hello; what i want to do is to get spectrum of an sinusoidal signal using the altera fft core i have generated a v8 fft with the following setting: fft point : 64 (to start with) architecture : brust data width : 16 twidle width : 16 i am including the matlab files in the attachment .. the input is a 64 point 2`compliment form. in the compare.m file i am doing three fft,,2 matlab fft and 1 altera core fft what i cant figure out is why the magnitude of the second and the the third power spectrum has bigger amplitude than the first fft at the signal fundmental frequency ? and why are those two are not so clean like the first one ? please have a look .. thanks to anyone who replies.

Hi Mesbah, if you plot(y2) or (y) you will see the distortion you are causing... good luck

i dont think its distortion i just raised the negitive part of the signal up... do u think i shouldnt do that ?..... should i feed y1(has pos & neg ) to the altera fft ? if so how can i express negative part of the signal in vhdl ? thanks

Hi Mesbah, you are causing fatal distortion to your nice sine waves by offsetting the negative halves up as this implies high frequencies due to sudden changes. Remember time domain and freq domain are related: sudden changes of amplitude or phase in time domain imply high frequencies, and high freq imply sudden changes in time domain... your ifft expects signed input, so I am told. The vhdl people allow you + representation. std_logic is read according to the way you decide(or ifft decides). If you want to store your sine data in lut then use mif and it accepts signed values. if you want to store 64 data on wires(as constants) then and to avoid the hassle of conversions use type "signed" then convert as required. Here is my adjustment of your code: clear; N = 64; Fs = 64; INVERSE = 0; y = round((2^15-1)*sin(2*pi*5*[0:63]/Fs)); Matlab_fft = fft(y); %[Y, exp_out] = fft_1_model(y,N,INVERSE); %Altera_fft = Y.*2.^(-exp_out); power_fft1 = abs(Matlab_fft); %power_fft2 = abs(Altera_fft); plot(power_fft1);hold %plot(power_fft2,'r') ************* if for any other reason you want to get rid of negative sine data then offset all values up, you will then get your frequency and some dc: y2 = y + abs(min(y)); figure;hold plot(y); plot(y2,'r')

I have a different but probably somewhat related question. I have a continuous digitized 1000Hz sine wave going to the FFT input. The FFT is configured with a streaming architecture. I get the expected two (almost identical) peaks from the output-one at the positive side and the other at the negative side. My question is how come the values for the peaks fluctuate constantly. In other words, the amplitudes for these two peaks keep changing, and the difference between the maximum and minimum amplitude is quite large.

Hi, You got a bug somewhere. One possible subtlety is this: Is your sine input continuous in phase. If you are reading sine LUT make sure there is phase continuity at the end of the cycle. You need to carefully choose your data points and not repeat the first value at the end. for a correct cycle I will do this: sin_data = sin(2*pi*[0:1023]/1024); plot(sin_data) then scale it for hardware. choose other than [0:1023]/1024 for your frequency case. You should see the first sample is 0 but the last sample is just under zero and shouldn't be zero. Having said that it is unlikley that phase will cause major amp fluctuations unless it is grossly wrong. So look for timing issues, your sampling clk and the way your input actually enters the fft.

Help with altera FFT | Altera Community

56 Replies

Altera_Forum
Honored Contributor
16 years ago
between your data source and your FFT input you need a multiplier, ROM, and state machine. the input data and ROM output are connected to the multiplier, and the output of the multiplier (truncated) is fed into the FFT. a counter steps through the ROM at the same rate as the input data is fed into the FFT. i will see if i can draw a block diagram, but it is fairly straight forward.
Altera_Forum
Honored Contributor
16 years ago
So you are just multiplying the input data with some constants before passing it to the FFT. What are the values in the ROM? And how large is the ROM? I guess the ROM size should be depenent on the FFT size?
Altera_Forum
Honored Contributor
16 years ago
you are correct, each input data is multiplied by a constant. this value depends on which input sample you are working with. the values in the ROM are the FFT window which you select. you can read about them here:

http://en.wikipedia.org/wiki/window_function

you can generate FFT windows with MATLAB or Octave (I think) which is free. i can generate a window and paste it here if you'd like. the length of the ROM is indeed equal to the length of the FFT.

you should do some reading on FFT windows to see which trade-offs you are willing to deal with. basically it comes down to dynamic range vs spectral resolution.
Altera_Forum
Honored Contributor
16 years ago
I'll do some reading on it. Given the following problem that I have...

I have an audio wave file that changes around 1000Hz going into the FFT (configured as streaming 256 points.) At the output of the FFT, I check for the amplitude in the bin for 1000Hz. My problem is that when I play the SAME audio file over and over again, most of the time I will get a large amplitude in that bin for 1000Hz (which is correct), but once in a while, I get a very low amplitude in that bin.

...should I use dynamic range or spectral resolution?
Can you please generate the window for me and tell me how to do it in Octave? Yes, Octave is GNU free software and is 100% compatible with Matlab.
Altera_Forum
Honored Contributor
16 years ago
i was looking at the images you posted and thought that it looked like a windowing issue. your last post sounds a little different.

are you displaying the magnitude spectrum in linear or log? if its linear then small variations in amplitude of the sinusoid will look like significant variations in the frequency domain.
Altera_Forum
Honored Contributor
16 years ago
I'm displaying it in linear. But I really don't care too much about the display; I was just using that to visualize the numbers that I was getting. What I want to do is to check whether there is a 1000Hz in the input sample. So I just constantly read the value that is in the FFT output bin for 1000Hz. If the value is greater than a certain threshold, I say that there is a 1000Hz in the sample. I play this same audio file over and over again, and most of the time, the value is greater than the threshold, but as I said, once in a while, it is much lower than the threshold.

The only difference that I can think of is that I play this same audio file at a different time in the FFT input cycle, and that is why I thought that your suggestion of a windowing problem might be the solution.
Altera_Forum
Honored Contributor
16 years ago
does your wave file sweep frequencies or is a constant tone?

maybe an FFT is overkill if you're just trying to detect presence of a 1 kHz tone.

is it possible that some of your FFT input data is silence until the wave file actually plays? that would explain why the amplitude varies below your threshold.
Altera_Forum
Honored Contributor
16 years ago
--- Quote Start ---
is it possible that some of your FFT input data is silence until the wave file actually plays? that would explain why the amplitude varies below your threshold.
--- Quote End ---

Well yes. I can't start the FFT and the wave file at the same time. I have to start the FFT first, and so the FFT is reading silence input before I play the wave file. If I can't do this then what am I supposed to do?
Altera_Forum
Honored Contributor
16 years ago
i would say that could affect the amplitude of your tones. one solution is a kind of squelch circuit that turns the FFT on/off based on an amplitude threshold in the time domain.
Altera_Forum
Honored Contributor
16 years ago
So the FFT should be turned on only at the first sound signal?
What if the FFT input is not silence but some other random sound?

Forum Discussion

Help with altera FFT

56 Replies

Recent Discussions

Agilex3/5 GTS Hard Ethernet IP 10G example design pin loc and io std wanted

Agilex 7 slew rate reconfiguration

Agilex-7 AXI MCDMA for PCIe hang

Constraints not being picked for DCFIFO

Can't generate F-Tile Ethernet Hard IP Design Example