as an analysing tool which 
helps you to chose filter frequencies. Filter order determents how many 
points of input buffer (file) is taken to calculate one output point. Higher 
order means higher quality of filtration, but it requires more time. Using 
of extremely high filter orders make sense when you want to perform 
extremely narrow band filtration (for example, you have sample 'polluted' 
with 50Hz and you want to cut out only very narrow band - for example 
from 40 - 60 Hz. In this case you have to set ORDER to at lest 2000 - 
4000 (need speedy machine)).  


Example DEMO.WAV:

This sample is mix of regular song and whistle. You can to separate this 
sample in two samples (song and whistle) by using WAV FILTER.
This is the on way how to do it:
1. Run WAV FILTER an OPEN  DEMO.WAV (file is going to be loaded 
in INPUT BUFFER)
2. Play the sample by selection PLAY INPUT BUFF in SOUND menu and 
try to determinate where is  whistle the most 'obvious'. (Let say this is 
at approximately  5th second from the beginning of the file).
3. You have to determinate frequency of whistle. Use DFT in ANALYSE 
menu ( chose 500-1500 (1500-2000 for fast machines) calculation points 
and start position at 5th second (please , if you enter this value in 
seconds put '*' as a first character in line)).
4. On the shown graph you can see spectral lines. Each line represents 
certain frequency in sample. Resolution is defined by number of 
calculation points and is displayed after DFT calculation - 'quantum'. 
UNIT is distance between two lines on axe. Now you can determinate 
whistle frequency ( peek on the right side/middle of DFT graph). Whistle 
frequency is approximately 1640 Hz.
5. Now, you can use BAND REJECT FILTER to remove whistle from 
sample. Chose filtration band from 1550 to 1750 Hz. Wait and play 
output buffer. If you want to 'hear' only whistle chose BAND PASS 
FILTER.
6. Try to use also HIGH and LOW pass filter to se the difference.

I hope you will like this program. If you think WAV FILTER is useful 
please register and send me $12 + $5 postage and package. If you 
think W_FILTER worth less than $12 send me amount you think I deserve (but,
please cover postage and disk costs - $5 (oversees mail)). To receive
source in C send additional $6.


					THANK YOU!
						      Denis

My address:
Denis Donlagic
Kersnikova 9
62000 Maribor
Slovenia
Evrope

E-mail: DONLAGIC@UNI-MB.SI 


  
