Tag: how lossless audio works

Audio and video analogic & digital

The appearance of multimedia systems, of course, brought about revolutionary changes in areas such as education, computer training, in many areas of professional activity, science, art, computer games, etc. But, you must agree, it is impossible to imagine the modern. multimedia systems without sound or video. In this work, I would like to dwell on the consideration of the fundamental differences in the representation of digital signals from analog, the characteristics of digital audio and video information, their compression algorithms (compression).

2. Differences between the digital representation of analog signals.

The traditional analog representation of signals is based on the similarity (similarity) of electrical signals (current and voltage changes) with the original signals represented by them (sound pressure, temperature, speed, etc.), as well as on the similarity of electrical signals. signal forms at various points in the transmission or amplification path. The shape of the electrical curve that describes (also called transfer) the original signal is as close as possible to the shape of the curve of this signal.

Such a representation is the most accurate, however, the slightest distortion of the shape of the electrical carrier signal will inevitably involve the same distortion of the shape and signal of the carrier. In terms of information theory, the amount of information in the carrier signal is exactly equal to the amount of information in the original signal, and the electrical representation does not contain redundancy that could protect the carried signal from distortion during storage, transmission. and amplification.

The digital representation of electrical signals is designed to add redundancy to avoid unwanted interference. For this, serious restrictions are imposed on the carrier electrical signal: its amplitude can take only two limit values: 0 and 1.

In this case, the entire zone of possible amplitudes is divided into three zones: the lower one represents zero values, the upper one, individual, and the middle one is prohibited, only interferences can enter. Therefore, any interference, the amplitude of which is less than half the amplitude of the carrier signal, does not affect the correct transmission of the values ​​0 and 1. Interference with a higher amplitude also does not affect whether the duration of the interference pulse is significantly shorter than the duration of the information pulse, and a filter is installed at the pulse noise input of the receiver.

The digital signal formed in this way can carry any useful information that is encoded in the form of a sequence of bits: zeros and ones; Electrical and sound signals are a special case of such information. Here, the amount of information in the digital carrier signal is much greater than in the original encoded signal, so that the carrier signal has some redundancy with respect to the original, and any distortion in the waveform of the carrier signal, which it still retains the ability of the receiver to correctly distinguish between zeros and ones, it does not affect the reliability of the signal transmitted by this information signal. However, in the case of significant interference, the shape of the signal can become so distorted that the precise transmission of the information being carried becomes impossible: errors appear in it, which, with a simple coding method, the receiver does not you can only correct, but also detect. To further increase the resistance of a digital signal to interference and distortion, two types of redundant digital coding are used: verification codes (EDC – Error Detection Code) and correction (ECC – Error Correction Code) . Digital encoding is simply adding extra bits to the original information and / or converting the original bit string into a longer string and other structure. EDC allows you to simply detect the fact of an error: a distortion or loss of a useful one or the appearance of a false digit, but the information transferred in this case is also distorted; ECC allows you to immediately correct detected errors, keeping the information that is transferred unchanged.

Each type of EDC / ECC has its own capacity limit to detect and correct errors, after which undetected errors and distortions of the information being transferred start anew. An increase in the amount of EDC / ECC relative to the amount of initial information generally increases the detection and correction capabilities of these codes.

Like EDC, the popular cyclic redundancy code CRC (Cyclic Redundancy Check), whose essence is the complex mixture of the initial information in the block and the formation of short binary words, whose bits have u

Audio compression algorithms for streaming purposes.

The problem of transmitting the necessary number of audio channels through a network of limited capacity forces us to resort to audio compression.

Despite the use of modern digital technologies, compression negatively affects sound quality and causes additional delay in signal transmission.

Currently, there are two fundamentally different approaches to compressing audio signals. This article will provide a general comparison between these two different compression principles. Also presented are graphs of the frequency response (amplitude frequency characteristic) of the sound sample in its original uncompressed form and after one cycle of encoding and decoding using MPEG Layer II and Enhanced apt-X.

Algorithms like MPEG and AAC use encoding using a psychoacoustic model of sound perception. Another approach is time encoding using Adaptive Differential PCM (ADPCM) in algorithms like Enhanced apt-X.

Linear PCM audio
Before compression, the audio is generally digitized in linear PCM format at 32 kHz, 44.1 or 48 kHz with a resolution of 16 or 24 bits.

The analog signal will be digitized in uncompressed digital PCM. The digital inputs of the codecs use oversampling to ensure conversion without timing issues. The uncompressed PCM signal is our benchmark for comparing compressed audio files.

MPEG Layer ll compression
MPEG 1 Layer ll is a widely used format. This is a typical example of a psychoacoustic perception coding algorithm that analyzes the incoming signal and compares it to a theoretical model to determine what frequency and what time domain information can be lost. The need to analyze the audio signal results in a mandatory delay, typically greater than 30 ms.

In theory, high compression ratios can be achieved, but even with relatively low compression, MPEG can seriously degrade audio quality. In Fig. 2 shows the frequency response after one pass of MPEG encoding of the source file.

Be aware of frequencies that are lost or distorted compared to original PCM audio.

Compression Enhanced Apt-X
Enhanced apt-X uses ADPCM audio processing technology. The signal is divided into four frequency bands that can be processed at a quarter of the original sample rate using a variable bit rate and variable quantization step. Since all processing is based on the time domain method, there is no delay other than the actual processing time required.

As a result, a 4: 1 compression ratio preserves the entire frequency content of the original signal with a coding delay of less than 3 ms. Frequency response graph in Fig. 3 shows the result of one pass encoding / decoding using Enhanced apt-X at 256 kbps and illustrates the high fidelity of Enhanced apt-X compared to the original uncompressed signal.

How Enhanced apt-X Works
The improved apt-X encoding algorithm passes the original PCM data through a specially designed two-stage Q-mirror filter to divide the signal into four subbands and reduce the clock frequency to 1/4 of the original clock frequency. The quantization procedure consists of processing four sub-signals to reduce each signal from 16 bits to 7 bits in subband 1, 4 bits in subband 2, 3 bits in subband 3 and 2 in subband 4.

The inverse quantizer and prediction scheme uses the above values ​​to predict the size of the next signal. This value is compared to the actual signal and the “difference” is measured. The encoder transmits this measured “difference” signal to the decoder. Each subband is processed in parallel and the output of the string quantizer and predictor is encoded with a predetermined resolution. The processing output of the four subbands is multiplexed into a single 16- or 24-bit enhanced apt-X signal. Then additional data and sync data is added to it for streaming.

What is digital audio data compression?

It could be said that there are two methods with which it is possible to compress the data in the case of digital audio.

On the one hand, the method known as lossy compression is intended to reduce psychoacoustic redundancy and the other method is to reduce statistical redundancy. And this is known as lossless compression.

Lossless compression

Many people wonder how lossless compression can be achieved.
The Huffman code takes into account the probability that levels of different magnitudes will appear, for example, the most probable values ​​to appear frequently are assigned shorter codes and, on the other hand, the values ​​whose probability of appearance is small are assigned they use longer code words.

If we think about it, we will realize that by replacing the signal values ​​that will appear very frequently with shorter words, this will save us space but it does not imply any loss of quality because no information is being discarded.

If we put an example perhaps very simplified to be able to explain this method, we could imagine that we are going to compress a text. Now suppose this text contains some words that may be repeated very frequently.
For this explanation suppose that our text contains the word “statistically” many times and suppose that we substitute 3 characters for it: xx ÷.
So in each place where the word “statistically” should appear, we will replace it with the characters “xx ÷” And if this word appears enough times we will reduce the size of the text.
If we do this with each of the words that will appear several times, we will be able to make a significant reduction of the text without having to discard any information.

Therefore, when we rebuild or decompress our file, we will obtain a file exactly the same as the initial one, without any loss.

The other method where if there is a loss what is sought is to discard what is audio information the ear cannot distinguish, for example in the masking effect.
This masking what happens in our ears due to the imperfection of the human ear, supposes that two sounds do occur at approximately the same moment And these sounds have a close frequency But one of them has a much higher volume, the ear will only perceive the one who has higher volume and the one with the lowest volume will not perceive. Therefore, it can be discarded and the human ear will not perceive the difference.

This type of compression does use the technique of discarding information, which is why the resulting file has some information loss.

In general terms, today those systems that act without loss of information are being highly valued.

Lossless vs lossy, what is the difference?

In a recent article on wireless audio, we addressed the topic of lossy and lossless digital audio encoding. Today we will dwell on this topic in more detail.

So, we have analog sound, which, during digital sound recording and / or for later storage in a computer and other electronic media, is digitized into an audio file, an electronic document consisting of information about the amplitude and frequency of the sound, with the help of which the digital-analog inverse conversion and reproduction of the sound contained in the file.

The sound format depends on the quantization method using an analog-to-digital converter (ADC), two types of quantization are widespread:

pulse code modulation (PCM, most MP3 to FLAC formats)

sigma-delta modulation (Delta-sigma, DSD format)

The main parameters of digital audio are the quantization bit (bit) and the sample rate (kHz / MHz), which are indicated for various recording and playback devices as the format to represent digital audio, for example, 24 bit / 192 kHz.

There are uncompressed audio formats (eg WAV, AIFF), but for more convenient storage / distribution, codecs that compress audio data are often used. Data compression (data compression) is performed in order to reduce the volume occupied by files and is based on eliminating the redundancy contained in the original data. There are two types of compressed formats:

Lossless: lossless compression (FLAC, ALAC, APE)

Lossy: Lossy compression (MP3, Ogg, AAC)

Lossless compression allows you to make a complete recovery of the original data, lossy compression allows you to recover data with certain distortions.

Lossy compression is significantly more efficient than lossless compression and is used when full compliance with original and recovered data is not required, and volume reduction is a priority.

A lossy encoded file is very different from the original on the level of byte comparison, but to an inexperienced human ear, the difference may not be as strong and sometimes even imperceptible. It does this by focusing lossy compression techniques on the physical characteristics of a person’s senses, such as a psychoacoustic model, which determines how much sound can be compressed without degrading the perceived quality of the person. Impairments caused by loss of compression that are perceptible to the human ear are considered compression artifacts.

MP3 spectrogram (left) and original file (right)

Examples of common lossy formats:
MP3: defined by the MPEG-1 specification, perhaps still the most common format

Ogg Vorbis: distinguished by the absence of patent restrictions and higher quality with the same bit rate as MP3

AAC, AAC + – Exists in various versions, defined by MPEG-2 and MPEG-4 specifications, it became widespread along with Apple technology

eAAC + is a format offered by Sony as an alternative to AAC and AAC +

WMA is a format developed by Microsoft

Dolby AC-3

DTS

Previously, lossless audio formats were most often used for archival data storage and in cases where distortion was unacceptable or undesirable, and most common listeners used music in lossy compressed formats. But the amount of memory in electronic devices is constantly growing and prices are falling, which is why more and more people are switching to listening to Lossless formats, which allow them to perceive music in its original form. In addition, the support for Lossless formats has now appeared on almost all consumer devices, even some streaming services are beginning to broadcast sound in lossless quality, for example, Deezer presented in Russia or Tidal, which is officially absent from us.

Examples of common lossless formats:
Free lossless audio codec: FLAC is the most common free format

ALAC – Apple Lossless Audio Codec – Apple variant

Lossless audio encoding, also known as MPEG-4 ALS

Direct Flow Transfer – DST

Dolby TrueHD

DTS-HD Master Audio

Meridian Lossless Packing – MLP

Monkey’s Audio – Monkey’s Audio APE

WavPack – Lossless WavPack

WMA Lossless – Windows Media Lossless