What audio formats do you need?
You are a music lover or a professional musician, or you simply use music for some of your own purposes (performances, dances, videos …), but you are not well versed in all these “computer” things: formats, files, bitrates, extensions … This article is for you …
The content is brief: of all the variety of audio formats, you need four:
1) for high-quality listening, a music library;
2) for “fuzzy” listening service purposes;
3) “for Wikipedia”;
4) for real-time audio processing.
Which ones, why and what to do about it – read below!
Digital sound
Audio Pixels Audio
compressed and uncompressed
To understand how one format differs from another, let’s briefly understand what
it is digital sound.
Sound “pixels”
You have surely seen an image of a sound wave (for example, in a sound editing program). Like any image, a sound graphic can be recorded digitally by dividing it into many small “pixel” parts. The smaller the pixels, the more horizontally and vertically, the better the image.
Only in the case of sound, horizontally, we plot the time axis and vertically, the value of the signal (in real sound propagating through the air, the value of the signal is the change in air pressure). Each second of sound is divided into many smaller identical time intervals, in each of which a signal value is recorded. The number of these segments per second is called the sampling frequency …. And the number of possible gradations of the signal value is determined by the bit depth; This parameter indicates how many bits are used to record a signal value: the more bits that are used, the more gradations of the value can be recorded. For example, our figure shows a bit width of 4 bits, which gives 2 4 = 16 value gradations.
Digital recording of a sound wave. The green curve is the original sound, the purple bars are its digital recording as separate signal values.
As a result of digital rendering, the sound wave looks like this:
Digital-sound-image__presentation_2.png
Digital representation of a sound wave.
This is of course a very rough “drawing”, but if you use a higher sample rate and bit depth you can make the “pitch” invisible to our ears (just like very small pixels are invisible for the eye).
The sample rate of an audio CD is 44,100 hertz (signal values per second) = 44.1 kHz (kilohertz). This is sufficient to encode all frequencies audible to the human ear. And its capacity is 16 bits, which allows to transmit 2 16 = 65,536 gradations of signal values.
Experts discuss whether you need a higher resolution for high-quality sound. Indirect evidence suggests yes. Hence, there is digital sound with higher sample rates (48, 96, 192, 384 kHz) and bit depths (24, 32 bits). This high resolution supports DVD Audio, many types of files (flac, wav, aiff …), some cinema sound formats (Dolby TrueHD, DTS-HD Master Audio …).
Uncompressed and compressed audio bit rate
However, even an audio CD takes up a lot of space. This is shown by the bit rate parameter (bit rate, data rate, kbit / s): it indicates how many bits (kilobits) are required to record 1 second of sound. For uncompressed “normal” sound, it is very easy to find out the bit rate: you need to multiply the bit depth (number of bits times 1 signal value) by the sample rate (number of values per second) and by the number of sound channels (let’s not forget that stereo sound has two sound channels, left and right, in each of which the sound wave is recorded separately).
Using our figure, the bitrate can be represented symbolically as follows:
Digital-sound-image__presentation_3.png
Uncompressed audio bit rate (blue area)
It is easy to calculate that the bit rate of an audio CD is
16 bits × 44 100 Hz × 2 channels = 1,411,200 bps = 1,411.2 kbps.
This is usually true for uncompressed audio file formats (wav, aiff).
So one second of uncompressed sound takes approx. 172 KB, one minute – 10 MB, one hour of music occupies 606 MB, the total sound volume of an audio CD is 74 minutes, ie approx. 750 MB.
[How do we calculate this? ]
Hence, mankind started looking for a way to reduce file size without reducing sample rate and bit depth.
