Finding exact frequences

Re: Finding exact frequences

Thank you very much.

[Craig Maier]

Re: Finding exact frequences

In the future, there will be a marker that you can maneuver with the mouse to determine the frequency and amplitude of any frequency domain signal that you choose to measure on the Spectrum Analyzer. We hope that this will be useful.

Re: Finding exact frequences

Thank you, I put your advice to the test, and although (in 4.18???) things are a little different, I worked out the answer.

First, the spectral chart in the continous noise filter allows you to set the FFT size, but not the resolution. The spectral analyzer (from the View drop-down) doesn't mention FFT size, but has a setting for number of bands, and DOES allow the setting of the resolution, although max bands + min resolution seems to turn the whole thing off.

BUT what I hadn't noticed is the "show peak" check-box in the analyzer, which displays the actual frequency of the highest peak, which seems to be the number of interest to me. I would still like it better if I could choose ANY peak and have it's frequency read off to me, but for the job at hand this will do.

Again, thanks for the pointer.

Cyril

[Craig Maier]

Re: Finding exact frequences

Are you sure that the spectrograph is useless for this? I have used it successfully. You need to make the FFT size maximum. You also need to make the resolution the smallest value (I think that it is 0.16 Hz. Then the full scale frequency becomes 500 Hz with 0.16 Hz resolution. Give it a try. If this does not work, another technique for identifying a noise frequency is to use the filter in inverse. Instead of adjust the filter in its normal mode, use the keep residue mode. Now you will be hearing only the noise terms. Adjust for the loudest buzz or hum. When these settings are found, turn off keep residue and run the filter.

I am working with material which has been transfered from records -> r2r tape -> r2r tape -> cassette -> wave files.

Some of the items contain "buzz", but the multiple transfers have caused the frequency to drift, usually in the range of 58 to 60 cycles/second. I have resorted to setting markers in a section of the file, counting the waves between them and computing the actual frequency. The spectographs are useless for this.

Is there any way to convince the spectographs to give the exact frequency of the peaks?

Cyril N. Alberga