in the above image, you’ll see that the pitch of person A is lower than the pitch of person B, now what I don’t know how to filter the pitch of person B out of the audio.
I Analysed ‘Plot spectrum’ and the results were as follows for A and B:
I thought that high pass filter will do the trick, but I don’t know what frequency I need to input in it which it will filter up. Please look at those spectrum plots and tell me which frequency i need to enter and where? Will high filter do the trick? or is there any other method/way to remove all voices below this pitch of person A from this audio.
Through out the audio everywhere, Pitch of A is lower than that of person B;
I noticed pitch is the only factor that distinguishes the voice of two person in the audio so is there any tool which could give me the parameter of different pitch of those 2 person and then help me filter the voice of the person having lower pitch in it?
Thanks in advance
If you run the size control in Plot Spectrum up way higher, you’ll find there is a terrific amount of overlap between human voices, even those that to your ear sound different.
For example, if both people said “Shshshshsh, be quiet,” then be quiet may be different, but shshshshs from the two voices would be exactly the same.
That brings us to: Audacity can’t separate mixed voices, instruments or sounds from each other. One of the ways to kill a podcast recording is to leave a TV running in the background by accident. It’s trash. No way out.
In addition, my spidey sense is telling me you’re trying to do surveillance, conflict resolution, or sub-rosa production and we can’t help you.
In addition, my spidey sense is telling me you’re trying to do surveillance, conflict resolution, or sub-rosa production and we can’t help you.
Koz
Actually it’s just a recording of a voice call between me and a friend of mine.
Now from a tool(i don’t know whether it’s against the policy of board or not to mention names here) I found that her pitch always remain above always 200 and mine on the other hand is around 130, So is there a way to filter the voice out having pitch less than say 180?
Please do help.
If you select a region of audio where the unwanted person is speaking, you can mute it by pressing “Ctrl”+“L” keys. Alternatively you can bleep them out using generate-tone, (1000Hz sine-wave is the traditional censor bleep).
Again select the unwanted region, generate tone will replace the selected audio with a bleep of the same duration.
You’ve zoomed-in vertically: those lines on the spectrogram (“A”&“B”) are only one small part of speech.
Just removing those one of those lines will not silence the voice.
If you must try that, small parts of the spectrogram can be erased using Audacity’s, spectral-edit tools …
It’s a slow process & I don’t think it will achieve your objective.
Ah. Actually, those screen shots are of Pitch EAC graph, from the same menu of that of waveform and spectrogram, anyway here is the ss of the spectrogram :https://imgur.com/L2mBxBM and when i saw it I wasn’t able to distinguish between the two voices by looking at it where as if you see this screen shot of pitch EAC graph https://imgur.com/u3adF0N you can clearly distinguish between the two voices on the basis of their pitch level.
and @koz, High pitch and low pitch filter aren’t doing the trick
The EAC is still a spectrogram, but one which only gives you the gist of the frequency content: the fundamental frequencies …
Removing the fundamentals still leaves all their harmonics.
I didn’t know I was allowed to give away the name.
You can do anything a forum elf asks you to do. We all all powerful [flexing muscles].
The actual restriction is never, ever try to sell us anything without notice. We have no trouble sending people off to an app that does something useful, many times whether Audacity can do that or not.
