how to use audio compression Archives - Page 5 of 6

Compression and compression methods for audio signals Part 2

Free Download Mp4Gain

Compression and compression methods for audio signals Part 2

audio compression

FLAC is a member of the Xiph.Org codec family. By the way, it also includes the well-known ogg vorbis, one of the best lossy music compression algorithms. As a container for audio data, of course, OGG (files with the extension .ogg) and another open source container – Matroska (files with the extension .mka) are used.

It should be noted right away that both the FLAC format and algorithm are fully open. They are not patented, so they can be used completely free of charge in any program. This is the reason for the wide support for FLAC in players – any serious gamer has a plugin for FLAC. In addition, there are hardware mp3 players that support the FLAC codec.

The FLAC encoder is compiled for most platforms in use, so there should be no compatibility issues on alternative Windows operating systems.

FLAC supports tags in its own “FlacTags” format. There is the ability to encode multi-channel audio, a great advantage over Monkey’s Audio. The format supports any sample rate in the range of 1 Hz (!) To 65,535 Hz. Audio bit depth from 4 (!) To 32 bits.

FLAC is believed to be the most efficient use of system resources when decoding (playing) audio compared to other lossless codecs. Unfortunately, this is achieved at the expense of a significant increase in encoding (compression) time.

The FLAC website is regularly updated and new versions of the codec are released. Overall, FLAC is without a doubt the leader in terms of development activity. This may make it the main format in the future. Well, let’s see …

FLAC is the best option for storing high quality music.

MIDI (Musical Instrument Digital Interface) is a standard for hardware and software that allows you to play (and record) music by executing / recording special commands, as well as the format of the files that contain those commands. The playback device or program is called a MIDI synthesizer (sequencer) and is actually an automatic musical instrument.

Unlike other formats, it does not store the digitized sound, but sets of commands (played notes, links to played instruments, variable sound parameter values) that can be played differently depending on the playback device. The convenience of the MIDI format as a data representation format enables devices that produce automatic arrangements according to given chords, as well as 3D sound visualization applications. Additionally, these files tend to be orders of magnitude smaller than digitized audio of comparable quality.

Monkey’s Audio is a popular lossless digital audio encoding format. Distributed for free along with open source and a suite of encoding and playback software, as well as plugins for popular players. Monkey’s audio files use the following extensions: .ape to store audio and .apl to store metadata. Despite being open source, Monkey’s Audio is not free, as its license imposes significant restrictions on its use.

Audio files compressed with the Monkey audio codec have the extension ‘APE’; As you can see, the monkeys are present not only in the logo or the name (from English monkey: monkey, primate).

The average bit rate in an audio file is 600 to 700 kbps; compare with 128 kbps in MP3. Average compression is 40-50%, depending on the genre of music: if classical or jazz pieces are compressed in the best way, then compositions in the style of trash-metal or something similar “electronic noise” will show the worst result. . For codecs with acceptable quality loss, compression is approximately 80%.

There are four levels of compression. Maximum compression may seem like the only correct solution, although the compression time is quite long. However, you must also take into account the resource consumption of the system that plays the file; for the most compressed file, it is relatively high.

The .APE format provides tag support for searching for songs in your music collection. Another advantage is the verification of the integrity of the file during decoding. Recovery of original compressed .APE wav files is supported.

Monkey’s Audio has a graphical interface for Windows, in other words, a convenient window program to manage the encoding process. The rest of the codecs require the use of the command line or third-party interfaces.

Free Download Mp4Gain

Mp4Gain Main Window

Mp4Gain Features

Free Download Mp4Gain

Compression and compression methods of audio signals

Compression and compression methods of audio signals (types, differences, use)

Audio Compression

Basics of the analog-to-digital conversion principle, sound conversion and compression method, existing sound storage formats. Programs to convert and process sound and audio files. Application of these programs in linguistic research.

Bit rate is the amount of information per unit of time. In general, the bit rate is the number of bits that we spend encoding a sound with a duration of 1 second.

Analog-to-digital converter (ADC): A device that converts an input analog signal into a binary code (digital signal). The reverse conversion is done using a DAC (digital-to-analog converter, DAC). Typically, an ADC is an electronic device that converts voltage into a binary digital code. However, some non-electronic devices with digital output must also be classified as ADCs, such as some types of angle-to-code converters. The simplest one-bit binary ADC is a comparator.

The circuit to convert an audio signal from analog to digital:

Sampling is the transformation of continuous images and sound into a set of discrete values in the form of codes.

Quantization is the process of aligning a set of musical notes to a grid.

Compression (compression) of audio data is a process of lowering the bit rate by reducing the statistical and psychoacoustic redundancy of a digital audio signal.

The underlying idea behind all lossy audio compression techniques is to neglect the subtle details of the original sound that are beyond the reach of the human ear.

Codec (CoDec) is an abbreviation for compressor and decompressor. Basically, a codec is a collection of files, drivers, and libraries required to package a video or audio file into a compressed format and play the compressed file.

Formats:

AAC (Advanced Audio Coding) is an audio file format with less quality loss when encoding than MP3 of the same size. The format also allows you to compress without losing the quality of the source (ALAC AAC profile).

AAC (Advanced Audio Coding) was originally created as a successor to MP3 with improved encoding quality. The AAC format, officially known as ISO / IEC 13818-7, was released in 1997 as the new seventh part of the MPEG-2 family. There is also the AAC format known as MPEG-4

Apple AIFF: This file type is standard for Apple Macintosh systems and sound processing systems built on top of it. Apple AIFF stands for Audio Interchange File Format, an audio interchange file format, it is somewhat similar to WAV. Its peculiarity is that it allows you to place additional information along with the sound wave, in particular WaveTable samples (examples of the instrument sound together with synthesizer parameters), which improves the quality of the final result. Although today Apple computers are capable of playing files of almost any format, including MP3.

FLAC (Free Lossless Audio Codec) is a popular free codec for audio compression. Unlike lossy Ogg Vorbis, MP3 and AAC codecs, it does not remove any information from the audio stream and is suitable for both daily listening and archiving of audio collection. Today, the FLAC format is compatible with many audio applications.

Digital audio compression methods

Digital audio compression methods

audio compression

Lossless compression

AUDIO COMPRESSION

Generally speaking, the meaning of lossless compression is as follows: some pattern is found in the original data, and taking this pattern into account, a second stream is generated, uniquely describing the original. For example, to encode binary sequences in which there are many zeros and few ones, we can use the following replacement:

00> 0
01> 10
10> 110
11> 111

In this case, sixteen bits:
00 01 00 00 11 10 00 00

will be converted to thirteen bits:
0 10 0 0 111 110 0 0

If we write a compressed string without spaces, we can still add spaces in it, which means restoring the original sequence.

FLAC (Free Lossless Audio Codec)
Coding principle: the algorithm tries to describe the signal with this function so that the result obtained after subtracting it from the original (called difference, remainder, error) can be encoded with the minimum number of bits.

When the model is fitted, the algorithm subtracts the approximation from the original to obtain a residual signal (error), which is then losslessly encoded.

Lossy compression (MP3, AAC, WMA, OGG)
Using a lossy compression algorithm, the size of an MP3 file with an average bit rate of 128 kbps is approximately 1/11 of the original file of an Audio CD (uncompressed audio in CD-Audio format has a rate bit rate of 1411.2 kbps). MP3 files can be created at high or low bit rates, which affects the quality of the result.

The principle of compression is to reduce the precision of some parts of the sound flow, which is almost indistinguishable for most people. The audio signal is divided into segments of equal length, each of which, after processing, is packed into its own frame (frame). Spectral decomposition requires continuity of the input signal; therefore the table above and below are also used for calculations. The audio signal contains harmonics with a lower amplitude and harmonics that are close to the strongest; Such harmonics are cut off, as the average human ear will not always be able to determine the presence or absence of such harmonics. This characteristic of hearing is called the masking effect. It is also possible to replace two or more nearby peaks with an averaged one (which, as a rule, leads to sound distortion). The cutoff criterion is determined by the outflow requirement. Since the entire spectrum is relevant, the high-frequency harmonics are not cut off, but are only selectively removed to reduce information flow due to spectrum sparsity. After spectral removal, mathematical compression and frame packing methods are applied.

Masking effect
In certain cases, a sound can be hidden by another sound. For example, talking near the railroad tracks can be completely impossible if a train passes. This type of effect is called masking. A weak sound is said to be masked if it becomes indistinguishable in the presence of a louder sound.

Simultaneous masking
Any two sounds when heard simultaneously have an impact on the perception of the relative volume between them. A louder sound reduces the perception of a weaker one, until the disappearance of your hearing. The closer the frequency of the masked sound is to the frequency of the masker, the more it will be hidden. The masking effect is not the same when the masked sound is shifted down or up in frequency with respect to masking. Low-frequency sound masks high-frequency sound. However, it is important to note that high-frequency sounds cannot mask low-frequency sounds.

Time masking
This phenomenon is similar to frequency masking, but time masking occurs here. When the masking sound is stopped, the masking remains inaudible for some time. Under normal conditions, the temporary masking effect lasts significantly less. The masking time depends on the frequency and amplitude of the signal and can be up to 100 ms.
In the case where the masking tone appears at a time after masking, the effect is called post-masking. When the masking tone appears before the masking (this is also possible), the effect is called premasking.

Post-stimulus fatigue
Often after exposure to loud, high-intensity sounds, a person’s hearing sensitivity drops dramatically. Recovery to normal thresholds can take up to 16 hours. This process is called “temporary change in hearing sensitivity threshold” or “post-stimulus fatigue.”

Digital audio compression methods

Digital audio compression methods

Audio Compression

Lossless compression

Generally speaking, the meaning of lossless compression is as follows: some pattern is found in the original data, and taking this pattern into account, a second stream is generated, uniquely describing the original. For example, to encode binary sequences with many zeros and few ones, we can use the following replacement:

00> 0
01> 10
10> 110
11> 111

In this case, sixteen bits:

00 01 00 00 11 10 00 00

will be converted to thirteen bits:

0 10 0 0 111 110 0 0

If we write a compressed string without spaces, we can still add spaces in it, which means restoring the original sequence.

FLAC (Free Lossless Audio Codec – Free Lossless Audio Codec)
Coding principle: the algorithm tries to describe the signal with this function so that the result obtained after subtracting it from the original (called difference, remainder, error) can be encoded with the minimum of bits.

When the model is fitted, the algorithm subtracts the approximation from the original to obtain a residual signal (error), which is then losslessly encoded.

Lossy compression (MP3, AAC, WMA, OGG)
Using a lossy compression algorithm, the size of an MP3 file with an average bit rate of 128 kbps is approximately 1/11 of the original file of an Audio CD (uncompressed audio in CD-Audio format has a rate 1411.2 kbps bit rate). MP3 files can be created at high or low bit rates, which affects the quality of the result.

The principle of compression is to reduce the precision of some parts of the sound flow, which is almost indistinguishable for most people. The audio signal is divided into segments of equal length, each of which, after processing, is packed into its own frame (frame). Spectral decomposition requires continuity of the input signal; therefore, the previous and next tables are also used for calculations. The audio signal contains harmonics with a lower amplitude and harmonics that are close to the strongest; Such harmonics are cut off, as the average human ear will not always be able to determine the presence or absence of such harmonics. This characteristic of hearing is called the masking effect. It is also possible to replace two or more close peaks with an averaged one (which, as a rule, leads to sound distortion). The cutoff criterion is determined by the outflow requirement. Since the entire spectrum is relevant, the high frequency harmonics are not cut off, but are only selectively removed to reduce information flow due to rarefaction of the spectrum. After spectral removal, mathematical compression and frame packing methods are applied.

Masking effect
In certain cases, a sound can be hidden by another sound. For example, talking next to a train track can be completely impossible if a train passes. This type of effect is called masking. A weak sound is said to be masked if it becomes indistinguishable in the presence of a louder sound.

Simultaneous masking
Any two sounds, when heard simultaneously, have an impact on the perception of the relative volume between them. A louder sound reduces the perception of a weaker one, until the disappearance of your hearing. The closer the frequency of the masked sound is to the frequency of the masker, the more it will be hidden. The masking effect is not the same when the masked sound is shifted down or up in frequency relative to masking. Low-frequency sound masks high-frequency sound. However, it is important to note that high-frequency sounds cannot mask low-frequency sounds.

Time masking
This phenomenon is similar to frequency masking, but time masking occurs here. When the masking sound is stopped, the masking remains inaudible for some time. Under normal conditions, the effect of temporary masking lasts much less. The masking time depends on the frequency and amplitude of the signal and can be up to 100 ms.
In the case where the masking tone appears later than the masking, the effect is called post-masking. When the masking tone appears before the masking (this is also possible), the effect is called premasking.

Post-stimulus fatigue
Often, after exposure to loud, high-intensity sounds, a person’s hearing sensitivity drops dramatically. Recovery of normal thresholds can take up to 16 hours. This process is called “temporary change in hearing threshold.”

Audio compression for music lovers

Audio compression for music lovers

the truth about high bitrate lossy compression

In the opinion of most people, the word music lover is most often associated with a person who not only loves and collects music, but also appreciates high-quality music, and not only in artistic and aesthetic terms, but also the quality of the recording of the phonogram itself. Just think, a few years ago, an audio CD was considered the standard for music quality, whereas a computer, even in dreams, could not compete with the quality of a CD. However, time is a great joker, and he often likes to turn things upside down. It would seem that quite a while, a year or two passed and … that’s it, the CD on the PC went into the background. Don’t ask “why?”, You know the answer to this question yourself. Everything is to blame for the revolution in the world of computer sound: audio compression (hereinafter referred to as audiolo compression which means lossy compression to reduce the size of the audio file), which made it possible to store music on disk hard, lots of music! In addition, it was possible to exchange it over the Internet. New sound cards have been released, capable of almost “squeezing” studio quality out of a piece of hardware that seems useless in terms of music. Today, even having a computer that is not very smart in performance, having bought a Creative SoundBlaster Live! and remembering that since Soviet times there is a good amplifier and good acoustics, you will get nothing but a high-quality music center, the sound of which is inferior only to very expensive audio equipment (average or even the highest Hi-Fi category ). Add to this the general availability of music files and you understand that you have the power in your hands. And then there is a revolution, and you understand that a compact disc is no longer so convenient, you are fascinated by something completely different: the magic “MP3” signs. You cannot eat or sleep; you are faced with the seemingly insoluble “chicken and egg” question: how to “squeeze” and, most importantly, how to “squeeze” …

This is where I will help you. This article is the beginning of my new series of informational materials on music on the computer. For over a year developing OrlSoft MPeg eXtension and maintaining an extensive database of MP3 files, I have accumulated a great deal of research on audio compression. It is these studies that I will try to share with you. Many articles have been written on audio compression by different respected authors, so I will try not to write what I can easily find in other sources of information. I would like to put my position on the subject we are considering simply and clearly. We will not consider audio compression to be as compact a tool as possible put audio information on your hard drive (so that you can record so many hours of music there). Yes, compression allows you to record music more compactly, but my goal is to minimize quality loss by converting “pure” audio to compressed audio. This is why only high bit rates and qualitatively compressing encoders are considered in these modes. So it is much more convenient to work with compressed audio – instant access to any track from any album, convenient software for playback. And, of course, the financial issue has not been forgotten either.

Of the audio compression formats that exist today, in my opinion, three deserve attention: MP3 (or MPEG-1 Audio Layer III), LQT (as representative of the MPEG-2 AAC / MPEG-4 family) and a Completely new OGG format (Ogg Vorbis) developed by a group of enthusiasts:

MP3 is by far the most used of these (mainly because it is free). Let me remind you that it was thanks to the MP3 format that the victorious procession of compressed audio took place. However, as often happens with pioneers, little by little it is losing ground and giving way to new and better formats.
The second format, LQT, is a representative of a new direction of audio coding algorithms, a representative of the AAC family. This is a fairly high quality, but commercial and highly classified format.
OGG became widely known to the public this summer and is currently developing rapidly, soon (with the launch of the Encoder and Decoder) it should beat MP3 with better sound quality with smaller file size.

Digital audio compression

Digital audio compression

Digital Audio Compression

The concept of loudness is close and understandable not only for a musician, but also for people who are not associated with music. The relationship between the volume of the parts of a piece and the volume of the instruments that are playing simultaneously is called the dynamic range. One of the main tools producers and musicians use to influence dynamic range is the compressor.

Digital Audio Compression

Although the compressor works with a known phenomenon, loudness, in most cases its use occurs spontaneously, randomly, without understanding the essence of what is happening. You can know the general principle of the compressor and the purpose of each handle, but this does not eliminate the stupor at the first experience.

Why do you need a compressor?

The main purpose of the compressor is to automatically change the signal level. It works roughly the same as if you kept your hand constantly on the volume fader, turning it up and down. The difference is that a compressor can react very quickly to changes, much faster and more accurately than a human.

Up to this point, the word compressor meant a whole class of dynamic devices. Using the same basic principles as a conventional compressor, various instruments work for different purposes: limiters, expanders, gates, etc. They are united by working with the volume of individual sounds or the mix as a whole.

The classic compressor is controversial by its very name. Everyone knows that he makes the loudest sound. But the name comes from compress, which means “compression”, and if you ask any sound engineer what a compressor does, you’ll hear the answer: “squash the signal.” The compressor reduces the amplitude of the dynamic bursts, makes them quieter. So what is the main purpose of the compressor: to make it quieter or louder? The answer is both at the same time.

Let’s take an example of voice recording. Very often, in the process of singing, syllables or sounds of different volume are heard. If the singer does not control the dynamics of his performance very well, then such differences create problems for the sound engineer and negatively affect the final result of the work. Silent syllables disappear into the mix, text becomes difficult to distinguish, and if you adjust the volume for a quiet area, in other places the voice begins to “stand out.”

This is where the compressor comes in. It allows you to suppress strong bursts, equalize them with silent fragments. Now you can turn up the volume of the track without fear of some syllables sticking out. So the compressor makes the sound lower and higher at the same time. Three images show the stages of working with sound: a source with large peaks (a), a compressed signal (b) and an increase in the volume level of the entire file (c).

It is especially important to apply compression when recording in a digital environment, when we are forced to adhere to a maximum level of 0 dB, because exceeding this threshold leads to clips and distortion. When clips appear, we lower the preamp level, which means we lower the volume of not only bursts, but quiet areas as well, leading to signal degradation due to quantization and aliasing noise.

The compressor, positioned between the preamp and the digital recording system, operates only on the loudest bursts, reducing their volume and ensuring a smooth soundtrack. Thanks to this, we have the opportunity not to reduce the overall volume of the recorded signal and to maintain the sound quality.

Unfortunately, many modern musicians, without going into the technical characteristics of the compressor, use it everywhere, believing that with its help you can “stretch” any sound in the mix. Also, compressors are often included on the road in extreme conditions. They are only used by experienced sound engineers when there is a real need.

The compressor helps avoid recording problems. The most common causes of problems can be the following:

Non-professionalism of the interpreter (dynamic unevenness).
Mismatched path (bad, mismatched, or inadequate microphones, preamps).
Disadvantages of the digital environment (limited to 0 dB).
Uncomfortable conditions for the singer (small and stuffy room, poor monitoring).
Low qualification of a recording engineer.
If a performer has a voice and can sing into a microphone, and a recording engineer knows her job well and knows how to properly position microphones and set up equipment, a compressor may not be required at all. But this is the ideal situation.

Digital audio compression

Digital audio compression

Digital Audio Compression

Audio data compression is a real problem today. There are two reasons for the need to compress audio data: memory savings when storing audio information, low bandwidth of remote digital information transmission channels. Compression effectively solves the two problems above. Data compression is an algorithmic transformation of data performed to reduce its volume.

Data Compression

It is used for a more rational use of data storage and transmission devices. Compression is based on eliminating the redundancy contained in the original data. To guarantee the parameters necessary for the transmission of voice signals (music) over modern low-speed digital communication channels and to guarantee the specified noise immunity, it is necessary to use highly efficient data compression algorithms. The transmission channel is characterized by a concept such as the capacity of the channel: And the signal – by the volume (signal): …

Both of the above features include dynamic range D, channel width (signal spectrum), and transit time T. Digital audio compressors are used to reduce dynamic range. To improve spectral efficiency, digital filters are used to limit the spectrum of the encoder output signal (according to Nyquist criteria). Among other things, encoders based on the principles of elimination of redundancy (Huffman codes) are used to guarantee a certain information transmission speed. The essence of which is as follows: codes based on the principle of assigning more probable values of the amplitudes of the codewords of shorter length than the improbable ones.

Let’s consider how the types of redundancy described above are eliminated.
Structure of a lossy audio compression encoder The original digital audio signal is divided into frequency subbands and time-segmented into a time-frequency segmentation block. The length of the encoded sample depends on the shape of the temporal function of the audio signal. In the absence of sharp peaks in amplitude, a long sample is used, which provides high-frequency resolution. In the case of abrupt changes in signal amplitude, the length of the encoded sample decreases dramatically, giving a higher time resolution. The decision to change the length of the coded sample is made by the psychoacoustic analysis unit, calculating the value of the psychoacoustic entropy of the signal.
After segmentation, the frequency subband signals are normalized, quantized, and encoded. In the most efficient compression algorithms, it is not the samples of the audio signal that are encoded, but the corresponding MDCT coefficients. (the differential between the coefficients is smaller) The accounting of the auditory perception patterns of a sound signal is carried out in the psychoacoustic analysis unit. Here, according to a special procedure, for each frequency sub-band, the maximum allowable level of quantization distortion (noise) is calculated, in which they are still masked by the useful signal of this sub-band.

The block of dynamic distribution of bits according to the requirements of the psychoacoustic model for each coding subband selects a minimum possible number of them, in which the level of distortions caused by quantization does not exceed the threshold of their audibility calculated by the model psychoacoustic.

This article will consider the functional diagrams of the audio data compression algorithms, based on µ-laws, A. The functional diagram of the compression algorithm based on the A-level compression law is shown in Fig.2. Figure 2. Functional diagram of the compression algorithm based on the A-level compression law A signal (discrete sine) is applied to the input of the compressor. After compression, the signal passes to the adder, where the noise is fed to the second input of the adder, thus simulating the additive noise of the transmission channel.

Then the noisy signal enters the input of the expander, at the output we get the reconstructed signal. The reconstructed and original signal is then fed to the adder, after which the power of the spectral noise is observed.

Simulation results (A = 87.6)
The following graphs are presented: 1-original signal, 2-signal passed through the compressor, 3-recovered signal, 4-noise power at the output of the noise generator, 5-noise power after the expander.

Improved efficiency of digital audio data compression algorithms.

Improved efficiency of digital audio data compression algorithms.

audio compression

The relevance of the work. Methods for encoding high quality (HS) audio signals have become very widespread in the last decade in the field of broadcasting, digital sound recording, and home audio and video equipment. There’s even a fast-growing new class of consumer electronics: portable MP3 players.

Audio Compression:

Digital television and radio transmission networks are being developed, providing consumers with high-quality images and sound with a wide coverage area. The popularity of radio and television broadcasts over the Internet and mobile phone networks is increasing. All these technological innovations have become economically viable, and in some cases even technically possible, thanks to the use of highly efficient digital video and audio data compression algorithms, such as MPEG-1 ISO / IEC 11172, MPEG-2 TSO / IEC 13818, MPEG-4 ISO / IEC FCD 14496, ATSC Dolby AC-3. At the same time, due to the economic advantages of using these algorithms, which make it possible to reduce the bandwidth requirements of the transmission channels or the capacity of the information carriers by an order of magnitude, it is necessary to compensate with a certain decrease in the sound quality. During the era of the dominance of digital audio CDs, consumers have created a requirement for high sound quality from any sound reproduction equipment. The efforts of algorithm developers for encoding audio signals have always been aimed at ensuring that the quality of decoded audio material is no worse than that of a CD. Sound quality is often the determining factor in the economic success of digital broadcasting services or digital sound distribution services like iTunes). Further,

It is obvious that the problem of improving the quality of audio coding is today one of the key problems for the sound recording industry, the audio broadcasting industry and the manufacturers of various multimedia systems.

The basic principle of operation of highly efficient audio coding systems is to use the properties of the human auditory system, mainly the phenomenon of masking. The phenomenon of psychoacoustic masking is due to the biophysical and neuronal processing of sound signals by the human auditory system [173]. At the same time, part of the sound information does not affect the acoustic perception of the sound signal due to the presence of components with greater intensity in it. Therefore, the strongest components of the audio signal form the so-called masking thresholds. Sound information with a signal energy level below the masking threshold is not perceived by the auditory system. In the traditional digital representation of audio signals using pulse code modulation (PCM), time-sampled samples of the original signal are represented using a specific number of bits in the code word. The finite precision of the instantaneous values of a continuous analog signal introduces an error in the signal, the so-called quantization noise. The idea of encoding audio signals with the elimination of psychoacoustic redundancy is to combine psychoacoustic analysis and the quantization mechanism of audio signals [112]. In this case, the digitally encoded signal is converted into a time-frequency representation, as close as possible to the time-frequency resolution of the human auditory system. Psychoacoustic analysis determines the masking thresholds at each point in the time-frequency representation of the encoded signal, and the quantizer re-quantizes the signal with the minimum possible number of bits per sample, in which the increasing quantization noise is still below the masking thresholds. Thus, a compact representation of audio signals can be achieved without subjective degradation of sound quality. It is obvious that the efficiency and quality of such systems depend mainly on the precision of the psychoacoustic analysis. a compact representation of audio signals can be achieved without subjective degradation of sound quality. It is obvious that the efficiency and quality of such systems depend mainly on the precision of the psychoacoustic analysis. a compact representation of audio signals can be achieved without subjective degradation of sound quality. It is obvious that the efficiency and quality of such systems depend mainly on the precision of the psychoacoustic analysis.

Audio compression

Audio compression

Audio Compression

Well-established data compression methods such as RLE, statistical and dictionary methods can be used to compress lossless audio files, but the result is highly dependent on the specific audio data. Some sounds will compress well with RLE, but poorly with statistical algorithms. Statistical compression is more suitable for other sounds, but with a dictionary approach, on the contrary, expansion can occur. Here is a brief overview of the effectiveness of these three methods for compressing audio files.

Audio Compression

RLE works well with sounds that contain long series of repeating sound chunks – samples. With 8-bit sampling, this can happen quite often. Remember that the voltage difference between two 8-bit samples n and n – 1 is approximately 4 mV. A few seconds of homogeneous music, in which the sound wave changes by less than 4 mV, will generate a sequence of thousands of identical samples. With 16-bit sampling, obviously long repeats are less common and therefore the RLE algorithm will be less efficient.

Statistical methods assign variable length codes to audio samples according to their frequency. With 8-bit sampling, there are only 256 different samples, so the samples can be distributed evenly in a large audio file. A file of this type cannot be compressed well with the Huffman method. With 16-bit sampling, more than 65,000 sound bites are allowed. In this case, some samples may be more common and others less common. With a strong probability skew, good results can be achieved with the help of arithmetic coding.

Dictionary-based methods assume that some phrases will appear frequently throughout the file. This occurs in a text file in which individual words or sequences of them are repeated many times. However, the sound is an analog signal and the values of the specific generated samples are highly dependent on the operation of the ADC. For example, with 8-bit sampling, an 8 mV waveform becomes a numeric sample of 2, but a nearby wave of, say 7.6 mV or 8.5 mV, can be converted to a different number. For this reason, voice snippets that contain overlapping phrases and sound the same to us may differ slightly when digitized. Then they will enter the dictionary in the form of different phrases, which will not give the expected compression. Therefore, dictionary methods are not very suitable for audio compression.

You can achieve better results in lossy audio compression by developing compression techniques that take into account the perception of sound. They remove the part of the data that remains inaudible to the audience. It is like compressing images, discarding information invisible to the eye. In both cases, we assume that the original information (image or sound) is analog, that is, part of the information has already been lost during quantization and digitization. Allowing a little more loss with care will not affect the quality of the uncompressed sound reproduction, which will not differ much from the original. We will briefly describe two approaches called silence suppression and compaction.

The idea behind silence suppression is to treat small samples as if they were not there (i.e. they are zero). Such a zeroing will generate a series of zeros, so the method of suppressing pauses is, in fact, a variant of RLE adapted to audio compression. This method is based on the peculiarity of sound perception, which consists of the tolerance of the human ear to rule out barely audible sounds. Audio files containing long stretches of quiet sound will be better compressed using the silence suppression method than files full of loud sounds. This method requires the participation of the user, who will control the parameters that establish the loudness threshold for the samples. This requires two more parameters, which are not necessarily controlled by the user. One parameter is used to determine the shortest sequences of silent samples, usually 2 or 3. And the second sets the smallest number of consecutive strong samples, when silence or pause occurs. For example, 15 silent samples can be followed by 2 strong and then 13 silent,

Consolidation is based on the property that the ear better distinguishes changes in the amplitude of soft sounds than loud sounds. A typical ADC for computer sound cards uses a linear conversion to convert the voltage into a numerical form. If the amplitude a became n, then the amplitude 2 a will become 2 n.

LEARN HOW AUDIO DATA COMPRESSION WORKS

LEARN HOW AUDIO DATA COMPRESSION WORKS

Audio Data Compression

MP3s Around Us Many, many years ago, the Internet was supposed to be the force that would democratize the music industry, physical distribution was supposed to become obsolete, and it was possible to publish music on the Internet and be heard by millions of audiences.

Audio Data Compression

In fact, enthusiasts and companies have created websites where fans can listen to new tunes, the MP3 format has made it easy to place songs for critics, and music demo pieces are now helping to sell a CD or LP. physical. It is not difficult to put your music on the Internet, but if you are not a star of the first magnitude, you will have to accept the placement of the data in compressed format to save space on the server, as well as save download time for those who download your masterpiece. While there are many critics of MP3, there are ways around some of the limitations of this format.

The MP3 format is based on the use of data compression algorithms that can reduce the amount of data required to play music. Compression algorithms in MP3 work with loss of data, they do not work like Zip or Rar compression algorithms that restore original file without data loss. MP3 algorithms discard “unnecessary” data. For example, if there is a lot of high-level sound on a track, the algorithm may assume that you cannot hear low-level material and think that only 24 dB of dynamic range is sufficient for that part of the audio material. It only requires 4 bits of data, a quarter of the data needed for 16-bit resolution. Unfortunately, it is difficult to preserve the sound quality of music when compressed, but it is possible. One way is to use algorithms, working without data loss, such as FLAC, or some algorithms offered by Microsoft and Apple for their audio formats. However, these algorithms do not lead to a significant reduction in file size; with complex music, the size reduction can be only 10-20%.

Although there are many algorithms for compressing audio data, only a few are the most common:

MP3. This format allows multiple levels of encoding, you can create audio files of almost any size with a smaller size with greater loss of precision. There are many free and shareware MP3 players (such as iTunes and Windows Media Player), to encode MP3, you can use iTunes and most digital audio editors.

AAC. As the native iPod format, this format is quite popular and sounds better than MP3 for the same file size according to most users. ITunes can convert files to AAC.

Windows Media Audio. The format is promoted by Windows, but is used less frequently than MP3 or AAC. WMA sound quality is generally better than MP3. While Microsoft does not offer users WMA playback software for the Mac platform, the Flip4Mac utility (free version available) can play Windows Media formats on Mac.

Ogg Vorbis. A great but rarely used format that sounds better than MP3 at the same bit rate, and unlike MP3, the encoding tools are free for developers. Ogg Vorbis files are not widely used yet, but they are popular with advanced technical users.

FLAC. This popular lossless format is not supported by many portable music players, but musicians often use FLAC to exchange files when working on collaborative projects. High sound quality is maintained.

Although MP3 does not offer the best quality, this format is most often used when placing audio files on the network. all players can play MP3. It is important to choose the correct MP3 settings. When encoding files to MP3, it is always best to use a high-quality source file without compression. Then select the compression settings. When saving in MP3 format, you can generally choose from a range of bit rates (bits per second), from 320 kbps stereo (great quality, but also a fairly large file) to 8 kbps mono (good enough for dictation) . In addition to the fixed settings, there is variable bit rate (VBR) encoding, which optimizes the bit stream according to the playback material. VBR encoding is not supported by all players.