Psychoacoustic Models in MP3 and AAC Encoding

Free Download Mp4Gain

Psychoacoustic Models in MP3 and AAC Encoding

Let’s talk about Psychoacoustic Models in MP3 and AAC Encoding

When it comes to digital audio compression, especially in MP3 and AAC formats, psychoacoustic models are the secret sauce that makes it all work. These models allow us to shrink large audio files into much smaller sizes without a noticeable loss in sound quality. In my years of working with audio encoding, I’ve seen how these models have revolutionized the way we perceive sound after compression. The core idea is simple: we don’t hear all sounds equally. Some frequencies and nuances are more noticeable than others, and psychoacoustic models exploit this fact to make compression more efficient.

Think of it like this: imagine you’re at a concert, and a loud bass guitar is playing alongside a softer violin. Your attention is drawn to the bass because it’s much louder, and the violin’s subtle details get masked. This is exactly what psychoacoustic models do—they remove or reduce sounds that are unlikely to be heard due to masking effects. In this article, I’ll walk you through how psychoacoustic models in MP3 and AAC encoding work and why they matter for audio quality and file size.

Understanding the Basics of Psychoacoustic Models

Psychoacoustic models are based on the science of how our ears and brain perceive sound. They take into account how different sounds mask each other, which frequencies we are most sensitive to, and how we interpret sound in different contexts. MP3 and AAC encoding use these models to compress audio by identifying and removing information that won’t be noticeable to the listener.

A simple analogy would be taking a photograph with a high-resolution camera and then reducing its size by removing some pixels. You won’t notice much difference in the quality of the image because you can’t see all the pixels. Similarly, these audio encoders remove frequencies or audio details that the human ear won’t detect, making the audio file smaller without compromising its perceived quality.

Frequency Masking

Frequency masking happens when a louder sound in one frequency range makes a softer sound in a nearby frequency range inaudible.
Psychoacoustic models use this to discard or reduce the quieter, masked sounds, optimizing compression.
For example, if a heavy guitar is playing at a loud volume, the model might remove the higher-pitched background notes that are masked by the louder guitar.

Temporal Masking

Temporal masking occurs when one sound, like a sharp drum hit, can mask a quieter sound that occurs immediately after it.
This type of masking is crucial for determining which transient sounds can be removed in compression.
For instance, a loud snare hit can mask a subtle violin note that comes milliseconds after, making it unnecessary to keep all the data for that note.

The Role of Psychoacoustic Models in MP3 Encoding

In MP3 encoding, psychoacoustic models play a critical role in reducing the file size while maintaining an acceptable level of sound quality. The MP3 codec was one of the first to use psychoacoustic models to exploit human hearing limitations, and it was revolutionary when it was introduced in the 1990s. The encoder divides audio into different frequency bands and applies masking principles to decide which data can be discarded.

What’s fascinating is that MP3 uses a hybrid of time-domain and frequency-domain processing. It first splits the audio into small segments and then performs a frequency analysis. Using this information, the encoder decides which frequencies can be reduced or eliminated entirely. By doing this, the model allows the MP3 format to achieve relatively small file sizes while preserving the overall listening experience.

MP3 and the Trade-off Between Compression and Quality

MP3 encoding sacrifices some of the finer audio details to reduce file size.
The trade-off is more noticeable at lower bitrates, where artifacts like compression noise or a “tinny” sound may become audible.
Higher bitrates, like 192 kbps or 256 kbps, provide better sound quality, though the file size increases.

AAC: The Next Generation of Psychoacoustic Modeling

While MP3 revolutionized audio compression, AAC (Advanced Audio Codec) takes things a step further. As a more advanced codec, AAC uses a refined psychoacoustic model that performs better at lower bitrates, providing higher-quality audio with less data. This is especially important for modern audio streaming services, which need to balance high-quality sound with efficient bandwidth usage.

The AAC psychoacoustic model is more sophisticated, taking into account additional factors like stereo imaging and spatial effects. It’s also more adept at handling complex audio, such as orchestral music or tracks with a wide range of dynamics. From my experience, AAC does a better job than MP3 in preserving the subtleties of sound, especially at lower bitrates, which is why I recommend it over MP3 when available.

Why AAC Outperforms MP3

AAC uses more advanced psychoacoustic techniques, making it more efficient at lower bitrates.
It better preserves transient sounds and complex audio elements, like the reverberations of a piano or the nuances of a singer’s voice.
With AAC, you can get excellent sound quality at 128 kbps, whereas MP3 may require 192 kbps or higher for a similar result.

How Psychoacoustic Models Help with Audio Quality at Low Bitrates

One of the most remarkable aspects of psychoacoustic models is how they enable high-quality audio at low bitrates. At lower bitrates, many codecs, including MP3 and AAC, might introduce artifacts such as distortion or loss of clarity. However, psychoacoustic models allow the encoder to focus on the most important elements of the sound—those that we are most likely to notice—while discarding the less important parts.

This is especially noticeable in AAC, where the advanced psychoacoustic model ensures that even at low bitrates, the encoding still captures essential auditory information, such as pitch, rhythm, and timbre. I’ve personally found that with AAC, even at 128 kbps, I can enjoy clear vocals and instruments without the harsh artifacts that often accompany MP3 at the same bitrate.

Latest Words on Psychoacoustic Models in MP3 and AAC Encoding

Psychoacoustic models are an integral part of both MP3 and AAC encoding, helping us achieve smaller file sizes while preserving audio quality. These models allow the encoder to reduce the file size by removing sounds that are less perceptible to the human ear, making the audio more efficient without sacrificing what matters most to the listener. While MP3 was groundbreaking in its time, AAC offers superior compression and better handling of complex audio, making it the better choice for modern audio applications.

As I’ve discussed throughout this article, these psychoacoustic models are crucial in ensuring that we can enjoy high-quality audio, even with file sizes that fit comfortably on our devices and bandwidth constraints. Whether you’re listening to your favorite album or streaming a podcast, psychoacoustic models are working behind the scenes to make your audio experience better. As the technology continues to improve, we can only expect even better performance in the future.

Frequently Asked Questions

What are psychoacoustic models in MP3 and AAC encoding?

Psychoacoustic models in MP3 and AAC encoding are based on the way humans perceive sound. These models analyze how different frequencies mask each other, allowing the codecs to remove or reduce the data for sounds that are less noticeable to the human ear. This process helps reduce file size without sacrificing audio quality. Essentially, psychoacoustic models optimize compression by focusing on the most important sounds in an audio file.

How do psychoacoustic models improve audio compression?

Psychoacoustic models improve audio compression by eliminating or reducing sounds that the human ear is less sensitive to. For example, louder sounds can mask softer ones, so the encoder can discard those quieter sounds, saving space without impacting the perceived quality of the audio. This makes it possible to compress audio files into smaller sizes while still delivering high-quality sound, especially in formats like MP3 and AAC.

What is the difference between MP3 and AAC in terms of psychoacoustic models?

The main difference between MP3 and AAC lies in the sophistication of their psychoacoustic models. AAC has a more advanced model that better handles complex audio, such as classical music or tracks with subtle dynamic changes. It also performs better at lower bitrates compared to MP3, providing higher sound quality at the same compression level. In short, AAC offers superior compression efficiency, especially when dealing with modern audio formats and streaming.

Why does AAC sound better than MP3 at lower bitrates?

AAC sounds better than MP3 at lower bitrates because it uses a more efficient psychoacoustic model. The AAC codec is designed to optimize the way it removes or reduces sounds, prioritizing the frequencies that are most important for human perception. This allows it to achieve a better balance between file size and audio quality, especially at bitrates like 128 kbps, where MP3 might begin to show noticeable artifacts.

How does temporal masking affect audio compression?

Temporal masking occurs when a loud sound at one moment in time masks a softer sound that follows it almost immediately. This effect is important for audio compression because it allows the encoder to discard these masked sounds without the listener noticing. This type of masking helps improve compression efficiency, especially in formats like MP3 and AAC, where transient sounds, like a snare hit or cymbal crash, may cover quieter background elements.

Can psychoacoustic models cause distortion in compressed audio?

While psychoacoustic models aim to reduce file size without degrading sound quality, they can sometimes introduce distortion, particularly at lower bitrates. This happens when the codec removes too much data, resulting in noticeable artifacts such as a “tinny” or metallic sound. However, with modern codecs like AAC, these artifacts are much less common, even at lower bitrates, thanks to more advanced psychoacoustic modeling.

Comments:

Wow, I had no idea how much science goes into these audio codecs. Your explanation about frequency and temporal masking really helped me understand why AAC sounds better at lower bitrates. Great article! – AudioFan77

I’ve always been a fan of MP3, but now I’m definitely considering switching to AAC for my music collection. The way you described the differences in psychoacoustic models makes it so much clearer! Thanks! – MusicJunkie88

This article is awesome! The real-life examples helped me visualize how psychoacoustic models work. I never understood how my music could sound so good at a low bitrate, but now I get it. Thanks for the great info! – SoundLover42

Can you talk more about how AAC handles high-frequency sounds compared to MP3? I’d love to know more about that! Great article though, very informative. – HighFreqFan

I didn’t realize how important these psychoacoustic models were in compressing audio. I always wondered how audio streaming services maintain such high-quality sound at lower bitrates. Now I know! – DeeJayDave

This is one of the most detailed articles on this topic I’ve found! I’ve been using AAC for a while now, but this article really made me appreciate how much better it is than MP3, especially for complex audio. – SoundEngineerX

Excellent breakdown of the differences between MP3 and AAC. I always assumed MP3 was “good enough” but now I realize AAC is the better choice, especially for lower bitrates. Thanks for clearing that up! – TechieTom

Great read, but I wish you would’ve gone deeper into how these psychoacoustic models impact the experience for listeners with hearing impairments. Any chance you can dive into that next? – ClearSound76

As a musician, I’ve always been picky about sound quality. After reading this, I’m convinced that AAC is worth the switch for my music files. Thanks for sharing your expertise! – MusicMaker24

I had no idea that psychoacoustic models were so important for compression. I always assumed audio codecs just “squished” the data and that was it! – CuriousGeorge

Very well-written article! I didn’t know much about psychoacoustics before, but now I understand why AAC sounds better at lower bitrates. Thanks for breaking it down so clearly! – TuneInExpert

Free Download Mp4Gain

Mp4Gain Main Window

Mp4Gain Features

Free Download Mp4Gain

MP3 Bit Allocation

What Are the Key Principles Behind MP3 Bit Allocation?

Latest Words on MP3 Bit Allocation

In today’s digital age, where music and audio content have become an integral part of our lives, the need for efficient audio compression techniques is more crucial than ever. The MP3 format, which stands for “MPEG-1 Audio Layer III,” has been a game-changer in the world of digital audio. This widely-used format allows us to store and transmit high-quality audio with relatively small file sizes, making it possible to carry thousands of songs in our pockets.

The magic behind the MP3 format lies in its bit allocation principles. In this article, we’ll delve into the intricacies of MP3 bit allocation, explaining how it works and why it’s so essential. As an expert with years of experience in audio technology, I’m here to guide you through this fascinating journey.

Let’s Talk About MP3 Bit Allocation

Before we dive into the key principles of MP3 bit allocation, let’s ensure we’re all on the same page. You might be wondering what “bit allocation” even means. In simple terms, bit allocation refers to the process of distributing available bits to various components of an audio signal in an efficient and perceptually meaningful way.

Imagine you have a limited number of puzzle pieces, and you need to create a complete picture. Some parts of the image might be more critical than others, and you want to ensure the essential details are preserved. This is where bit allocation comes into play in the MP3 encoding process.

Now, let’s get deeper into the principles behind MP3 bit allocation.

The Psychoacoustic Model: A Vital Component

At the core of MP3 bit allocation is the psychoacoustic model. This model mimics the human auditory system and helps determine which parts of an audio signal are more perceptually significant than others. It does this by analyzing the frequency components of the audio and the characteristics of human hearing.

Imagine you’re in a room filled with people talking at various volumes. Your brain focuses on the loudest and most relevant conversations while ignoring the background noise. Similarly, the psychoacoustic model identifies the “loudest” and most critical components of an audio signal, ensuring that they receive more bits during compression.

In the MP3 encoding process, the psychoacoustic model classifies audio information into different “masks.” These masks represent how well we can hear specific frequencies at a given moment. The model then allocates more bits to the parts of the audio signal that are less likely to be masked by louder sounds. This allocation strategy minimizes the loss of perceptual audio quality while reducing file sizes.

Masking Effect: An Everyday Analogy

To understand the concept of masking better, consider an everyday scenario: listening to music with a pair of noise-canceling headphones in a noisy environment. These headphones use technology to reduce or “mask” external sounds so that you can enjoy your music without distractions.

Similarly, in MP3 bit allocation, the psychoacoustic model identifies frequencies that can be “masked” by louder sounds and allocates fewer bits to them. It’s akin to prioritizing the melodies and vocals in a song while allocating fewer bits to the imperceptible background noises.

This approach is what makes MP3 compression so efficient. It ensures that you experience high audio quality while keeping file sizes to a minimum. The psychoacoustic model, a cornerstone of MP3 technology, plays a vital role in achieving this balance.

The Bit Reservoir: Ensuring Smooth Playback

Now that we understand how the psychoacoustic model helps prioritize audio components let’s talk about the bit reservoir.

Comments:

Comment 1.

I really enjoyed this article! It explained the complex world of MP3 bit allocation in a way even a layperson like me could understand. Great job!

Comment 2.

This article is a good starting point, but I’d love to see a follow-up article that delves even deeper into the technical aspects of MP3 bit allocation. Keep up the good work!

Comment 3.

Kudos to the author for making such a technical topic accessible. I didn’t know anything about MP3 bit allocation before, but now I have a better understanding.

Comment 4.

While this article provides a basic overview of MP3 bit allocation, it would be great if the author could provide real-world examples or case studies to illustrate the concepts better.

Comment 5.

Great explanation! It’s nice to read an article written by someone who knows their stuff. Keep writing more on audio technology, please.

Comment 6.

This article covers the fundamentals well. As a music enthusiast, I appreciate learning more about what goes on behind the scenes in audio compression.

Comment 7.

Wow, I had no idea MP3s were so complex. The part about the psychoacoustic model was fascinating. I look forward to reading more from this author.

Comment 8.

This article could benefit from more practical applications. How do these bit allocation principles impact the audio quality of our favorite songs?

Comment 9.

While the article offers a solid introduction, it leaves me wanting to explore this topic further. It’s a compelling read that piques curiosity.

Comment 10.

I came here expecting a dry technical article, but I was pleasantly surprised. The analogy with noise-canceling headphones was spot on.

Comment 11.

I appreciate the clear and concise language in this article. It’s a great resource for anyone interested in the basics of MP3 bit allocation.

Comment 12.

More, please! I can’t get enough of this topic now. Looking forward to part two. Thanks for making this accessible to the average reader.

The Science of Audio Encoding: Technical Aspects

Audio encoding is the process of converting analog sound into digital data. This data can then be stored or transmitted in a variety of formats, such as WAV, MP3, or AAC.

There are two main types of audio encoding: lossless and lossy. Lossless encoding preserves all of the original sound data, resulting in high-quality audio but large file sizes. Lossy encoding removes some of the original sound data, resulting in smaller file sizes but lower sound quality.

The process of audio encoding can be divided into three main steps: sampling, quantization, and compression.

Sampling

The first step in audio encoding is sampling. In this step, the analog sound signal is converted into a series of discrete values. The number of times per second that the sound signal is sampled is called the sample rate. Higher sample rates result in more accurate representations of the original sound signal, but they also result in larger file sizes.

Quantization

The second step in audio encoding is quantization. In this step, each sample value is rounded to the nearest integer value. The number of bits used to represent each sample value is called the bit depth. Higher bit depths result in more accurate representations of the original sound signal, but they also result in larger file sizes.

Compression

The third and final step in audio encoding is compression. In this step, the digital audio data is compressed to reduce its file size. There are a number of different compression algorithms that can be used, each with its own advantages and disadvantages.

The most common compression algorithms for audio encoding are:

MP3: MP3 is a lossy compression algorithm that is widely used for storing and transferring audio files. MP3 files are typically much smaller than WAV files, while still providing good sound quality.
AAC: AAC is another lossy compression algorithm that offers better sound quality than MP3. AAC files are typically slightly larger than MP3 files, but they offer a noticeable improvement in sound quality.
FLAC: FLAC is a lossless compression algorithm that offers similar sound quality to WAV, but with much smaller file sizes. FLAC files are a good choice for people who want the best possible sound quality without sacrificing file size.

Final Words

Audio encoding is a complex process that involves converting analog sound into digital data. The quality of the audio that is encoded can be affected by a number of factors, including the sample rate, bit depth, and compression of the audio file.

If you are looking for the best possible sound quality, you should use a lossless audio format such as WAV or FLAC. However, if you need to store or transfer audio files over a network, you should use a lossy audio format such as MP3 or AAC.

MP3 finally goes into the public domain

MP3 finally goes into the public domain

mp3

Open Source

Mp3 Public Domain

Perhaps many did not think so, but the mp3 standard so well known to all had problems with the purity of patents. On April 23, 2017, the last patents expired and the format was finally free. Technicolor has officially stopped collecting royalties from manufacturers of software and embedded solutions.

License

Although hardware mp3 decoding is built into all other coffee machines, until recently its use in commercial projects required royalties from the developer: Fraunhofer Society. In 2005 alone, the amount paid was one hundred million euros. Most of the patents became invalid in the European Union in 2012. However, some of them continued to operate in the United States due to peculiarities of local law. What does this news bring to the community? At least now it will be possible to compile Gentoo and listen to music at the same time immediately on the base distribution. Many distributions will be able to provide support for the standard to the main repository. Now, for example, Ubuntu itself requires the installation of non-free components from a separate Ubuntu Restricted Extras meta-package to support mp3.

Bourbon vanilla vs vanillin

How does this standard, which has been the main standard in this area for 24 years, despite many more advanced free options? mp3 is in many ways similar in principle to its cousin in the photo world: JPEG. Due to the imperfection of our hearing aid and the peculiarities of psychoacoustics, it is possible to “discard” those parts of the audio spectrum that do not make a significant contribution to the musical pattern. In particular, in the illustration above, you can see how the amount of information encoded in the high-frequency region increases.

High frequencies are often sacrificed for the sake of preserving detail in the lower region – vocals, most instruments (thanks for the comment, KorDen32). Standard values of cutoff frequencies for the lame encoder:

CBR 096 kbps: 14000 – 15000 Hz;
CBR 112 kbps: 15000-15600 Hz;
CBR 128 kbps: 16000 – 16500 Hz;
CBR 160 kbps: 16500-17500 Hz;
CBR 192 kbps: 18000-18700 Hz;
CBR 224 kbps: 19000-19400 Hz;
CBR 256 kbps: 19500-19700 Hz;
CBR 320 kbps: 20,000 – 21,000 Hz.

The method can be compared to the creativity of flavor chemists. You’ve probably noticed that strawberry gum is very conventionally strawberry, and there isn’t enough lemon in synthetic lemon tea. Any natural flavoring composition contains dozens and even hundreds of chemical compounds. But the main core generally creates only a very limited amount. So, for example, vanillin defines most of the aroma of natural vanilla, and if you don’t appreciate the subtle nuances too much, the remaining components can be neglected. mp3 uses the same principles, removing insignificant portions of the spectrum. Most people cannot tell the lossless formats by ear from the normally encoded 320kbps mp3s, which saves a lot of space when storing your media library.

Audio Coding: Secrets Revealed Part 2

Audio Coding: Secrets Revealed Part 2

Bit Depth

Bit depth

audio encoding

Along with the sample rate, there is the bit depth or depth of the sound. Bit depth is the number of bits of digital information to encode each sample. Simply put, the bit depth determines the “accuracy” of the input signal measurement. The larger the digit capacity, the smaller the error will be for each individual conversion from the magnitude of an electrical signal to a number and vice versa. With the smallest possible bit depth, there are only two options for measuring sound accuracy: 0 for full silence and 1 for full sound. If the bit width is 8 (16), then by measuring the input signal, 2 8 = 256 (2 16 = 65,536) different values can be obtained.

Bit depth is fixed in the PCM codec, but for codecs that assume compression (eg MP3 and AAC), this parameter is calculated during encoding and may vary from sample to sample.

Bitrate
Bit rate is an indicator of the amount of information that one second of sound encodes. The higher it is, the less distortion and the closer the encoded composition is to the original. For linear PCM, the bit rate is very easy to calculate.

bitrate = sample rate × bit depth × channels

For systems like the Epiphan Pearl Mini that encode 16-bit (16-bit) linear PCM, this calculation can be used to determine how much additional bandwidth the PCM audio might require. For example, for stereo (two channels), the signal is digitized at 44.1 kHz at 16 bits and the bit rate is calculated as follows:

44.1 kHz × 16 bit × 2 = 1411.2 kbps

Meanwhile, audio compression algorithms like AAC and MP3 have fewer bits to transmit the signal (that’s their purpose), so they use low bit rates. Typically, the values are in the range of 96 kbps to 320 kbps. For these codecs, the higher the bit rate you choose, the more audio bits you get per sample and the better the sound quality.

Sample rate, bit depth and bit rates in real life.
Audio CDs, one of the most popular early inventions for the general public for storing digital audio, used 44.1 kHz (20 Hz – 20 kHz, human ear range) and 16 bits. These values were chosen to be able to save as much audio as possible to disk with good sound quality.

When video was added to audio and DVD and then Blu-ray discs came along, a new standard was created. DVD and Blu-Ray recordings typically use 48 kHz (stereo) or 96 kHz (5.1 surround) linear PCM format and 24-bit depth. These settings have been selected as ideal for keeping audio in sync with video while obtaining the best possible quality using the additional available disk space.

Our recommendations
CDs, DVDs, and Blu-Ray discs all have one goal: to provide the consumer with a high-quality playback engine. The goal of all developments was to provide high-quality audio and video without worrying about file size (if only it could fit on disk). Such quality could be provided by linear PCM.

In contrast, mobile media and streaming media have a completely different goal: to use the lowest bit rate, as low as possible, while still being sufficient to maintain acceptable quality for the listener. Compression algorithms are best suited for this task. You can follow the same principles for your records.

When recording audio from a video …
In case the record is used for the next on-ra-ki-bot, choose the 48 kHz PCM codec and the maximum bit depth (16 or 24) to provide the best audio quality. We recommend these parameters for Epiphan Pearl Mini.

When streaming audio from video …
With streaming or recording for later translation, good sound can be obtained with less bandwidth, using MP3 or AAC codecs with a frequency of 44.1 kHz and a bit rate of 128 kbit / s or higher. These parameters ensure that the sound is good enough without affecting the quality of the transmission.

Audio encoding: secrets revealed

Audio encoding: secrets revealed

Audio Encoding

Audio settings for video capture and transmission.

audio and video encoding

As people directly related to the AV sphere, we constantly talk about audio coding and audio codecs, but what is it? An audio codec is essentially a device or algorithm that can encode and decode a digital audio signal.

In practice, the audio waves that travel through the air are continuous analog signals. The signals are converted to digital form by a device called an analog-to-digital converter (ADC), and the reverse converter is called a digital-to-analog converter (DAC). The codec lies between these two functions and it is he who allows you to adjust some important parameters for the successful capture, recording and transmission of an audio signal: the codec algorithm, the sampling frequency, the bit width and the speed of the audio signal. data.

The three most popular audio codecs are Pulse-Code Modulation (PCM), MP3, and Advanced Audio Coding (AAC). The choice of codec determines the compression rate and the recording quality. PCM is a codec used by computers, CDs, digital phones, and sometimes SACD. The PCM signal source is sampled at regular intervals, and each sample is the digital amplitude of the analog signal. PCM is the simplest option for digitizing an analog signal.

With the correct parameters, this digitized signal can be completely converted back to analog without any loss. But this codec, which provides an almost complete identity with the original audio, is unfortunately not very cheap, which translates into very large file sizes, and such files are not suitable for streaming. We recommend using PCM to record digital images for your sources or when doing audio post-processing.

Fortunately, we always have the option of choosing a different codec that can compress digital data (rather than PCM) based on some helpful observations on the behavior of sound waves. But in this case, you have to make a compromise: all alternative algorithms are associated with “losses”, since it is impossible to completely restore the original signal, but nevertheless the result is still so good that most users will not be able to to catch the difference.

MP3 is an audio encoding format that uses a digital data compression algorithm that allows you to save the audio signal in smaller files. The MP3 codec is the most used by users to record and store music files. We recommend using MP3 to stream audio content as it requires less network bandwidth.

AAC is a newer audio encoding algorithm that is the successor to MP3. AAC has become the standard for MPEG-2 and MPEG-4 formats. In fact, this is also a digital data compression codec, but with less quality loss than MP3 when encoded with the same bit rate. We recommend using this codec for online streaming.

Sampling frequency (kHz, kHz)
Sample rate (or sample rate): the frequency with which the signal is digitized, stored, processed, or converted from analog to digital. Time sampling means that the signal is represented by several of its samples (samples) taken at regular intervals.

Measured in hertz (Hz, Hz) or kilohertz (kHz, kHz,) 1 kHz equals 1000 Hz. For example, 44,100 samples per second can be labeled 44,100 Hz or 44.1 kHz. The selected sample rate will determine the maximum playback frequency and, as follows from Kotelnikov’s theorem, to fully restore the original signal, the sample rate must be twice the highest frequency in the signal spectrum.

As you know, the human ear is capable of picking up frequencies between 20 Hz and 20 kHz. Given these parameters and the values shown in the table below, you can understand why 44.1 kHz was chosen as the sampling frequency for CD and is still considered a very good frequency for recording.

There are several reasons for choosing a higher sample rate, although it may seem like a waste of time and effort to reproduce sound outside the range of human hearing. At the same time, 44.1 – 48 kHz will suffice for the average listener for a high-quality solution to most problems.

Audio encoding and processing

Audio encoding and processing

Encoding

Sound information.

ENCODING

Sound is a wave that travels through air, water, or other medium with a continuously changing intensity and frequency.

A person perceives sound waves (air vibrations) with the help of hearing in the form of sound of different volume and pitch. The higher the intensity of the sound wave, the louder the sound, the higher the frequency of the wave, the higher the pitch of the sound

The human ear perceives sound at a frequency of 20 vibrations per second (low sound) to 20,000 vibrations per second (high sound).

A person can perceive sound in a wide range of intensities, in which the maximum intensity is 10 14 times greater than the minimum (one hundred thousand billion times). A special unit “decibel” (dbl) is used to measure the volume of sound (Table 5.1). Decreasing or increasing the volume of the sound by 10 dB corresponds to a decrease or increase in the intensity of the sound by 10 times.

Table 5.1. Sound volume
Sonar Volume in decibels
Lower limit of human ear sensitivity 0
Whisper of Leaves 10
Conversation 60
Horn 90
Jet engine 120
Pain threshold 140
Sound time sampling. For a computer to process sound, a continuous audio signal must be converted to a discrete digital form using time sampling. A continuous sound wave is divided into separate small time sections, for each section a certain value of sound intensity is set.

Therefore, the continuous dependence of the loudness of the sound at time A (t) is replaced by a discrete sequence of loudness levels.

Sampling frequency.

A microphone connected to the sound card is used to record analog sound and convert it to digital format. The quality of the digital sound obtained depends on the number of measurements of the sound volume level per unit time, that is, the sampling frequency. The more measurements that are made in 1 second (the higher the sampling frequency), the more accurately the “ladder” of the digital audio signal repeats the curve of the dialogue signal.

The audio sample rate is the number of measurements of the volume of a sound in one second.

The audio sample rate can range from 8000 to 48000 sound volume measurements per second.

Audio encoding depth. Each “step” is assigned a specific value for the volume level of the sound. Loudness levels of sound can be viewed as a set of possible states N, for which a certain amount of information is needed to encode, which is called audio encoding depth.

Audio encoding depth is the amount of information required to encode the discrete volume levels of digital audio.

If the known encoding depth, the number of digital audio volume levels can be calculated using the formula N = 2 I. Let the sound encoding depth be 16 bit, then the number of sound volume levels is:

N = 2 I = 2 16 = 65 536.

During the encoding process, each sound volume level is assigned its own 16-bit binary code, the smallest sound level will correspond to the code 0000000000000000 and the highest, 1111111111111111.

The quality of digitized sound. The higher the sound sampling frequency and depth, the better the digitized sound will sound. The lowest quality of digitized sound, corresponding to the quality of telephone communication, is obtained at a sampling rate of 8000 times per second, a sampling rate of 8 bits, and by recording an audio track (“mono” mode). The highest quality digitized audio, corresponding to the quality of an audio CD, is achieved with a sampling rate of 48,000 times per second, a sampling rate of 16 bits, and the recording of two audio tracks (“stereo” mode ).

It should be remembered that the higher the quality of the digital sound, the greater the volume of information in the audio file. It is possible to estimate the information volume of a digital stereo sound file with a duration of 1 second with an average sound quality (16 bits, 24,000 measurements per second). To do this, the encoding depth must be multiplied by the number of measurements in 1 second and multiplied by 2 (stereo sound):

16 bits × 24,000 × 2 = 768,000 bits = 96,000 bytes = 93.75 KB.

Audio Coding: Secrets Revealed – Part 2

Audio Coding: Secrets Revealed – Part 2

AUDIO ENCODING

Audio settings for video capture and transmission.

AUDIO ENCODING

There are several reasons for choosing a higher sample rate, although it may seem like a waste of time and effort to reproduce sound outside the range of the human ear. At the same time, 44.1 – 48 kHz will suffice for the average listener for a high-quality solution to most problems.

Bit depth
Along with the sample rate, there is the bit depth or depth of sound. Bit depth is the number of bits of digital information to encode each sample. Simply put, the bit depth determines the “accuracy” of the input signal measurement. The larger the digit capacity, the smaller the error for each individual conversion from the magnitude of an electrical signal to a number and vice versa. With the smallest possible bit depth, there are only two options for measuring sound accuracy: 0 for full silence and 1 for full sound. If the bit width is 8 (16), then by measuring the input signal, 2 8 = 256 (2 16 = 65,536) different values can be obtained.

Bit depth is fixed in the PCM codec, but for codecs that assume compression (eg MP3 and AAC), this parameter is calculated during encoding and may vary from sample to sample.

bitrate = sample rate × bit depth × channels

For systems such as the Epiphan Pearl Mini that encode 16-bit (16-bit) linear PCM, this calculation can be used to determine how much additional bandwidth the PCM audio might require. For example, for stereo (two channels), the signal is digitized at 44.1 kHz at 16 bits and the bit rate is calculated as follows:

44.1 kHz × 16 bit × 2 = 1411.2 kbps

When video was added to audio and DVD and then Blu-ray discs came along, a new standard was created. DVD and Blu-Ray recordings typically use 48 kHz (stereo) or 96 kHz (5.1 surround) linear PCM format and 24-bit depth. These settings have been chosen as ideal for keeping the audio in sync with the video while obtaining the best possible quality using additional available disk space.

By contrast, mobile media and streaming media have a completely different goal: to use the lowest bit rate possible, while still being sufficient to maintain acceptable quality for the listener.

Audio encoding: secrets revealed

Audio encoding: secrets revealed

audio encoding

Audio settings for video capture and transmission.

AUDIO ENCODING

With the correct parameters, this digitized signal can be completely converted back to analog without any loss. But this codec, which provides an almost complete identity with the original audio, is unfortunately not very cheap, which results in very large file sizes, and such files are not suitable for streaming. We recommend using PCM to record digital images for your sources or when doing audio post-processing.

AAC is a newer audio encoding algorithm that is the successor to MP3. AAC has become the standard for the MPEG-2 and MPEG-4 formats. In fact, this is also a digital data compression codec, but with less quality loss than MP3 when encoded with the same bit rate. We recommend using this codec for online streaming.

Digital audio encoding

Digital audio encoding

Digital audio encoding

PC-based audio coding is based on the process of converting air vibrations into electrical current fluctuations and the subsequent sampling of an analog electrical signal.

DIGITAL AUDIO ENCODING

The encoding and reproduction of audio information is carried out using special programs. The quality of reproduction of the encoded sound depends on the sampling frequency and its resolution (sound encoding depth – the number of levels).

Digital audio is an analog audio signal represented by discrete numerical values of its amplitude.

Sound digitization is a technology with a divided time step and subsequent recording of the values obtained in numerical form. Another name for digitizing audio is analog to digital audio conversion, which includes the following operations:

Bandwidth limiting is done by using a low pass filter to suppress spectral components that are more than half the sample rate.

Time sampling, that is, replacing a continuous analog signal with a sequence of its values at discrete moments of time: samples.

Level quantization is the replacement of the signal’s reference value with the closest value of a set of fixed values: quantization levels.

Encoding or digitization, as a result of which the value of each quantized sample is represented as a number corresponding to the ordinal number of the quantization level.

This is done as follows: a continuous analog signal is “cut” into sections with a sample rate, a discrete digital signal is obtained, which goes through the quantization process with a certain bit depth, and is then encoded, that is, it is replaced by a sequence of code symbols. To record sound in a 20-20,000 Hz frequency band, a sampling frequency of 44.1 and higher is required (today there are ADCs and DACs with a sampling frequency of 192 and even 384 kHz). To obtain a high-quality recording, 16-bit is sufficient, however, to expand the dynamic range and improve the quality of the sound recording, 24 (less often 32) bits are used.

Sound coding methods (of course an electrical signal coming from a microphone) are based on the fact that, theoretically, any complex sound can be decomposed into a sequence of simpler harmonic signals of different frequencies, each of which it is a sinusoid, called the spectrum of the original signal. The task of encoding sound, like any other analog signal, is to represent it in the form of another analog or digital signal, which is more convenient for its transmission or storage in each specific case. Real sound sources have a limited spectrum width, therefore, for encoding, transformation methods are used that transform the original signal into one, the spectrum of which is more suitable for transmission on the selected channel. Representing an analog signal as another analog signal is commonly referred to as modulation and digitally as encoding. This division is very arbitrary. An analog signal can be represented as a harmonic signal (that is, a sinusoid), the parameters of which change depending on the value of the original signal. In the event that the amplitude of the sinusoid changes with a change in the original signal, it is amplitude modulation (AM). If, depending on the value of the original signal, the frequency or phase of the sinusoid changes, we are dealing with frequency modulation (FM) or phase modulation (PM). Amplitude and frequency modulation, for example, is widely used to transmit sound by radio. These types of modulation, of course, are not the decomposition of the original signal into harmonics. The development of digital technology and the use of computer processing and information storage has led to the widespread use of pulse encoding or modulation methods. Such types of modulation are, for example, pulse code modulation, in which the value of the original signal at regular intervals is represented in code form. The vast majority of “computer sound” is precisely the recording of the binary code of the received signal in short equal time intervals, determined by the sampling frequency. For storage and transmission through communication channels, this signal is usually compressed (reducing the volume by discarding unnecessary or insignificant information). In addition to pulse code modulation, other types of digital modulation (pulse width, pulse frequency, etc.) are also used to encode sound.