March 4, 2015 at 15:56We check in practice the meaninglessness of high sampling rates
Much has been written about the unnecessary sampling frequency of 192,000 Hz in audio files intended for listening. But arguments usually refer to theorems, for a correct understanding of which one needs to be pretty good at mathematics. But there is another way to test some statements - to carry out the corresponding experiments many times.
First you need to formulate a statement that will be verified in the future. If the frequency of 192000 Hz is excessive, then there must be a lower sampling frequency at which the correct restoration of the audio signal occurs. Suppose this is a frequency of 44100 Hz. If the correct recovery occurs at 192000 Hz, and the correct recovery occurs at 44100 Hz, then in both cases the output should be the same. To test this in practice, you need expensive equipment, which not everyone has. So let's go a little different way. If the files 44100 Hz and 192000 Hz contain the same amount of information about the sound, then this means that the conversion from 192000 Hz to 44100 Hz is lossless compression, therefore, There must be a way to recover from the 44100 Hz file the original 192000 Hz file. Now everyone can check it on any modern computer.
As the source, I chose a piece of music with a sampling frequency of 192000 Hz. If I came across some wrong material, in which initially there was nothing important that should be lost when converting from 192000 Hz to 44100 Hz, then anyone can do what is described in this article with any other file. All actions will be performed in the free Audacity editor with standard effects. All files received in the process will be saved in the FLAC format with a bit capacity of 24 bits.
The source file is stored in the “A.FLAC” file and looks like this:
But its spectrum looks like this:
We are only interested in sound information, so we will remove the ultrasound using the equalizer.
And we get the following spectrum:
We export the result to the “B.FLAC” file - it is with it that we will compare the file that will be obtained at the end of all conversions.
Before converting the sampling frequency, make sure that the maximum quality is set in the settings:
Next, select the new project sampling frequency and export it to the file “C.FLAC”
Then open the file “C.FLAC”, set the project sampling frequency to 192000 Hz and export it to the file “D .FLAC ".
And the main stage remained: open the files “B.FLAC”, “D.FLAC” and compare them:
Let's take a closer look:
There is no difference. Compare better - invert one of the tracks
And put everything in one track:
Silence! The complete silence!
But in order to see the difference, you need to increase the amplitude by 96 dB!
The difference is so quiet that it cannot be heard, which means that when converting from 192000 Hz to 44100 Hz in the audio range, information is not lost. So, without deep knowledge in mathematics, using the software available to everyone, you can check the sufficiency of the sampling frequency of 44100 Hz for storing music files.
First you need to formulate a statement that will be verified in the future. If the frequency of 192000 Hz is excessive, then there must be a lower sampling frequency at which the correct restoration of the audio signal occurs. Suppose this is a frequency of 44100 Hz. If the correct recovery occurs at 192000 Hz, and the correct recovery occurs at 44100 Hz, then in both cases the output should be the same. To test this in practice, you need expensive equipment, which not everyone has. So let's go a little different way. If the files 44100 Hz and 192000 Hz contain the same amount of information about the sound, then this means that the conversion from 192000 Hz to 44100 Hz is lossless compression, therefore, There must be a way to recover from the 44100 Hz file the original 192000 Hz file. Now everyone can check it on any modern computer.
As the source, I chose a piece of music with a sampling frequency of 192000 Hz. If I came across some wrong material, in which initially there was nothing important that should be lost when converting from 192000 Hz to 44100 Hz, then anyone can do what is described in this article with any other file. All actions will be performed in the free Audacity editor with standard effects. All files received in the process will be saved in the FLAC format with a bit capacity of 24 bits.
The source file is stored in the “A.FLAC” file and looks like this:
But its spectrum looks like this:
We are only interested in sound information, so we will remove the ultrasound using the equalizer.
And we get the following spectrum:
We export the result to the “B.FLAC” file - it is with it that we will compare the file that will be obtained at the end of all conversions.
Before converting the sampling frequency, make sure that the maximum quality is set in the settings:
Next, select the new project sampling frequency and export it to the file “C.FLAC”
Then open the file “C.FLAC”, set the project sampling frequency to 192000 Hz and export it to the file “D .FLAC ".
And the main stage remained: open the files “B.FLAC”, “D.FLAC” and compare them:
Let's take a closer look:
There is no difference. Compare better - invert one of the tracks
And put everything in one track:
Silence! The complete silence!
But in order to see the difference, you need to increase the amplitude by 96 dB!
The difference is so quiet that it cannot be heard, which means that when converting from 192000 Hz to 44100 Hz in the audio range, information is not lost. So, without deep knowledge in mathematics, using the software available to everyone, you can check the sufficiency of the sampling frequency of 44100 Hz for storing music files.