audio library to check cutoff frequency

[Loe]

Hi,
I just read in google about fake lossless audiofile.
Say it is flac or high bitrate mp3, ex 512kbps mp3.
It fake because its was originaly loss or low bitrate up converted file.

The article also talk about relation between cutoff frequency & bitrate.
We can check it real lossless or high bitrate at it sprctrum or cutoff frequency.

The real flac/320kbps mp3 cutoff freq about 20khz or more. 128kbps mp3 cutoff about 16khz

So here my question, is there any audio lib for fb to check cutoff frequency of audio file?

If it combine with audio encoder lib, we can eliminate fake lossless/high bitrate audio file to downgrade it bitrate so we can save harddisk space.

Any suggestion or direction waiting

[St_W]

I guess you can use nearly any audio library that gives you access to the raw audio sample data and/or implements a FFT. BASS is a very often used audio library, unfortunately it's only free if you do not plan to make money with your software.
With BASS you would probably use the BASS_ChannelGetData function to get the FFT data. I cannot give details here without trying it myself, but I should be rather easy (provided that you understand the signal processing theory behind it, but I assume that's the case).

[sancho2]

Wouldn't arbitrarily reducing the bit rate further degrade the file?
When the fake is created, it is 'recorded' from a lossy mp3 (usually 128 or 96) and at that point the damage is done (false or missing data is part of the data in the file). Reducing the bit rate is still working with incorrect data.
It has been a long time since I was involved with this. We used to look at a spectrum analysis and could see the fake at a glance of the graph.
We would discard the fakes.

[Loe]

With BASS you would probably use the BASS_ChannelGetData function to get the FFT data. I cannot give details here without trying it myself, but I should be rather easy (provided that you understand the signal processing theory behind it, but I assume that's the case).

Yes st_w signal processing theory is my weakness, but I prefer to answer i dont want to reinvent the wheel:)
There are many tools to view spectrum, so im sure there sould be audio lib to check/guess cutoff frequency.

Wouldn't arbitrarily reducing the bit rate further degrade the file?
When the fake is created, it is 'recorded' from a lossy mp3 (usually 128 or 96) and at that point the damage is done (false or missing data is part of the data in the file). Reducing the bit rate is still working with incorrect data.
It has been a long time since I was involved with this. We used to look at a spectrum analysis and could see the fake at a glance of the graph.
We would discard the fakes.

Yes sancho2, but if we had many file, it should be easy to separate real lossless/high bitrate with batch operation not with visualy look at it spectrum.
Also it better to reduce it quality a bit after 'optimize' rather than to keep 20mb 512kbps fake file, whic is 4mb after optimize:)