블루투스 공식 홈페이지(한국어) : https://www.bluetooth.org/ko-kr/home

Bluetooth 공식 홈페이지(영어) : https://www.bluetooth.org/en-us

Bluetooth SIG는 전세계적으로 마케팅 엔지니어링 및 조직운영 전문 직원들을 보유하고 있습니다.

This profile defines how multimedia audio can be streamed from one device to another over a Bluetooth connection (so, it is also called Bluetooth Audio Streaming). For example, music can be streamed from a mobile phone, to a wireless headset, hearing aid/cochlear implant streamer, car audio, or from a laptop/desktop to a wireless headset; also, voice can be streamed from a microphone device to a recorder on a PC. The Audio/Video Remote Control Profile (AVRCP) is often used in conjunction with A2DP for remote control on devices such as headphones, car audio systems, or stand-alone speaker units. These systems often also implement Headset (HSP) or Hands-Free (HFP) profiles for telephone calls, which may be used separately.

Each A2DP service, of possibly many, is designed to uni-directionally transfer an audio stream in up to 2 channel stereo, either to or from the Bluetooth host.[2] This profile relies on AVDTP and GAVDP. It includes mandatory support for the low-complexity SBC codec (not to be confused with Bluetooth's voice-signal codecs such as CVSDM), and supports optionally MPEG-1 Part 3/MPEG-2 Part 3 (MP2 and MP3), MPEG-2 Part 7/MPEG-4 Part 3 (AAC and HE-AAC), and ATRAC, and is extensible to support manufacturer-defined codecs, such as aptX.[3]

Some Bluetooth stacks enforce the SCMS-T digital rights management (DRM) scheme. In these cases, it is impossible to connect certain A2DP headphones for high quality audio.

고급 리눅스 사운드 아키텍처(Advanced Linux Sound Architecture, ALSA)는 사운드 카드용 장치 드라이버를 제공하기 위한 리눅스 커널의 구성 요소이다. ALSA 프로젝트의 목표들 가운데 하나는 사운드 카드 하드웨어를 자동으로 구성하고 시스템에 여러 개의 사운드 장치를 멋지게 관리하는 것이다. JACK과 같은 두 개의 다른 프레임워크들은 ALSA를 사용하여 낮은 레이턴시로 고급 수준의 오디오 편집과 소리 조절을 수행할 수 있게 한다. Jaroslav Kysela가 이끄는 이 프로젝트는 1998년의 그레비스 울트라사운드 사운드 드라이버용 리눅스 장치 드라이버로 처음 시작하였으며 2002년에 2.5 개발 버전(2.5.4-2.5.5) 이후부터 리눅스 커널로부터 떨어져나온 뒤에 계속 개발되고 있다. 2.6 버전부터 이전 시스템인 오픈 사운드 시스템(OSS)을 기본으로 대체하고 있다.

원저자: Jaroslav Kysela

개발자: ALSA team

발표일: 1998

최근 버전: 1.1.1 / 2016년 03월 31일

프로그래밍 언어: C

운영 체제: 리눅스

종류: 오디오

라이선스: GPL 및 LGPL

웹사이트: www.alsa-project.org

공식 웹사이트 : http://www.alsa-project.org/main/index.php/Main_Page

http://www.alsa-project.org/alsa-doc/alsa-lib/pcm.html

PCM (digital audio) interface

Although abbreviation PCM stands for Pulse Code Modulation, we are understanding it as general digital audio processing with volume samples generated in continuous time periods.

The analog signal is recorded via analog to digital converters (ADC). The digital value (de-facto a volume at a specific time) obtained from ADC can be further processed. The following picture shows a perfect sinus waveform:

Next image shows digitized representation:

As you may see, the quality of digital audio signal depends on the time (recording rate) and voltage resolution (usually in an linear integer representation with basic unit one bit).

The stored digital signal can be converted back to voltage (analog) representation via digital to analog converters (DAC).

One digital value is called sample. More samples are collected to frames (frame is terminology for ALSA) depending on count of converters used at one specific time. One frame might contain one sample (when only one converter is used - mono) or more samples (for example: stereo has signals from two converters recorded at same time). Digital audio stream contains collection of frames recorded at boundaries of continuous time periods.

FramesPeriods

A frame is equivalent of one sample being played, irrespective of the number of channels or the number of bits. e.g.

• 1 frame of a Stereo 48khz 16bit PCM stream is 4 bytes.
• 1 frame of a 5.1 48khz 16bit PCM stream is 12 bytes.

A period is the number of frames in between each hardware interrupt. The poll() will return once a period.

The buffer is a ring buffer. The buffer size always has to be greater than one period size. Commonly this is 2*period size, but some hardware can do 8 periods per buffer. It is also possible for the buffer size to not be an integer multiple of the period size.

Now, if the hardware has been set to 48000Hz , 2 periods, of 1024 frames each, making a buffer size of 2048 frames. The hardware will interrupt 2 times per buffer. ALSA will endeavor to keep the buffer as full as possible. Once the first period of samples has been played, the third period of samples is transfered into the space the first one occupied while the second period of samples is being played. (normal ring buffer behaviour).

Additional example

Here is an alternative example for the above discussion.

Say we want to work with a stereo, 16-bit, 44.1 KHz stream, one-way (meaning, either in playback or in capture direction). Then we have:

• 'stereo' = number of channels: 2
• 1 analog sample is represented with 16 bits = 2 bytes
• 1 frame represents 1 analog sample from all channels; here we have 2 channels, and so:
• 1 frame = (num_channels) * (1 sample in bytes) = (2 channels) * (2 bytes (16 bits) per sample) = 4 bytes (32 bits)

• To sustain 2x 44.1 KHz analog rate - the system must be capable of data transfer rate, in Bytes/sec:
• Bps_rate = (num_channels) * (1 sample in bytes) * (analog_rate) = (1 frame) * (analog_rate) = ( 2 channels ) * (2 bytes/sample) * (44100 samples/sec) = 2*2*44100 = 176400 Bytes/sec

(출처 : https://www.alsa-project.org/main/index.php/FramesPeriods)

미국 5개사, 스마트 음성인식 시장 80% 이상 장악

빅데이터와 클라우드 컴퓨팅 등 산업이 발전함에 따라 스마트음성을 응용한 기술 수요도 함께 증가하고 있다.

11일 중국음성산업연맹에 따르면 오는 2017년 세계 스마트음성산업은 105억달러(약 12조3000억원)까지 커질 것이란 전망이다. 중국음성산업연맹에 따르면 지난해 세계 스마트음성시장에서 미국의 5개 기업이 점유율 80% 이상을 차지하고 있는 것으로 나타났다. 기업별 시장 점유율은 뉘앙스 커뮤니케이션이 31.6%로 1위를 차지했으며 구글이 28.4%로 뒤를 이었다. 

음성인식 시스템 시리를 개발해 일찌감치 시장에 뛰어든 애플은 15.4%, 마이크로소프트(MS)와 IBM은 각각 8.1%와 4.1% 점유율을 보였다. 삼성전자 S보이스와 LG전자 Q보이스 개발에 참여한 것으로 알려진 중국 기업 아이플라이테크(iFLYTEK)는 시장점유율 4.5%를 차지했다. 아이플라이테크는 미국 기업 외에 유일하게 순위권에 이름을 올렸다. 국내 기업은 큰 점유율을 차지하지 못하고 있다.

(기사 출처 및 전체 기사 보기 : http://www.dt.co.kr)

http://www.audacityteam.org/

The Hands-Free Profile (HFP) 1.5 specification   <=-  스펙문서를 받으시려면 여기를 클릭하세요.

The Hands-Free Profile (HFP) 1.6 specification   <=-  스펙문서를 받으시려면 여기를 클릭하세요.

Currently in version 1.6, this is commonly used to allow car hands-free kits to communicate with mobile phones in the car. It commonly uses Synchronous Connection Oriented link (SCO) to carry a monaural audio channel with continuously variable slope delta modulation or pulse-code modulation, and with logarithmic a-law or μ-law quantization. Version 1.6 adds optional support for wide band speech with the mSBC codec, a 16 kHz monaural configuration of the SBC codec mandated by the A2DP profile.

 (출처: https://en.wikipedia.org/wiki/List_of_Bluetooth_profiles)

The following roles are defined for this profile: 

Audio Gateway(AG) – This is the device that is the gateway of the audio, both for input and output. Typical devicesacting as Audio Gateways are cellular phones. 

Hands-Free unit (HF) – This is the device acting as the Audio Gateway’s remote audio input and output mechanism. It also providessome remote control means.  

4.33.3 Bluetooth Defined AT Capabilities

The new Bluetooth specific AT capabilities are listed below: 

AT+BVRA (Bluetooth Voice Recognition Activation) 

Syntax: AT+BVRA=<vrec> 

Description: 

Enables/disables the voice recognition function in the AG. 

Only support for execution command is mandated. Neither the read nor test commands are mandatory. 

Values: 

<vrec>: 0, 1, entered asinteger values, where 

0 = Disable Voice recognition in the AG, 1 = Enable Voice recognition in the AG 

+BVRA (Bluetooth Voice Recognition Activation)

Syntax: +BVRA: <vrect> 

Description: 

Unsolicited result code used to notify the HF when the voice recognition function in the AG is activated/deactivated autonomously from the AG. 

The unsolicited +BVRA:1 result code shall not be sent by the AG to the HF if the corresponding voice recognition activation has been initiated by the HF. Likewise, the unsolicited +BVRA:0 result code shall not be sent by the 

AG to the HF if the corresponding voice recognition deactivation has been initiated by the HF, regardless of which side initiated the voice recognition activation. 

Values: 

<vrect>: 0, entered as integer value, where 

0 = Voice recognition is disabled in the AG, 1 = Voice recognition is enabled in the AG 

4.25 Voice Recognition Activation 

The HF, via the AT+BVRA command, or the AG autonomously, may activate/deactivate the voice recognition function resident in the AG. Beyond the audio routing and voice recognition activation capabilities, the rest of the voice recognition functionality is implementation dependent. 

Whenever the AG supports a voice recognition function it shall support the AT+BVRA command as described in the procedures in this section. 

If the HF issues the AT+BVRA command, the AG shall respond with the OK result code if it supports voice recognition, then initiate an Audio Connection to the HF (if the Audio Connection does not already exist) and begin the voice input sequence. 

If the AG does not support voice recognition, the AG shall respond with the ERROR indication. 

When the voice recognition function is activated from the AG, it shall inform the HF via the +BVRA: 1 unsolicited result code and the AG shall initiate an Audio Connection to the HF (if the Audio Connection does not already exist) and begin the voice input sequence.

Once activated, depending upon the voice recognition implementation, the AG shall then keep the voice recognition function enabled: 

•  For the duration of time supported by the implementation (“momentary on” voice recognition implementation). In this case, the AG shall notify the HF by sending a +BVRA: 0 unsolicited result code. 

•  Or until the AT+BVRA command is issued to disable voice recognition from the HF. 

•  Or until the current Service Level Connection between the AG and the HF is dropped for any reason.

Sensitivity: -32dBVpk

SNR: 69dB (1kHz, 1Pa, A-Weighted)

Max Input SPL: 110dB SPL

Sensitivity indicates how well the microphone converts acoustic pressure to output voltage. A high sensitivity microphone creates more voltage and so needs less amplification at the mixer or recording device. This is a practical concern but is not directly an indication of the microphone's quality, and in fact the term sensitivity is something of a misnomer, "transduction gain" being perhaps more meaningful, (or just "output level") because true sensitivity is generally set by the noise floor, and too much "sensitivity" in terms of output level compromises the clipping level. There are two common measures. The (preferred) international standard is made in millivolts per pascal at 1 kHz. A higher value indicates greater sensitivity. The older American method is referred to a 1 V/Pa standard and measured in plain decibels, resulting in a negative value. Again, a higher value indicates greater sensitivity, so −60  dB is more sensitive than −70 dB.

(출처: en.wikipedia.org)

EIN = Max Input( = acoustic overload point) − dynamic range

EIN = 94 dB − SNR

dynamic range = Max Input - EIN

오푸스(Opus)는 국제 인터넷 표준화 기구(IETF)가 개발한 손실 오디오 압축 포맷이며, 특히 인터넷 상의 인터렉티브 실시간 응용 프로그램에 적합하게 만들어졌다. 

RFC 6716으로 표준화된 오픈 포맷으로서, 참조 구현이 3절 BSD 라이선스 하에 제공된다. 오푸스를 포함하는 모든 알려진 소프트웨어 특허는 무로열티 조건 하에 허가되어 있다. 

오푸스는 음성지향 SILK와 저지연 CELT 두가지 코덱의 기술을 포함하고있다. 오푸스는 높은 비트 레이트와 낮은 비트 레이트에 모두 매끄럽게 적용될 수 있다. 내부적으로 낮은 비트레이트의 선형 예측 코덱과 높은 비트레이트의 변형 코덱의 사이를 전환한다. 

오푸스는, 자연스러운 대화와 네트워크 음악 공연 혹은 립싱크를 가능하게 하는 낮은 오디오 지연시간 의사소통 연결의 부분으로서의 사용을 위해 필수적인, 아주 낮은 알고리즘 지연시간(기본 22.5 ms)을 갖고 있다. 오푸스는 5 ms 이하의 더 작은 알고리즘 지연시간을 달성하기 위해 품질 및 비트 레이트의 균형을 맞추는 것을 허용한다.현재까지는 5ms까지로 지연시간을 줄일 수 있다.

(출처: ko.wikipedia.org)

테크니컬 컬럼 #2 - 24 Bit / 192 KHz 는 반드시 음질이 좋을까요?

An orthography is a set of conventions for how to write a language. It includes rules of spelling, hyphenation, capitalization, word breaks, emphasis, and punctuation.

[출처: en.wikipedia.org] 

A phoneme /ˈfoʊniːm/ is a basic unit of a language's phonology, which is combined with other phonemes to form meaningful units, morphemes. The phoneme can be described as "The smallest contrastive linguistic unit which may bring about a change of meaning". In this way the difference in meaning between the English words kill and kiss is a result of the exchange of the phoneme /l/ for the phoneme /s/. Two words that differ in meaning through a contrast of a single phoneme form a minimal pair.

[출처: en.wikipedia.org]

A grapheme is the smallest unit used in describing the writing system of a language, originally coined by analogy with the phoneme of spoken languages. A grapheme may or may not carry meaning by itself, and may or may not correspond to a single phoneme. Graphemes include alphabetic letters, typographic ligatures, Chinese characters, numerical digits,punctuation marks, and other individual symbols of any of the world's writing systems.

[출처: en.wikipedia.org]

WAV 또는 WAVE는 웨이브폼 오디오 포맷(웨이브 오디오 포맷, Waveform audio format)의 준말로 개인용 컴퓨터에서 오디오를 재생하는 마이크로소프트와 IBM오디오 파일 포맷 표준이다. 덩어리째로 데이터를 저장하기 위한 RIFF 비트스트림 포맷 방식에서 변화한 것으로, 아미가에 쓰이는 IFFF와 매킨토시 컴퓨터에 쓰이는 AIFF 포맷에 매우 가깝다. 가공되지 않은 오디오를 위한 윈도 시스템에 쓰이는 기본 포맷이다. 윈도 상에서는 WAV가 비압축 형식만 있는 것이 아니고 여러 압축 코덱을 이용할 수 있다. 그러나 윈도 비스타부터는 기본 지원 포맷이 WMA으로 바뀌었다.

출처: https://ko.wikipedia.org/wiki/WAV

파스칼(기호 Pa)는 압력에 대한 SI 유도 단위이다. 1 파스칼은 1 제곱미터 당 1 뉴턴의 힘이 작용할 때의 압력에 해당한다. 단위의 이름은 프랑스의 수학자 블레즈 파스칼의 이름을 땄다.

정의[편집]

여기서 N 은 뉴턴, m 은 미터, kg 은 킬로그램, s 는 초.

dBSPL(sound pressure level, 음압) : 음압(소리의 매질인 공기의 압력의 실효출력)의 단위는 파스칼(Pa)이지만, 2×10⁻⁵ Pa를 기준값(0 dBSPL)으로 한다.

The Lombard effect or Lombard reflex is the involuntary tendency of speakers to increase their vocal effort when speaking in loud noise to enhance the audibility of their voice. This change includes not only loudness but also other acoustic features such as pitch, rate, and duration of sound syllables. This compensation effect results in an increase in the auditory signal-to-noise ratio of the speaker’s spoken words. 

The effect links to the needs of effective communication as there is a reduced effect when words are repeated or lists are read where communication intelligibility is not important. Since the effect is also involuntary it is used as a means to detect malingering in those simulating hearing loss. Research upon birds and monkeys finds that the effect also occurs in the vocalizations of nonhuman animals. 

(출처: http://en.wikipedia.org/wiki/Lombard_effect)