Slow FFT in ksh93 and awk. Post: 303023480

Sponsored Content

Operating Systems OS X (Apple) Slow FFT in ksh93 and awk. Post 303023480 by wisecracker on Wednesday 19th of September 2018 01:19:42 AM

09-19-2018

Registered User

Quote:

Originally Posted by Corona688

Absolutely. I don't post code for people to not use it.

It's a kind of naive implementation though. Usually better to read values from stdin, but knowing how audioscope works I knew you'd need batches. That'd obscure how it really works a bit though. Maybe better as this, people can populate the arrays as they please.

It's not truly an FFT either, just a DFT -- which gives the same result but takes much more time to do so. Notice how it must do N*N loops for N input. FFT accomplishes the same with N log N loops.

Still, I hope it has its uses. And I'm grateful for you writing one in shell. That's the first DFT I've seen written in a language I really understand.

Yes it is technically a DFT but most people don't know what the difference is and FFT is what they know most, so I used that term. I will upload to AMINET as a DFT however.
And naive it is too, but I could stay with python and use the builtin import cmath version I have got to eliminate the 'scipy' and 'scipyIO' dependencies. I really didn't want to use python although my code works from Python 1.4.0 to the current 3.7.0.
Thanks a lot matey, will post the URL for you to see when I have uploaded it...
Bazza...

------ Post updated 19-09-18 at 06:19 AM ------

Hi Corona688...

Uploaded to the AMINET site at around 9:30pm UK local time last night, 18-09-2018.

At around 6:00am today 19-09-2018 UK local time it was added and has had 2 downloads, dls, already. It will be on show for 14 days and I expect at least 150 dls in that time.

Aminet - dev/gcc/DFT-FFT.awk.txt

Thanks a lot...

Bazza.

(Consider this thread wrapped up.)

I have no idea why it added my separate post to the end of this one...

This User Gave Thanks to wisecracker For This Post:

wisecracker

View Public Profile for wisecracker

Find all posts by wisecracker

10 More Discussions You Might Find Interesting

1. UNIX for Dummies Questions & Answers

echo is too slow. HELP with Awk

Hello All, Below is a simple script i worte to find the 208th char in a file. If the char = "C" then I re-direct the line to a file called change.txt. If it is not "C" then I re-direct it to a file called delete.txt. My problem is I have a file 0f 500K lines. this script is very slow. I am...

2. Shell Programming and Scripting

ksh93 deprecation...

Any means of running ksh93 in a ksh88-mode? Might sound odd, but I want/need to restrict U/Win-developed scripts to correspond to the ksh88 version on my Solaris environment(s). Thanks.

3. Shell Programming and Scripting

ksh88 or ksh93

Hi all! Does anybody know how can I check if any UNIX installation has implemented ksh88 or ksh93? Thanks in advance. N�stor.

4. Shell Programming and Scripting

The builtin split function in AWK is too slow

I have a text file that contains 4 million lines, each line contains 2 fields(colon as field separator). as shown: 123:444,555,666,777,888,345 233:5444,555,666,777,888,345 623:454,585,664,773,888,345 ...... Here I have to split the second field(can be up to 40,000 fields) by comma into an...

5. Shell Programming and Scripting

cut, sed, awk too slow to retrieve line - other options?

Hi, I have a script that, basically, has two input files of this type: file1 key1=value1_1_1 key2=value1_2_1 key4=value1_4_1 ... file2 key2=value2_2_1 key2=value2_2_2 key3=value2_3_1 key4=value2_4_1 ... My files are 10k lines big each (approx). The keys are strings that don't...

6. Shell Programming and Scripting

Making a faster alternative to a slow awk command

Hi, I have a large number of input files with two columns of numbers. For example: 83 1453 99 3255 99 8482 99 7372 83 175 I only wish to retain lines where the numbers fullfil two requirements. E.g: =83 1000<=<=2000 To do this I use the following...

7. UNIX for Advanced & Expert Users

Ksh93 on Linux compatible with ksh93 on AIX

Hi Experts, I have several shell scripts that have been developed on a Linux box for korn ksh93. If we want to run this software on an AIX 6.1 box that runs ksh88 by default can we just change the she-bang line to reference /bin/ksh93 which ships with AIX as its "enhanced shell" to ensure...

8. OS X (Apple)

FFT for the AMIGA through ksh88 shell.

I don't know if anyone is interested but I have been meddling with FFT for the AMIGA. (Sadly we AMIGAns don't have these luxuries through any scripting language. Below is a Python snippet that uses the builtin 'cmath' module to work with the lowly Python 2.0.1 for the AMIGA. It is part of a...

9. AIX

Ksh93/AIX compatibility

Hi everyone ! Im trying to know from wich version of AIX KSH93 is available ? Internet tell me 6.x and 7.x AIX are available, bue what about 5.x ? Is KSH93 available on AIX 5.x ? Is it the same way to manipulate variables as KSH93 on 7.x ? Thanks for your support and have a nice day !

10. OS X (Apple)

FFT for Python 2.0.x to 3.7.0.

Hi guys... This is code that was originally designed to work on an upgraded AMIGA A1200 using Python 2.0.x. Unfortunately it broke inside much later versions, NOT because of the print statement/function but other minor subtleties. So this is the final result tested on various machines including...

LEARN ABOUT DEBIAN

gr_plot_psd_f

GR_PLOT_PSD_F(1)						   User Commands						  GR_PLOT_PSD_F(1)

NAME

       gr_plot_psd_f - plot power spectral density using GNU Radio

SYNOPSIS

       gr_plot_psd_f: [options] input_filename

DESCRIPTION

       Takes a GNU Radio binary file (with specified data type using --data-type) and displays the I&Q data versus time as well as the power spec-
       tral density (PSD) plot. The y-axis values are plotted assuming volts as the amplitude of the I&Q streams and converted	into  dBm  in  the
       frequency  domain  (the	1/N  power adjustment out of the FFT is performed internally). The script plots a certain block of data at a time,
       specified on the command line as -B or --block. The start position in the file can be set by specifying -s or --start  and  defaults  to  0
       (the start of the file). By default, the system assumes a sample rate of 1, so in time, each sample is plotted versus the sample number. To
       set a true time and frequency axis, set the sample rate (-R or --sample-rate) to the sample rate used when capturing the samples.  Finally,
       the  size  of  the FFT to use for the PSD and spectrogram plots can be set independently with --psd-size and --spec-size, respectively. The
       spectrogram plot does not display by default and is turned on with -S or --enable-spec.

OPTIONS

       -h, --help
	      show this help message and exit

       -B BLOCK, --block=BLOCK
	      Specify the block size [default=8192]

       -s START, --start=START
	      Specify where to start in the file [default=0]

       -R SAMPLE_RATE, --sample-rate=SAMPLE_RATE
	      Set the sampler rate of the data [default=1.0]

       --psd-size=PSD_SIZE
	      Set the size of the PSD FFT [default=1024]

       --spec-size=SPEC_SIZE
	      Set the size of the spectrogram FFT [default=256]

       -S, --enable-spec
	      Turn on plotting the spectrogram [default=False]

SEE ALSO

       gr_plot_char(1)	gr_plot_const(1)  gr_plot_fft_c(1)  gr_plot_fft_f(1)  gr_plot_float(1)	 gr_plot_int(1)   gr_plot_iq(1)   gr_plot_psd_c(1)
       gr_plot_psd_f(1)  gr_plot_qt(1)	gr_plot_short(1)

gr_plot_psd_f 3.5						   December 2011						  GR_PLOT_PSD_F(1)