 | This articleneeds additional citations forverification. Please helpimprove this article byadding citations to reliable sources. Unsourced material may be challenged and removed. Find sources: "Speech coding" – news ·newspapers ·books ·scholar ·JSTOR(January 2013) (Learn how and when to remove this message) |
Speech coding is an application ofdata compression todigital audio signals containingspeech. Speech coding uses speech-specificparameter estimation usingaudio signal processing techniques to model the speech signal, combined with generic data compression algorithms to represent the resulting modeled parameters in a compact bitstream.[1]
Common applications of speech coding aremobile telephony andvoice over IP (VoIP).[2] The most widely used speech coding technique in mobile telephony islinear predictive coding (LPC), while the most widely used in VoIP applications are the LPC andmodified discrete cosine transform (MDCT) techniques.[citation needed]
The techniques employed in speech coding are similar to those used inaudio data compression andaudio coding where appreciation ofpsychoacoustics is used to transmit only data that is relevant to the human auditory system. For example, invoiceband speech coding, only information in the frequency band 400 to 3500 Hz is transmitted but the reconstructed signal retains adequateintelligibility.
Speech coding differs from other forms of audio coding in that speech is a simpler signal than other audio signals, and statistical information is available about the properties of speech. As a result, some auditory information that is relevant in general audio coding can be unnecessary in the speech coding context. Speech coding stresses the preservation of intelligibility andpleasantness of speech while using a constrained amount of transmitted data.[3] In addition, most speech applications require low coding delay, aslatency interferes with speech interaction.[4]
Speech coders are of two classes:[5]
TheA-law andμ-law algorithms used inG.711 PCMdigital telephony can be seen as an earlier precursor of speech encoding, requiring only 8 bits per sample but giving effectively 12bits of resolution.[7] Logarithmic companding are consistent with human hearing perception in that a low-amplitude noise is heard along a low-amplitude speech signal but is masked by a high-amplitude one. Although this would generate unacceptable distortion in a music signal, the peaky nature of speech waveforms, combined with the simple frequency structure of speech as aperiodic waveform having a singlefundamental frequency with occasional added noise bursts, make these very simple instantaneous compression algorithms acceptable for speech.[citation needed][dubious –discuss]
A wide variety of other algorithms were tried at the time, mostlydelta modulation variants, but after careful consideration, the A-law/μ-law algorithms were chosen by the designers of the early digital telephony systems. At the time of their design, their 33% bandwidth reduction for a very low complexity made an excellent engineering compromise. Their audio performance remains acceptable, and there was no need to replace them in the stationary phone network.[citation needed]
In 2008,G.711.1 codec, which has a scalable structure, was standardized by ITU-T. The input sampling rate is 16 kHz.[8]
Much of the later work in speech compression was motivated by military research into digital communications forsecure military radios, where very low data rates were used to achieve effective operation in a hostile radio environment. At the same time, far moreprocessing power was available, in the form ofVLSI circuits, than was available for earlier compression techniques. As a result, modern speech compression algorithms could use far more complex techniques than were available in the 1960s to achieve far higher compression ratios.
The most widely used speech coding algorithms are based onlinear predictive coding (LPC).[9] In particular, the most common speech coding scheme is the LPC-basedcode-excited linear prediction (CELP) coding, which is used for example in theGSM standard. In CELP, the modeling is divided in two stages, alinear predictive stage that models the spectral envelope and a code-book-based model of the residual of the linear predictive model. In CELP, linear prediction coefficients (LPC) are computed and quantized, usually asline spectral pairs (LSPs). In addition to the actual speech coding of the signal, it is often necessary to usechannel coding for transmission, to avoid losses due to transmission errors. In order to get the best overall coding results, speech coding and channel coding methods are chosen in pairs, with the more important bits in the speech data stream protected by more robust channel coding.
Themodified discrete cosine transform (MDCT) is used in the LD-MDCT technique used by theAAC-LD format introduced in 1999.[10] MDCT has since been widely adopted invoice-over-IP (VoIP) applications, such as theG.729.1wideband audio codec introduced in 2006,[11]Apple'sFaceTime (using AAC-LD) introduced in 2010,[12] and theCELT codec introduced in 2011.[13]
Opus is afree software audio coder. It combines the speech-oriented LPC-basedSILK algorithm and the lower-latency MDCT-based CELT algorithm, switching between or combining them as needed for maximal efficiency.[14][15] It is widely used for VoIP calls inWhatsApp.[16][17][18] ThePlayStation 4 video game console also uses Opus for itsPlayStation Network system party chat.[19]
A number of codecs with even lowerbit rates have been demonstrated.Codec 2, which operates at bit rates as low as450 bit/s, sees use in amateur radio.[20] NATO currently usesMELPe, offering intelligible speech at600 bit/s and below.[21] Neural vocoder approaches have also emerged:Lyra by Google gives an "almost eerie" quality at3 kbit/s.[22] Microsoft'sSatin also uses machine learning, but uses a higher tunable bitrate and is wideband.[23]