I tried finding frequency by various tools such as audacity (got frequency 0-16000 Hz). Tried with Librosa(Python package) and got the frequency of human recorded audio files in the range of 0-8000 Hz.
Please suggest any way to find out the frequency of audio file
Dear Akshay,
Hope you are well,
The audio frequency is ranging between 20 to 20 kHz. This is different from the speech. The bandwidth of the speech is normally much less than the audio frequency. One hears more bandwidth than he speaks.
The standard speech bandwidth is from 300 Hz to 3400 Hz.
If you have a digitally recorded sound , then it is band-limited audio signals.
and also samples. As the colleagues said the sampling frequency is at least twice the maximum frequency contained in the analog audio signal.
If you want to study the audio signals produced naturally you must have an analog wide band audio electronics and microphones. Then you can analyse it by audio spectrum analyzer .
Best wishes
There are actually two frequencies in audio files you need to differentiate. The first one is the sampling frequencies. It is probably what audacity is putting out. Because of the Nyquist theorem the highest recorded frequency is only half of that. A low pass filter is used to filter out higher frequencies. If you wouldn't do this aliasing would occur, i.e. you would hear sound artifacts which were not present when recording.
I am not sure what exactly you tried to determine frequencies. However, I guess that both is right: The sampling frequency might be 16000Hz and the highest frequency in the recording is thus 8000Hz.
It is a completely different question if 8000Hz is sufficient to capture human voice. This entirely depends on your application. I expect it would not sound entirely natural to human ears. The base frequency of human spoken voice might be below 260Hz. However, this is only the frequency range you use to distinguish humans. It does not carry any speech information at all. Some important information about vowels might be in the range between 3000 and 4000Hz. I am not an expert in this, though. So, I cannot help with the question what frequency range you would need for intellegibility of speech.
If you want further help with how to figure out anything about the frequencies in your sound files you should mention in which format they are saved in. This could help in suggesting proper tools.
Dear Akshay,
Hope you are well,
The audio frequency is ranging between 20 to 20 kHz. This is different from the speech. The bandwidth of the speech is normally much less than the audio frequency. One hears more bandwidth than he speaks.
The standard speech bandwidth is from 300 Hz to 3400 Hz.
If you have a digitally recorded sound , then it is band-limited audio signals.
and also samples. As the colleagues said the sampling frequency is at least twice the maximum frequency contained in the analog audio signal.
If you want to study the audio signals produced naturally you must have an analog wide band audio electronics and microphones. Then you can analyse it by audio spectrum analyzer .
Best wishes
Mohammed Deeb so I was getting the correct result? i.e frequency between(0-8kHz) but as per wiki page, it has to be below 255 for the human voice. Actually I want to classify different voices on basic of frequency, will that be possible?
The early of the communications by phone stabilished a band of 3kHz to each voice channel. That's why you hear the voices like the speakers are inside a can, this effect is mostly in older equipment. You loose some of the pitch and female voice may sound a childish voice.
I do not know your intention for voice recording, but your limit will be the technology of your instruments. You should use the maximum frequency you can handle. Broader frequency ranges will require more memory and more processing capability. Test some acquisition frequencies and see the results in term of quality and time to process as you wish.
- Badges
- Science topic
- Similar topics
- Filing