Amazon Transcribe & Translate Ready For Work
Written by Mike James   
Friday, 06 April 2018

Amazon has two new services with APIs ready to be built into your apps. Transcribe will convert speech to text and Translate translates text into another language.

AmazonAI

Amazon Transcribe is a speech-to-text API that you can easily build into your own apps and now it is out of the trial stage and ready to be used by anyone. Transcribe claims to be based on deep learning, so there is probably a neural network in there somewhere. Basically all you have to do is capture some audio and send the file stored on AWS S3 to the server. It processes it and sends back text.

The service is able to detect different speakers in your audio with high accuracy and confidence to produce intelligible transcriptions so that you can clearly understand who spoke when. You can also expand and tailor the vocabulary of the speech engine by uploading a custom vocabulary to improve the accuracy of the speech recognition for product names, domain-specific terminology, or names of individuals.

It can return a time-stamped transcription, which makes it useful for things like captioning and cataloging. Of course, you will need to try it to discover if it is good enough for your purpose and there is free tier for 12 months use with a limit of 60 minutes of audio per month. After that it is $0.0004 per second of audio and if that means nothing to you it is equivalent to about $0.72 for 30 minutes.

Amazon Translate is also based on deep learning and can work with six languages - Spanish, Chinese, French, German, Arabic, and Portuguese - and there are plans to add six more in the coming months - Japanese, Russian, Italian, Traditional Chinese, Turkish, and Czech.

Neural machine translation is a form of language translation automation that uses deep learning models to deliver more accurate and more natural sounding translation than traditional statistical and rule-based translation algorithms. Amazon Translate allows you to localize content - such as websites and applications - for international users, and to easily translate large volumes of text efficiently. 

To try it out there is a free tier for 12 months with a limit of 2 million characters per month. After that you pay $15 per million characters. To give you some idea of how much this costs, a typical book is 300,000 characters and that would cost around $5 to translate.

As well as the fact that the services are available and add to the AI offerings Amazon already has, it is interesting to compare its approach with Google. Google Translate is often viewed as a free service because it is almost instantly available to end users via Chrome or Android apps. The availability of free translation is one of the things that gives Chrome its edge over Firefox. If you are likely to get an email in a foreign language, then Chrome is what you want. Amazon now has a translator but it is only free for 12 months. If the aim is to dent Google, perhaps Amazon should team up with Mozilla to provide free translation to end users. Google's existing paid translation service is more expensive than Amazon's at $20 per million characters.

AmazonAI

 

Microsoft also has a translation service as part of Azure and it offers a free service up to 2 million characters per month and then charges $7.454 per million characters. There are various other deals available but overall Microsoft's looks to be the cheapest.

Slowly but surely, it looks as if translation and language services are where AI is going to become mainstream. The fight for your AI dollars doens't quite look like the battle over office suites of the past, but Microsoft, Google and Amazon are rolling out products that you are probably going to choose as a set rather than mix and match. Amazon being top cloud supplier has an advantage. If you have an AWS account why not try the AI services? The same is true for Microsoft, but not quite as much. Google, on the other hand, has Chrome and provides translation to the end user for free.  Amazon gives away transcription and voice synthesis as part of Alex - whoever said battle lines were easy.

awslogo

More Information

Amazon Translate is Now Generally Available

Amazon Transcribe is Now Generally Available

Related Articles

Amazon AI Services 

Amazon Rekognition Can Now Estimate Your Age

What Makes A Bot? 

New AWS Services

Amazon Opens Up Alexa To Developers and Third Parties

Amazon Uses Machine Learning For Fraud Detection


To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on, Twitter, FacebookGoogle+ or Linkedin.

 

Banner


Microsoft Developer Blogs On The Move
26/10/2018

Microsoft is in the process of moving its developer blogs to a new blogging platform. The aim is to provide "a clean design and powerful features that will make it easy for you to discover and share g [ ... ]



AI News Anchor - A First For China
10/11/2018

Two AI avatars that can read news in English and Mandarin were unveiled last week. This adds TV News Anchor to the long list of jobs in which machines will soon replace humans.


More News

Python

 



 

Comments




or email your comment to: comments@i-programmer.info

 

 

Last Updated ( Friday, 06 April 2018 )