Speech recognition

This article has multiple issues. Please help improve it or discuss these issues on the talk page. (Learn how and when to remove these messages) This article needs additional citations for verification. Please help improve this article by adding citations to reliable sources. Unsourced material may be challenged and removed. Find sources: "Speech recognition" – news · newspapers · books · scholar · JSTOR (July 2025) (Learn how and when to remove this message) This article may require copy editing for grammar, style, cohesion, tone, or spelling. You can assist by editing it. (July 2025) (Learn how and when to remove this message) This article may be too technical for most readers to understand. Please help improve it to make it understandable to non-experts, without removing the technical details. (July 2025) (Learn how and when to remove this message) (Learn how and when to remove this message)

Speech recognition is an interdisciplinary sub-field of computer science and computational linguistics focused on developing computer-based methods and technologies for translating spoken language into text. It is also known as automatic speech recognition (ASR), computer speech recognition, or speech-to-text (STT).^[1]

Speech recognition applications include voice user interfaces such as voice commands for dialing, call routing, home automation, and aircraft control (usually called direct voice input). There are also productivity applications for speech recognition such as searching audio recordings and creating transcripts. Similarly, speech-to-text processing can allow users to write via dictation for word processors, emails, or data entry.

Speech recognition can be used in determining speaker characteristics.^[2] Automatic pronunciation assessment is used in education, such as for spoken language learning.

The term voice recognition^[3][4][5] or speaker identification^[6][7][8] refers to identifying the speaker, rather than what they are saying. Recognizing the speaker can simplify the task of translating speech in systems trained on a specific person's voice, or it can be used to authenticate or verify the speaker's identity as part of a security process.

The technology behind speech recognition has a long and innovative history, benefiting significantly from recent advances in deep learning and the availability of big data. These advances are complemented by an increase in academic papers on the topic,^[9] and greater adoption in the design and deployment of speech recognition systems.^[10]

Key areas of growth have included larger vocabulary size, more accurate recognition for speakers that models were not trained on directly (called speaker independence), and faster processing speed.

• 1952 – Three Bell Labs researchers, Stephen Balashek,^[11] R. Biddulph, and K. H. Davis, built a system called "Audrey"^[12] for single-speaker digit recognition. Their system located the formants in the power spectrum of each utterance.^[13]
• 1960 – Gunnar Fant developed and published the source-filter model of speech production.
• 1962 – IBM demonstrated its 16-word "Shoebox" machine's speech recognition capability at the 1962 World's Fair.^[14]
• 1966 – Linear predictive coding (LPC), a speech coding method, was first proposed by Fumitada Itakura of Nagoya University and Shuzo Saito of Nippon Telegraph and Telephone (NTT), while working on speech recognition.^[15]
• 1969 – Funding at Bell Labs came to a halt for several years after the company's head engineer, John R. Pierce, wrote an open letter criticizing speech recognition research.^[16] This defunding lasted until Pierce retired and James L. Flanagan took over.

Raj Reddy was the first person to work on continuous speech recognition as a graduate student at Stanford University in the late 1960s. Previous systems required users to pause after each word, while Reddy's system issued spoken commands for playing chess.

Around this time, Soviet researchers invented the dynamic time warping (DTW) algorithm and used it to create a recognizer capable of operating on a 200-word vocabulary.^[17] DTW processed speech by dividing it into short frames (e.g., 10ms segments) and treating each frame as a single unit. Although DTW would be superseded by later algorithms, the technique carried on. Achieving speaker independence, however, remained unsolved during this time.

• 1971 – DARPA funded a five-year speech recognition research project, Speech Understanding Research, seeking a minimum vocabulary size of 1,000 words. The project considered speech understanding a key to achieving progress in speech recognition, which was later proved untrue.^[18] BBN, IBM, Carnegie Mellon, and Stanford Research Institute all participated in the program.^[19][20] This project revived speech recognition research following John Pierce's letter.
• 1972 – The IEEE Acoustics, Speech, and Signal Processing group held a conference in Newton, Massachusetts.
• 1976 – The first ICASSP was held in Philadelphia, which has since become a major venue for publishing on speech recognition.^[21]

During the late 1960s, Leonard Baum developed the mathematics of Markov chains at the Institute for Defense Analysis. A decade later, at CMU, Raj Reddy's students James Baker and Janet M. Baker began using the hidden Markov model (HMM) for speech recognition.^[22] James Baker had learned about HMMs while working a summer job at the Institute for Defense Analysis during his undergraduate education.^[23] The use of HMMs enabled researchers to combine different sources of knowledge, such as acoustics, language, and syntax, in a unified probabilistic model.

• By the mid-1980s, IBM's Fred Jelinek's team created a voice-activated typewriter called Tangora, which could handle a 20,000-word vocabulary.^[24] Jelinek's statistical approach placed less emphasis on emulating the way the human brain processes and understands speech in favor of using statistical modelling techniques like HMMs. (Jelinek's group independently discovered the application of HMMs to speech.^[23]) This was controversial among linguists since HMMs are too simplistic to account for many common features of human languages.^[25] However, the HMM proved to be a highly useful way for modelling speech and replaced dynamic time warping to become the dominant speech recognition algorithm in the 1980s.^[26][27]
• 1982 – Dragon Systems, founded by James and Janet M. Baker,^[28] was one of IBM's few competitors.

The 1980s also saw the introduction of the n-gram language model.

• 1987 – The back-off model enabled language models to use multiple-length n-grams, and CSELT^[29] used HMM to recognize languages (both in software and hardware specialized processors, e.g., RIPAC).

Much of the progress in the field is owed to the rapidly increasing capabilities of computers. At the end of the DARPA program in 1976, the best computer available to researchers was the PDP-10 with 4 MB of RAM.^[30] It could take up to 100 minutes to decode just 30 seconds of speech.^[31]

Two practical products were:

• 1984 – the Apricot Portable was released with up to 4096 words support, of which only 64 could be held in RAM at a time.^[32]
• 1987 – a recognizer from Kurzweil Applied Intelligence
• 1990 – Dragon Dictate, a consumer product released in 1990.^[33][34] AT&T deployed the Voice Recognition Call Processing service in 1992 to route telephone calls without the use of a human operator.^[35] The technology was developed by Lawrence Rabiner and others at Bell Labs.

By this point, the vocabulary of the typical commercial speech recognition system was larger than the average human vocabulary.^[30] Raj Reddy's former student, Xuedong Huang, developed the Sphinx-II system at CMU. The Sphinx-II system was the first to do speaker-independent, large vocabulary, continuous speech recognition, and it had the best performance in DARPA's 1992 evaluation. Handling continuous speech with a large vocabulary was a major milestone in the history of speech recognition. Huang went on to found the speech recognition group at Microsoft in 1993. Raj Reddy's student Kai-Fu Lee joined Apple, where, in 1992, he helped develop a speech interface prototype for the Apple computer, known as Casper.

Lernout & Hauspie, a Belgium-based speech recognition company, acquired several other companies, including Kurzweil Applied Intelligence in 1997 and Dragon Systems in 2000. The L&H speech technology was used in the Windows XP operating system. L&H was an industry leader until an accounting scandal brought an end to the company in 2001. The speech technology from L&H was bought by ScanSoft, which became Nuance in 2005. Apple originally licensed software from Nuance to provide speech recognition capability to its digital assistant Siri.^[36]

In the 2000s, DARPA sponsored two speech recognition programs: Effective Affordable Reusable Speech-to-Text (EARS) in 2002, followed by Global Autonomous Language Exploitation (GALE) in 2005. Four teams participated in the EARS program: IBM; a team led by BBN with LIMSI and the University of Pittsburgh; Cambridge University; and a team composed of ICSI, SRI, and the University of Washington. EARS funded the collection of the Switchboard telephone speech corpus, which contains 260 hours of recorded conversations from over 500 speakers.^[37] The GALE program focused on Arabic and Mandarin broadcast news speech. Google's first effort at speech recognition came in 2007 after recruiting researchers from Nuance.^[38] Its first product, GOOG-411, was a telephone-based directory service whose recordings helped improve Google's recognition systems. Google Voice Search now supports more than 30 languages.

Since at least 2006, the U.S. National Security Agency has made use of a type of speech recognition for keyword spotting, allowing analysts to index large volumes of recorded conversations and isolate mentions of keywords.^[39] Other government research programs focused on intelligence applications of speech recognition, such as DARPA's EARS program and IARPA's Babel program.

In the early 2000s, speech recognition was still dominated by traditional approaches such as hidden Markov models combined with feed-forward artificial neural networks.^[40] More recently, many aspects of speech recognition have been taken over by a deep learning method called Long short-term memory (LSTM), a recurrent neural network published by Sepp Hochreiter & Jürgen Schmidhuber in 1997.^[41] LSTM RNNs avoid the vanishing gradient problem and can learn "Very Deep Learning" tasks^[42] that require memories of events that happened thousands of discrete time steps ago, which is important for speech.

Around 2007, LSTM trained with Connectionist Temporal Classification (CTC)^[43] began to outperform traditional approaches.^[44] In 2015, Google reported a 49 percent error‑rate reduction in its speech recognition via CTC‑trained LSTM, now available through Google Voice to all smartphone users.^[45] Transformers, a type of neural network based solely on attention, have been widely adopted in computer vision^[46][47] and language modelling,^[48][49] sparking the interest of adapting such models to new domains, including speech recognition.^[50][51][52] Some recent papers reported superior performance levels using transformer models for speech recognition, but these models usually require large scale training datasets to reach high performance levels.

The use of deep feed-forward (non-recurrent) networks for acoustic modelling was introduced during the later part of 2009 by Geoffrey Hinton and his students at the University of Toronto, and by Li Deng^[53] and colleagues at Microsoft Research, initially in the collaborative work between Microsoft and the University of Toronto, which was subsequently expanded to include IBM and Google (hence "The shared views of four research groups" subtitle in their 2012 review paper).^[54][55][56] A Microsoft research executive described this innovation as "the most dramatic change in accuracy since 1979".^[57] In contrast to the steady incremental improvements of the past few decades, the application of deep learning decreased word error rate by 30%.^[57] This innovation was quickly adopted across the field. Researchers have now begun to use deep learning techniques for language modelling as well, leading to the development of models such as OpenAI's ChatGPT, Anthropic's Claude, and Google's Gemini.

In the long history of speech recognition, both shallow forms and deep forms (e.g., recurrent nets) of artificial neural networks had been explored for many years during the 1980s, 1990s, and the 2000s.^[58][59][60] However, these methods never won over the non-uniform internal-handcrafting Gaussian mixture model/hidden Markov model (GMM-HMM) technology based on generative models of speech trained discriminatively.^[61] Several key difficulties had been methodologically analyzed in the 1990s, including gradient diminishing^[62] and weak temporal correlation structure in the neural predictive models.^[63][64] All these difficulties were in addition to the lack of big training data and big computing power in their early days. Most speech recognition researchers who understood these barriers subsequently moved away from neural nets to pursue generative modelling approaches until the recent resurgence of deep learning starting around 2009–2010, which had overcome all these difficulties. Hinton et al. and Deng et al. reviewed part of this recent history about how their collaboration with each other and then with colleagues across four groups (University of Toronto, Microsoft, Google, and IBM) ignited a renaissance of applications of deep feed-forward neural networks for speech recognition.^{[55][56][65][66]}

By early the 2010s, speech recognition, also called voice recognition,^[67][68][69] was clearly differentiated from speaker recognition, and speaker independence was considered a major breakthrough. Until then, systems required a "training" period. A 1987 ad for a doll had carried the tagline "Finally, the doll that understands you," despite the fact that it was described as "which children could train to respond to their voice."^[14]

In 2017, Microsoft researchers reached a historical human parity milestone of transcribing conversational telephony speech on the widely benchmarked Switchboard task. Multiple deep learning models were used to optimize speech recognition accuracy. The speech recognition word error rate was reported to be as low as 4 professional human transcribers working together on the same benchmark, which was funded by IBM Watson speech team on the same task.^[70]

Both acoustic modelling and language modelling are important parts of modern statistically-based speech recognition algorithms. Hidden Markov models (HMMs) are widely used in many systems. Language modelling is also used in many other natural language processing applications, such as document classification or statistical machine translation.

Modern general-purpose speech recognition systems are based on hidden Markov models. These are statistical models that output a sequence of symbols or quantities. HMMs are used in speech recognition because a speech signal can be viewed as a piecewise stationary signal or a short-time stationary signal. In a short time scale (e.g., 10 milliseconds), speech can be approximated as a stationary process. Speech can be thought of as a Markov model for many stochastic purposes.

Another reason why HMMs are popular is that they can be trained automatically and are simple and computationally feasible to use. In speech recognition, the hidden Markov model would output a sequence of n-dimensional real-valued vectors (with n being a small integer, such as 10), outputting one of these every 10 milliseconds. The vectors would consist of cepstral coefficients, which are obtained by taking a Fourier transform of a short time window of speech and decorrelating the spectrum using a cosine transform, then taking the first (most significant) coefficients. The hidden Markov model will tend to have, in each state, a statistical distribution that is a mixture of diagonal covariance Gaussians, which will give a likelihood for each observed vector. Each word, or (for more general speech recognition systems), each phoneme, will have a different output distribution; a hidden Markov model for a sequence of words or phonemes is made by concatenating the individual trained hidden Markov models for the separate words and phonemes.

Described above are the core elements of the most common, HMM-based approach to speech recognition. Modern speech recognition systems use various combinations of a number of standard techniques in order to improve results over the basic approach described above. A typical large-vocabulary system would need context dependency for the phonemes (so that phonemes with different left and right context would have different realizations as HMM states). It would use cepstral normalization to normalize for a different speaker and recording conditions. For further speaker normalization, it might use vocal tract length normalization (VTLN) for male-female normalization and maximum likelihood linear regression (MLLR) for more general speaker adaptation. The features would have so-called delta and delta-delta coefficients to capture speech dynamics, and in addition might use heteroscedastic linear discriminant analysis (HLDA); or might skip the delta and delta-delta coefficients and use splicing and an LDA-based projection, followed perhaps by HLDA or a global semi-tied covariance transform (also known as maximum likelihood linear transform, or MLLT). Many systems use so-called discriminative training techniques that dispense with a purely statistical approach to HMM parameter estimation and instead optimize some classification-related measure of the training data. Examples are maximum mutual information (MMI), minimum classification error (MCE), and minimum phone error (MPE).

Decoding of the speech (the term for what happens when the system is presented with a new utterance and must compute the most likely source sentence) would probably use the Viterbi algorithm to find the best path, and here there is a choice between dynamically creating a combination hidden Markov model, which includes both the acoustic and language model information, and combining it statically beforehand (the finite state transducer, or FST, approach).

A possible improvement to decoding is to keep a set of good candidates instead of only keeping the best candidate, and to use a better scoring function (re-scoring) to rate these good candidates so that we may pick the best one according to this refined score. The set of candidates can be kept either as a list (the N-best list approach) or as a subset of the models (a lattice). Re-scoring is usually done by trying to minimize the Bayes risk^[71] (or an approximation thereof). Instead of taking the source sentence with maximal probability, we try to take the sentence that minimizes the expectancy of a given loss function with regards to all possible transcriptions (i.e., we take the sentence that minimizes the average distance to other possible sentences weighted by their estimated probability). The loss function is usually the Levenshtein distance, though it can be different distances for specific tasks; the set of possible transcriptions is, of course, pruned to maintain tractability. Efficient algorithms have been devised to re-score lattices represented as weighted finite-state transducers with edit distances represented as a finite-state transducer, verifying certain assumptions.^[72]

Dynamic time warping is an approach that was historically used for speech recognition, but has now largely been displaced by the more successful HMM-based approach.

Dynamic time warping is an algorithm for measuring similarity between two sequences that may vary in time or speed. For instance, similarities in walking patterns would be detected, even if in one video the person was walking slowly and in another they were walking more quickly, or even if there were accelerations and decelerations during the course of one observation. DTW has been applied to video, audio, and graphics – indeed, any data that can be turned into a linear representation can be analyzed with DTW.

A well-known application has been automatic speech recognition, to cope with different speaking speeds. In general, it is a method that allows a computer to find an optimal match between two given sequences (e.g., time series) with certain restrictions. That is, the sequences are "warped" non-linearly to match each other. This sequence alignment method is often used in the context of hidden Markov models.

Neural networks emerged as an attractive acoustic modelling approach in ASR in the late 1980s. Since then, neural networks have been used in many aspects of speech recognition, such as phoneme classification,^[73] phoneme classification through multi-objective evolutionary algorithms,^[74] isolated word recognition,^[75] audiovisual speech recognition, audiovisual speaker recognition, and speaker adaptation.

Neural networks make fewer explicit assumptions about feature statistical properties than HMMs and have several qualities that make them more attractive models for speech recognition. When used to estimate the probabilities of a speech feature segment, neural networks allow discriminative training in a natural and efficient manner. However, in spite of their effectiveness in classifying short-time units such as individual phonemes and isolated words,^[76] early neural networks were rarely successful for continuous recognition tasks because of their limited ability to model temporal dependencies.

One approach to this limitation was to use neural networks as a pre-processing, feature transformation, or dimensionality reduction^[77] step prior to HMM-based recognition. However, more recently, LSTM and related recurrent neural networks (RNNs),^{[41][45][78][79]} Time Delay Neural Networks(TDNN's),^[80] and transformers^[50][51][52] have demonstrated improved performance in this area.

Researchers are also exploring deep neural networks (DNNs) and denoising autoencoders^[81] .A DNN is a type of artificial neural network that includes multiple hidden layers between the input and output. ^[55] Like simpler neural networks, DNNs can model complex, non-linear relationships. However, their deeper architecture allows them to build more sophisticated representations by combining features from earlier layers. This gives them a powerful ability to learn and recognize complex patterns in speech data.^[82]

A major breakthrough in using DNNs for large vocabulary speech recognition came in 2010. In a collaboration between industry and academia, researchers used DNNs with large output layers based on context-dependent Hidden Markov Model (HMM) states, which were created using decision trees.^{[83][84] [85]} This approach significantly improved performance. For a detailed overview of this development and the state of the field as of 2014, you can refer to a Springer book published by Microsoft Research.^[86] Additional background on automatic speech recognition and the role of deep learning can be found in recent review articles.^[87][88]

One of the core ideas behind deep learning is to eliminate the need for manually designed features and instead learn directly from raw data. This was first demonstrated using deep autoencoders trained on raw spectrograms or linear filter-bank features.^[89] These models outperformed traditional Mel-Cepstral features, which rely on fixed transformations. More recently, researchers have shown that using raw audio waveforms as input can lead to excellent results in large-scale speech recognition tasks.^[90]

Since 2014, there has been much research interest in "end-to-end" ASR. Traditional phonetic-based (i.e., all HMM-based model) approaches required separate components and training for the pronunciation, acoustic, and language model. End-to-end models jointly learn all the components of the speech recognizer. This is valuable because it simplifies the training and deployment processes. For example, a n-gram language model is required for all HMM-based systems, and a typical n-gram language model often takes several gigabytes in memory making them impractical to deploy on mobile devices.^[91] Consequently, modern commercial ASR systems from Google and Apple (as of 2017^[update]) are deployed on the cloud and require a network connection as opposed to the device locally.

The first attempt at end-to-end ASR was with Connectionist Temporal Classification (CTC)-based systems introduced by Alex Graves of Google DeepMind and Navdeep Jaitly of the University of Toronto in 2014.^[92] The model consisted of recurrent neural networks and a CTC layer. Jointly, the RNN-CTC model learns the pronunciation and acoustic model together, however, it is incapable of learning the language model due to conditional independence assumptions, similar to a HMM. Consequently, CTC models can directly learn to map speech acoustics to English characters, but the models make many common spelling mistakes and must rely on a separate language model to clean up the transcripts. Later, Baidu expanded on the work with extremely large datasets and demonstrated some commercial success in Chinese Mandarin and English.^[93] In 2016, the University of Oxford presented LipNet,^[94] the first end-to-end sentence-level lipreading model, using spatiotemporal convolutions coupled with an RNN-CTC architecture, surpassing human-level performance in a restricted grammar dataset.^[95] A large-scale CNN-RNN-CTC architecture was presented in 2018 by Google DeepMind, achieving 6 times better performance than human experts.^[96] In 2019, Nvidia launched two CNN-CTC ASR models, Jasper and QuarzNet, with an overall performance WER of 3%.^[97][98] Similar to other deep learning applications, transfer learning and domain adaptation are important strategies for reusing and extending the capabilities of deep learning models, particularly due to the high costs of training models from scratch, and the small size of available corpus in many languages and/or specific domains.^{[99][100][101]}

An alternative approach to CTC-based models are attention-based models. Attention-based ASR models were introduced simultaneously by Chan et al. of Carnegie Mellon University and Google Brain, and Bahdanau et al. of the University of Montreal in 2016.^[102][103] The model named "Listen, Attend and Spell" (LAS), literally "listens" to the acoustic signal, pays "attention" to different parts of the signal and "spells" out the transcript one character at a time. Unlike CTC-based models, attention-based models do not have conditional-independence assumptions and can learn all the components of a speech recognizer including directly. This means that during deployment, there is no need to carry around a language model, making it very practical for applications with limited memory.

By the end of 2016, the attention-based models had seen considerable success, including outperforming the CTC models (with or without an external language model).^[104] Various extensions have been proposed since the original LAS model. Latent Sequence Decompositions (LSD) was proposed by Carnegie Mellon University, MIT and Google Brain to directly emit sub-word units which are more natural than English characters;^[105] the University of Oxford and Google DeepMind extended LAS to "Watch, Listen, Attend and Spell" (WLAS) to handle lip reading surpassing human-level performance.^[106]

Typically, a manual control input, e.g. a finger control on the steering-wheel, enables the speech recognition system and is signaled to the driver by an audio prompt. Following the audio prompt, the system has a "listening window" during which it may accept a speech input for recognition. ^{[citation needed]}

Simple voice commands may be used to initiate phone calls, select radio stations, or play music from a compatible smartphone, MP3 player, or music-loaded flash drive. Voice recognition capabilities vary between car make and model. Some recent car models offer natural-language speech recognition in place of a fixed set of commands, allowing the driver to use full sentences and common phrases. With such systems, there is no need for the user to memorize a set of fixed commands.^{[citation needed]}

Automatic pronunciation assessment is the use of speech recognition to verify the correctness of pronounced speech,^[107] as distinguished from manual assessment by an instructor or proctor.^[108] Also called speech verification, pronunciation evaluation, and pronunciation scoring, the main application of this technology is computer-aided pronunciation teaching (CAPT) when combined with computer-aided instruction for computer-assisted language learning (CALL), speech remediation, or accent reduction. Pronunciation assessment does not determine unknown speech (as in dictation or automatic transcription) but instead, knowing the expected word(s) in advance, it attempts to verify the correctness of the learner's pronunciation and ideally their intelligibility to listeners,^[109][110] sometimes along with often inconsequential prosody such as intonation, pitch, tempo, rhythm, and stress.^[111] Pronunciation assessment is also used in reading tutoring, for example in products such as Microsoft Teams^[112] and from Amira Learning.^[113] Automatic pronunciation assessment can also be used to help diagnose and treat speech disorders such as apraxia.^[114]

Assessing authentic listener intelligibility is essential for avoiding inaccuracies from accent bias, especially in high-stakes assessments,^{[115][116][117]} from words with multiple correct pronunciations,^[118] and from phoneme coding errors in machine-readable pronunciation dictionaries.^[119] In 2022, researchers found that some newer speech to text systems, based on end-to-end reinforcement learning to map audio signals directly into words, produce word and phrase confidence scores very closely correlated with genuine listener intelligibility.^[120] In the Common European Framework of Reference for Languages (CEFR) assessment criteria for "overall phonological control", intelligibility outweighs formally correct pronunciation at all levels.^[121]

In the health care sector, speech recognition can be implemented in front-end or back-end of the medical documentation process. Front-end speech recognition is where the provider dictates into a speech-recognition engine, the recognized words are displayed as they are spoken, and the dictator is responsible for editing and signing off on the document. Back-end or deferred speech recognition is where the provider dictates into a digital dictation system, the voice is routed through a speech-recognition machine, and the recognized draft document is routed along with the original voice file to the editor, where the draft is edited and report finalized. Deferred speech recognition is widely used in the industry.

One of the major issues relating to the use of speech recognition in healthcare is that the American Recovery and Reinvestment Act of 2009 (ARRA) provides for substantial financial benefits to physicians who utilize an Electronic Health Record (EHR) according to "Meaningful Use" standards. These standards require that a substantial amount of data be maintained by the EHR. The use of speech recognition is more naturally suited to the generation of narrative text, as part of a radiology/pathology interpretation, progress note or discharge summary; the ergonomic gains of using speech recognition to enter structured discrete data (e.g., numeric values or codes from a list or a controlled vocabulary) are relatively minimal for people who are sighted and who can operate a keyboard and mouse.

A more significant issue is that most EHRs have not been expressly tailored to take advantage of voice-recognition capabilities. A large part of a clinician's interaction with EHR involves navigation through the user interface using menus and tab/button clicks, and is heavily dependent on keyboard and mouse; voice-based navigation provides only modest ergonomic benefits. By contrast, many highly customized systems for radiology or pathology dictation implement voice "macros", where the use of certain phrases – e.g., "normal report", will automatically fill in a large number of default values and/or generate boilerplate, which will vary with the type of the exam – e.g., a chest X-ray vs. a gastrointestinal contrast series for a radiology system.

Prolonged use of speech recognition software in conjunction with word processors has shown benefits to short-term-memory restrengthening in brain AVM patients who have been treated with resection. Further research needs to be conducted to determine cognitive benefits for individuals whose AVMs have been treated using radiologic techniques.^{[citation needed]}

Substantial efforts have been devoted to the test and evaluation of speech recognition in fighter aircraft. Of particular note have been the US programme in speech recognition for the Advanced Fighter Technology Integration (AFTI)/F-16 aircraft (F-16 VISTA), the programme in France for Mirage aircraft, and other programmes in the UK dealing with a variety of aircraft platforms. In these programmes, speech recognizers have been operated successfully in fighter aircraft, with applications including setting radio frequencies, commanding an autopilot system, setting steer-point coordinates and weapons release parameters, and controlling flight display.

Working with Swedish pilots flying in the JAS-39 Gripen cockpit, Englund (2004) found recognition deteriorated with increasing g-loads. The report also concluded that adaptation greatly improved the results in all cases and that the introduction of models for breathing was shown to improve recognition scores significantly. Contrary to what might have been expected, no effects of the broken English of the speakers were found. It was evident that spontaneous speech caused problems for the recognizer, as might have been expected. A restricted vocabulary, and above all, a proper syntax, could thus be expected to improve recognition accuracy substantially.^[122]

The Eurofighter Typhoon, currently in service with the UK RAF, employs a speaker-dependent system, requiring each pilot to create a template. The system is not used for any safety-critical or weapon-critical tasks, such as weapon release or lowering of the undercarriage, but is used for a wide range of other cockpit functions. Voice commands are confirmed by visual and/or aural feedback. The system is seen as a major design feature in the reduction of pilot workload,^[123] and even allows the pilot to assign targets to his aircraft with two simple voice commands or to any of his wingmen with only five commands.^[124]

Speaker-independent systems are also being developed and are under test for the F-35 Lightning II (JSF) and the Alenia Aermacchi M-346 Master lead-in fighter trainer. These systems have produced word accuracy scores in excess of 98%.^[125]

The problems of achieving high recognition accuracy under stress and noise are particularly relevant in the helicopter environment as well as in the jet fighter environment. The acoustic noise problem is actually more severe in the helicopter environment, not only because of the high noise levels, but also because the helicopter pilot, in general, does not wear a facemask, which would reduce acoustic noise in the microphone. Substantial test and evaluation programmes have been carried out with speech recognition systems applications in helicopters, notably by the U.S. Army Avionics Research and Development Activity (AVRADA) and by the Royal Aerospace Establishment (RAE) in the UK. Work in France has included speech recognition in the Puma helicopter. There has also been much useful work in Canada. Results have been encouraging, and voice applications have included control of communication radios, setting of navigation systems, and control of an automated target handover system.

As in fighter applications, the overriding issue for voice in helicopters is the impact on pilot effectiveness. Encouraging results are reported for the AVRADA tests, although these represent only a feasibility demonstration in a test environment. Much remains to be done both in speech recognition and in overall speech technology in order to consistently achieve performance improvements in operational settings.

Training for air traffic controllers (ATC) represents an excellent application for speech recognition systems. Many ATC training systems currently require a person to act as a "pseudo-pilot", engaging in a voice dialog with the trainee controller, which simulates the dialog that the controller would have to conduct with pilots in a real ATC situation. Speech recognition and synthesis techniques offer the potential to eliminate the need for a person to act as a pseudo-pilot, thus reducing training and support personnel. In theory, air controller tasks are also characterized by highly structured speech as the primary output of the controller, hence reducing the difficulty of the speech recognition task should be possible. In practice, this is rarely the case. The FAA document 7110.65 details the phrases that should be used by air traffic controllers. While this document gives less than 150 examples of such phrases, the number of phrases supported by one of the simulation vendors speech recognition systems is in excess of 500,000.

The USAF, USMC, US Army, US Navy, and FAA as well as a number of international ATC training organizations such as the Royal Australian Air Force and Civil Aviation Authorities in Italy, Brazil, and Canada are currently using ATC simulators with speech recognition from a number of different vendors.^{[citation needed]}

ASR is now commonplace in the field of telephony and is becoming more widespread in the field of computer gaming and simulation. In telephony systems, ASR is now being predominantly used in contact centers by integrating it with IVR systems. Despite the high level of integration with word processing in general personal computing, in the field of document production, ASR has not seen the expected increases in use.

The improvement of mobile processor speeds has made speech recognition practical in smartphones. Speech is used mostly as a part of a user interface, for creating predefined or custom speech commands.

This section may require copy editing. (July 2025) (Learn how and when to remove this message)

Speech recognition programs can provide much benefit to those with disabilities. For individuals that are deaf or hard of hearing, speech recognition software can be used to generate captions of conversations.^[126] Additionally, individuals who are blind (see blindness and education) or have poor vision can benefit from listening to textual content, as well as garner more functionality from a computer by issuing commands with their voice.^[127]

The use of voice recognition software, in conjunction with a digital audio recorder and a personal computer running word-processing software, has proven useful for restoring damaged short-term memory capacity in individuals who have suffered a stroke or have undergone a craniotomy.^{[citation needed]}

Speech recognition have proven very useful for those who have difficulty using their hands due to causes ranging from mild repetitive stress injuries to disabilities that preclude using conventional computer input devices. Individuals with physical disabilities can use voice commands and transcription to navigate electronics hands-free.^[127] In fact, people who developed RSI from keyboard use became an urgent early market for speech recognition.^[128][129] Speech recognition is used in deaf telephony, such as voicemail to text, relay services, and captioned telephone. Individuals with learning disabilities who have problems with thought-to-paper communication can possibly benefit from the software, but the fallibility of the product remains an important consideration for many.^[130] In addition, speech to text technology is only an effective aid for those with intellectual disabilities if the proper training and resources are provided (eg., in the classroom setting).^[131]

This type of technology can help those with dyslexia, but the potential benefits regarding other disabilities are still in question. Mistakes made by the software hinder its effectiveness, since misheard words take more time to fix.^[132]

This section needs additional citations for verification. Please help improve this article by adding citations to reliable sources in this section. Unsourced material may be challenged and removed. (July 2025) (Learn how and when to remove this message)

• Aerospace, e.g., NASA's Mars Polar Lander used speech recognition technology from Sensory, Inc. in the Mars microphone on the lander^[133]
• Automatic subtitling with speech recognition
• Automatic emotion recognition^[134]
• Automatic shot listing in audiovisual production
• Automatic translation
• E-discovery
• Hands-free computing
• Home automation
• Interactive voice response
• Mobile telephony, including mobile email
• Multimodal interaction^[66]
• Real-time captioning^[135]
• Robotics
• Security, including usage with other biometric scanners for multi-factor authentication^[136]
• Speech to text
• Telematics, e.g., vehicle navigation systems
• Transcription
• Video games like Tom Clancy's EndWar and Lifeline
• Virtual assistant such as Siri

The performance of speech recognition systems is usually evaluated in terms of accuracy and speed.^[137][138] Accuracy is usually rated with word error rate (WER), whereas speed is measured with the real time factor. Other measures of accuracy include Single Word Error Rate (SWER) and Command Success Rate (CSR).

Speech recognition by machine is complicated by many properties of speech. Vocalizations vary in terms of accent, pronunciation, articulation, roughness, nasality, pitch, volume, and speed. Speech is distorted by background noise, echoes, and electrical characteristics. Accuracy of speech recognition may vary with the following:^{[139][citation needed]}

• Vocabulary size and confusability
• Speaker dependence versus independence
• Isolated, discontinuous, or continuous speech
• Task and language constraints
• Read versus spontaneous speech
• Adverse conditions

The accuracy of speech recognition may vary depending on the following factors:

• Error rates increase as the vocabulary size grows:

e.g. the 10 digits "zero" to "nine" can be recognized essentially perfectly, but vocabulary sizes of 200, 5000, or 100000 may have error rates of 3%, 7%, or 45% respectively.

• Vocabulary is hard to recognize if it contains confusing letters:

e.g. the 26 letters of the English alphabet are difficult to discriminate because they are confusing words (most notoriously, the E-set: "B, C, D, E, G, P, T, V, Z (when "Z" is pronounced "zee" rather than "zed", depending on the English region); an 8% error rate is considered good for this vocabulary.^[140]

• Speaker dependence vs. independence:

A speaker-dependent system is intended for use by a single speaker.

A speaker-independent system is intended for use by any speaker (more difficult).

• Isolated, Discontinuous or continuous speech

With isolated speech, single words are used, therefore it becomes easier to recognize the speech.

With discontinuous speech, full sentences separated by silence are used. Because of this, the speech becomes easier to recognize similar to isolated speech.
With continuous speech naturally spoken sentences are used, therefore it becomes harder to recognize the speech, different from both isolated and discontinuous speech.

• Task and language constraints
  • e.g. Querying application may dismiss the hypothesis "The apple is red."
  • e.g. Constraints may be semantic; rejecting "The apple is angry."
  • e.g. Syntactic; rejecting "Red is apple the."

Constraints are often represented by grammar.

• Read vs. Spontaneous Speech – When a person reads it's usually in a context that has been previously prepared, but when a person uses spontaneous speech, it is difficult to recognize the speech because of the disfluencies (like "uh" and "um", false starts, incomplete sentences, stuttering, coughing, and laughter) and limited vocabulary.
• Adverse conditions – Environmental noise (e.g. Noise in a car or a factory). Acoustical distortions (e.g. echoes, room acoustics)

Speech recognition is a multi-leveled pattern recognition task.

• Acoustical signals are structured into a hierarchy of units, e.g. phonemes, words, phrases, and sentences;
• Each level provides additional constraints;

e.g. Known word pronunciations or legal word sequences, which can compensate for errors or uncertainties at a lower level;

• This hierarchy of constraints is exploited. By combining decisions probabilistically at all lower levels, and making more deterministic decisions only at the highest level, speech recognition by a machine is a process broken into several phases. Computationally, it is a problem in which a sound pattern has to be recognized or classified into a category that represents a meaning to a human. Every acoustic signal can be broken into smaller more basic sub-signals. As the more complex sound signal is broken into the smaller sub-sounds, different levels are created, where at the top level we have complex sounds, which are made of simpler sounds on the lower level, and going to lower levels, even more, we create more basic and shorter and simpler sounds. At the lowest level, where the sounds are the most fundamental, a machine would check for simple and more probabilistic rules of what sound should represent. Once these sounds are put together into more complex sounds on upper level, a new set of more deterministic rules should predict what the new complex sound should represent. The most upper level of a deterministic rule should figure out the meaning of complex expressions. In order to expand our knowledge about speech recognition, we need to take into consideration neural networks. There are four steps of neural network approaches:
• Digitize the speech that we want to recognize

For telephone speech, the sampling rate is 8000 samples per second;

• Compute features of spectral-domain of the speech (with Fourier transform);

computed every 10ms, with one 10ms section called a frame;

Analysis of four-step neural network approaches can be explained by further information. Sound is produced by air (or some other medium) vibration, which we register by ears, but machines by receivers. Basic sound creates a wave which has two descriptions: amplitude (how strong is it), and frequency (how often it vibrates per second). Accuracy can be computed with the help of word error rate (WER). Word error rate can be calculated by aligning the recognized word and referenced word using dynamic string alignment. The problem may occur while computing the word error rate due to the difference between the sequence lengths of the recognized word and referenced word.

The formula to compute the word error rate (WER) is:

${\displaystyle WER={(s+d+i) \over n}}$

where s is the number of substitutions, d is the number of deletions, i is the number of insertions, and n is the number of word references.

While computing, the word recognition rate (WRR) is used. The formula is:

${\displaystyle WRR=1-WER={(n-s-d-i) \over n}={h-i \over n}}$

where h is the number of correctly recognized words:

${\displaystyle h=n-(s+d).}$

Speech recognition can become a means of attack, theft, or accidental operation. For example, activation words like "Alexa" spoken in an audio or video broadcast can cause devices in homes and offices to start listening for input inappropriately, or possibly take an unwanted action.^[141] Voice-controlled devices are also accessible to visitors in the same building, or even those outside of the building if they can be heard inside. Attackers may be able to gain access to personal information, like calendars, address book contents, private messages, and documents. They may also be able to impersonate the user to send messages or make online purchases.

Two attacks have been demonstrated that use artificial sounds. One transmits ultrasound and attempts to send commands without nearby people noticing.^[142] The other adds small, inaudible distortions to other speech or music that are specially crafted to confuse the specific speech recognition system into recognizing music as speech, or to make what sounds like one command to a human sound like a different command to the system.^[143]

Popular speech recognition conferences held regularly include SpeechTEK and SpeechTEK Europe, ICASSP, Interspeech/Eurospeech, and the IEEE ASRU. Conferences in the field of natural language processing, such as ACL, NAACL, EMNLP, and HLT, are beginning to include papers on speech processing. Important journals include the IEEE Transactions on Speech and Audio Processing (later renamed IEEE Transactions on Audio, Speech and Language Processing and since Sept 2014 renamed IEEE/ACM Transactions on Audio, Speech and Language Processing—after merging with an ACM publication), Computer Speech and Language, and Speech Communication.

Books like Fundamentals of Speech Recognition by Lawrence Rabiner can be useful to acquire basic knowledge, but may not be fully up to date (1993). Another good source can be Statistical Methods for Speech Recognition by Frederick Jelinek and Spoken Language Processing (2001) by Xuedong Huang et al., Computer Speech by Manfred R. Schroeder, second edition published in 2004, and Speech Processing: A Dynamic and Optimization-Oriented Approach, published in 2003 by Li Deng and Doug O'Shaughnessey. The updated textbook Speech and Language Processing (2008) by Jurafsky and Martin presents the basics and the state of the art for ASR. Speaker recognition also uses the same features, most of the same front-end processing, and classification techniques as are used in speech recognition. A comprehensive textbook, Fundamentals of Speaker Recognition, is an in depth source for up to date details on the theory and practice.^[144] A good insight into the techniques used in the best modern systems can be gained by paying attention to government sponsored evaluations, such as those organised by DARPA. The largest speech recognition-related project ongoing as of 2007 is the GALE project, which involves both speech recognition and translation components.

A good and accessible introduction to speech recognition technology and its history is provided by the general audience book The Voice in the Machine. Building Computers That Understand Speech by Roberto Pieraccini (2012).

A more recent book on speech recognition is Automatic Speech Recognition: A Deep Learning Approach (Publisher: Springer), written by Microsoft researchers D. Yu and L. Deng and published near the end of 2014, with highly mathematically-oriented technical detail on how deep learning methods are derived and implemented in modern speech recognition systems based on DNNs and related deep learning methods.^[86] A related book published earlier in 2014, Deep Learning: Methods and Applications by L. Deng and D. Yu, provides a less technical but more methodology-focused overview of DNN-based speech recognition during 2009–2014, placed within the more general context of deep learning applications, including not only speech recognition but also image recognition, natural language processing, information retrieval, multimodal processing, and multitask learning.^[82]

In terms of freely available resources, Carnegie Mellon University's Sphinx toolkit is one starting point for learning about and experimenting with speech recognition. Another resource (free but copyrighted) is the HTK book and its accompanying toolkit. For more recent and state-of-the-art techniques, Kaldi toolkit can be used.^[145] In 2017, Mozilla launched the open source project Common Voice^[146] to gather a database of voices that would help build free speech recognition project DeepSpeech (available free at GitHub),^[147] using Google's open source platform TensorFlow.^[148] When Mozilla redirected funding away from the project in 2020, it was forked by its original developers as Coqui STT^[149] using the same open-source license.^[150][151]

Google Gboard supports speech recognition on all Android applications. It can be activated through the microphone icon.^[152] Speech recognition can be activated in Microsoft Windows operating systems by pressing Windows logo key + Ctrl + S.^[153]

The commercial cloud based speech recognition APIs are broadly available.

• Cole, Ronald; Mariani, Joseph; Uszkoreit, Hans; Varile, Giovanni Battista; Zaenen, Annie; Zampolli; Zue, Victor, eds. (1997). Survey of the state of the art in human language technology. Cambridge Studies in Natural Language Processing. Vol. XII–XIII. Cambridge University Press. ISBN 978-0-521-59277-2.
• Junqua, J.-C.; Haton, J.-P. (1995). Robustness in Automatic Speech Recognition: Fundamentals and Applications. Kluwer Academic Publishers. ISBN 978-0-7923-9646-8.
• Karat, Clare-Marie; Vergo, John; Nahamoo, David (2007). "Conversational Interface Technologies". In Sears, Andrew; Jacko, Julie A. (eds.). The Human-Computer Interaction Handbook: Fundamentals, Evolving Technologies, and Emerging Applications (Human Factors and Ergonomics). Lawrence Erlbaum Associates Inc. ISBN 978-0-8058-5870-9.
• Pieraccini, Roberto (2012). The Voice in the Machine. Building Computers That Understand Speech. The MIT Press. ISBN 978-0262016858.
• Pirani, Giancarlo, ed. (2013). Advanced algorithms and architectures for speech understanding. Springer Science & Business Media. ISBN 978-3-642-84341-9.
• Signer, Beat; Hoste, Lode (December 2013). "SpeeG2: A Speech- and Gesture-based Interface for Efficient Controller-free Text Entry". Proceedings of ICMI 2013. 15th International Conference on Multimodal Interaction. Sydney, Australia.
• Woelfel, Matthias; McDonough, John (26 May 2009). Distant Speech Recognition. Wiley. ISBN 978-0470517048.