TheBackShed.com

Forum Index : Microcontroller and PC projects : CMM2: Option Base bug in MATH FFT ?

Author

Message

thwill

Guru

Joined: 16/09/2019
Location: United Kingdom
Posts: 3833

Posted: 08:24am 19 Sep 2021

Copy link to clipboard

Print this post

Hi Peter,

Found this whilst working with Volhout's MATH test code:

> 480MHz Colour Maximite 2 G2
MMBasic Version 5.07.01RC2
Copyright 2011-2021 Geoff Graham
Copyright 2016-2021 Peter Mather

> list "fft_bug.bas"
Option Default None
Option Explicit

Const BASE% = Val(Mm.CmdLine$)
Option Base BASE%

Const NUM_SAMPLES% = 512
Dim sin_amp2!(NUM_SAMPLES% + BASE% - 1)
Dim dft!(2 + BASE% - 1, NUM_SAMPLES% + BASE% - 1)
Dim i%
For i% = BASE% to NUM_SAMPLES% + BASE% - 1
sin_amp2!(i%) = 2 * Sin(2 * Pi * i% / NUM_SAMPLES%)
Next

Math Fft sin_amp2!(), dft!()

Print Math(Max dft!())

> run "fft_bug.bas", 1
511.9614473

> run "fft_bug.bas", 0
512

Assuming it's a thing (i.e. I haven't just confused the hell out of myself with Option Base) and you fix it could I please have an updated copy of the MATH source to integrate with MMB4L.

Thanks,

Tom
Edited 2021-09-19 18:33 by thwill

Game*Mite, CMM2 Welcome Tape, Creaky old text adventures

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 8569

Posted: 09:02am 19 Sep 2021

Copy link to clipboard

Print this post

I'll leave you to find the bug in your test program (HINT try comparing sin_amp2!() in both cases)

thwill

Guru

Joined: 16/09/2019
Location: United Kingdom
Posts: 3833

Posted: 09:04am 19 Sep 2021

Copy link to clipboard

Print this post

matherp said I'll leave you to find the bug in your test program (HINT try comparing sin_amp2!() in both cases)

D'oh! Thanks Peter.

Best wishes,

Tom

Game*Mite, CMM2 Welcome Tape, Creaky old text adventures

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 8569

Posted: 09:12am 19 Sep 2021

Copy link to clipboard

Print this post

If you look at the imaginary component of the entry with 511.9614473 it is 6.2830

take the square root of the sum of squares of both numbers and you get 512. Your error introduced a phase shift hence the difference and the non-zero imaginary component

thwill

Guru

Joined: 16/09/2019
Location: United Kingdom
Posts: 3833

Posted: 08:37am 20 Sep 2021

Copy link to clipboard

Print this post

matherp said If you look at the imaginary component of the entry with 511.9614473 it is 6.2830

take the square root of the sum of squares of both numbers and you get 512. Your error introduced a phase shift hence the difference and the non-zero imaginary component

Thanks Peter.

I vaguely remember someone saying "Fourier Transform" in my first year of university, but for the remainder and the 20+ years since they have never reared their head in my career.

Best wishes,

Tom

Game*Mite, CMM2 Welcome Tape, Creaky old text adventures

Print this page

To reply to this topic, you need to log in.