Best speech to text software

Voice recognition technology continues to evolve, with AI and virtual assistant technology applied to services like speech-to-text and transcription software.

This software has wide-ranging applications, from assisting people with disabilities to helping people learn languages, as well as being particularly handy for journalists tasked with transcribing long interviews.

Here we examine some of the best speech to text software available right now.

Read next: Best software tools for journalists

Otter
iStock

Otter

Otteris a note-taking and collaboration app that can record and transcribe meetings, interviews and other voice interactions in real-time.

Created by AISense, Otter is built with Ambient Voice Intelligence making it particularly smart and accurate. It provides transcriptions within minutes, providing an opportunity to share files amongst co-workers.

Users can also record conversations from the phone or computer and is integrated with Zoom, for easy recordings of conference calls.

Otter is free for 600 minutes of transcription a month and pricing starts from $9.99 (£7.65) a month for 6000 minutes of transcription a month.

The app is available to download on iOS and Android.

Verbit
iStock

Verbit

Verbit, an Israeli-based AI transcription company, provides a smart transcription and captioning solution using adaptive algorithms.

The company claims to provide the only technology which generates speech-to-text files with over 99 percent accuracy at a high speed.

It has over 100 customers utilising its services in the legal and higher education sectors. Verbit’s solution has also been of help for university students that suffer from hearing impairments.

The technology also uses human intelligence via an on-demand network of transcribers, with on-demand CART services for real-time results using AI technology.

In January 2019, Verbit announced the close of a $23 million Series A round led by Viola Ventures, Vertex Ventures, HV Ventures and other big names.

Verbit’s pricing is available on request.

Braina
iStock

Braina

Brainais an AI-based virtual assistant software that includes voice recognition for Windows PC users.

The voice recognition software enables users to convert voice to text on any website or word document, and supports over 100 languages.

Braina also includes natural language commands for ease of use across multiple computer software and websites. Users can also control a PC using the Android or iOS app, allowing tasks to be completed wirelessly from a smartphone or tablet device.

Similarly, the voice recognition software also offers text to speech to listen to e-books, emails and webpage content read aloud by the virtual assistant.

Braina Pro, which offers a one-year download, is available at $49 (£37.59) whilst the lifetime version is available at $139 (£106.64).

SpeechTexter
iStock

SpeechTexter

SpeechTexteris free multilingual speech-to-text software for Chrome desktop browser and Android OS.

Its transcription service is available to transcribe any type of speech in over 60 languages that can be formatted into books, reports, and blog posts using speech dictation.

The core features include continuous real-time speech recognition and a custom dictionary that allows users to add their own commands or punctuation.

GoTranscribe
iStock

GoTranscribe

GoTranscribeoffers cloud-based speech-to-text software which is powered by AI.

Supported by any iOS and Android device, the software will automatically transcribe any audio file within minutes after it is uploaded onto the platform.

It also allows users to easily access specific parts of the text using the online editor. All transcripts can be downloaded in different formats such as Microsoft Word, PDF and SubRip-SRT.

GoTranscribe is available at 18p per min on the pay as you go plan, and pricing starts at £36 a month for an individual, £72 a month for business and £300 a month for the corporate package.

Speechmatics
iStock

Speechmatics

Speechmaticsis a Cambridge-based company that develops automatic speech recognition (ASR) software based on neural networks and statistical language modeling.

Founded in 2012, the software is available in over 70 languages. It offers a batch of ASR products, both on-premise and cloud-based with a transcription service or audio and video recorded files.

According to Speechmatics, the addition of machine learning provides high accuracy levels for several use cases.

Its cloud-based transcription service is available at 6p per minute, and blocks of £10 or £100 for multiple files. It also offers free trials for up to 60 minutes of audio.

Dragon Naturally Speaking
© TechSpective

Dragon Naturally Speaking

Nuance’s Dragon Naturally Speaking (DNS), also known as Dragon for PC, is speech recognition software available in eight languages. It claims a 99 percent accurate speech to text transcription and is recognised as one of the most accurate pieces of software on the market.

DNS includes voice commands which enable users to format and edit text using voice, with the addition of custom commands for computer actions and transcribing files. It can be run on Windows and Macs, as well as iOS and Android devices.

DNS integrates its voice commands and editing features together to ensure the text is provided just as the user delivers the speech, with specific punctuation and formatting options inserted upon command by the speaker.

The newest version of the software, DNS 13, is available in home, premium and professional editions, with the latter allowing users to import and export custom word lists to add any specific words and acronyms that may not already be in the vocabulary.

Nuance charges £79.99 for its Home Edition, £125.99 for premium and the professional edition starts at £279.99.

Transcribe (Wreally)
iStock

Transcribe (Wreally)

The name gets poor marks for imagination, but it does convey the clean functionality with which Transcribe converts your (clear, English) audio into text. 

The automatic transcriber effortlessly converts your recordings into written text files. However, if your sound quality isn't clear, you can also slow the text down and read it aloud, and the software will capture what you're saying - for a range of languages.

As a last resort, you can use the application to support manual transcribing, by playing the sound file within the software and employing a number of useful controls to streamline the process.

The basic self-transcribe package comes in at $20 (£15) per year (although if only opting for this, there are free alternatives such as oTranscribe out there) and $6 (£4.50) per 60 minutes of automatic transcription.

IBM Speech to text
iStock

IBM Speech to text

You can trial IBM's Watson-powered speech to text service for free on their Lite Plan, which allows for up to 100 minutes of free audio input per month. The full package offers automatic real-time transcription for seven languages, even for lower quality audio with powerful speech recognition technology.

You can customise the service to recognise particular product or peoples names, for example. This software is designed to be adaptable and scalable for many different use cases spanning from real-time transcription from a microphone to analysing thousands of recordings from your business's call centre.

If you need to go beyond a Lite Plan, a Standard plan offers the first 1000 minutes free, and after that, between one and 250,000 minutes are charged at $0.02 (£0.015) per minute.

Amazon Transcribe

Amazon Transcribe

Amazon Transcribeis an automatic speech recognition service which is powered by deep learning. It is best suited for developers that may want to add speech-to-text to their apps.

It can also be used to transcribe audio and video files, including a detailed transcript of customer service calls. This makes it a good choice for businesses that may need to transcribe client meetings.

The built-in API is simple to use and provides a quick transfer of transcripts from files stored in Amazon S3. It also includes a timestamp for each word, as well as the opportunity to customise vocabulary to add any new words or acronyms that may not be in there already.

Amazon Transcribe provides a free tier of 60 minutes a month for one year, following that the pricing is $0.0004 (£0.00028) per second.

Temi
© Temi

Temi

Temiuses automated software to provide a detailed speech to text transcription in five minutes.

The app allows users to make recordings, edit and share them with the ability to order transcripts. It is available on Apple’s app store and Google Play for free download.

Transcripts are priced at $0.10 (7p) per minute. This includes custom timestamps, speaker identification and the opportunity to export transcripts to MS Word and other formats directly from the app. Temi also delivers transcripts directly to the user's email.

Windows speech recognition
© YouTube

Windows speech recognition

Windows Speech Recognition(WSR) offers speech recognition for Microsoft Windows Vista and later.

The speech to text feature allows users to dictate text within documents and email messages. It is available in eight languages including UK/US English, Chinese, and Spanish.

The accuracy of WSR is best with good-quality audio. As it is automatically built into the Windows OS, there is no additional price to access the speech recognition software.

However, compared to other speech to text platforms, WSR users will need to select a 'play audible feedback' option after the transcribed text is selected to ensure accuracy.

Speech Recognition SoundWriter
Getty Images

Speech Recognition SoundWriter

Speech Recognition SoundWriteris a voice typing and recognition software addon for Google Docs.

As long as users have a microphone installed, SoundWriter can be used to write all Google Doc files in any language. It also includes custom configuration, which provides the opportunity to add words as commands, as well as any specific characters and punctuation styles.

Although the addon only works with the Google Chrome browser, SoundWriter can be a good option for regular users of Google Docs. It also has voice typing integrated and is available for free download.

Google Docs Voice Typing
© YouTube

Google Docs Voice Typing

Google Docs Voice Typingprovides web-based dictation and speech recognition for Google Docs users, particularly for those that may only require the speech to text feature across their Google Suite platform.

It can be accessed in 43 different languages for voice typing, but voice commands are only available in English. It is free to install to the Google Chrome web browser.

It is easy to use as clicking on the microphone icon controls start and stop of dictation and all delivered speech is transcribed directly as you speak. So, it is a good option for instant transcription of client meetings and recordings.

With the added commands, Google Docs Voice Typing follows instant instructions to delete or edit text as it is commanded.

Trint
© Trint

Trint

Trint is speech to text transcription software which provides automatic transcripts of audio and video files using artificial intelligence.

Users can access a quick search and edit feature that is embedded into the software to ensure detailed accuracy. The software also makes it easy to share and collaborate transcribed content, so a good choice if working in a team.

It is available in 12 languages, plus three English accents. There are three price plans available for Trint, these are pay as you go at £13.20 an hour, basic at £26 a month and supercharged at £100 a month.

The pay as you go plan provides unlimited credit for 12 months, the basic plan offers three hours a month and supercharged provides 10 hours a month.

Copyright © 2019 IDG Communications, Inc.