In the AI Startup of the Week, the editorial staff of ai.nl is featuring promising AI startups, their innovations, solutions and challenges. In this eighteenth episode, we are taking a look at Amberscript, an Amsterdam-based AI startup making it easier and accessible for anyone to transform audio and video into text.
Last year, Amberscript raised €8.65M in November from Endeit Capital with the goal of transforming audio and video to text and subtitles using AI. With the funding, Amberscript established itself as a major player in the field of speech to text and showed how AI could disrupt this industry.
While there are a number of companies that offer translation or transcription services, Amberscript is one of those startups looking at speech to text from an accessibility lens. This ability of Amberscript to challenge the conventional wisdom in an industry dominated by human and machine intelligence makes it a must follow tech company in the AI landscape.
Transforming audio and video to text and subtitles
For the past few years, creators as well as big organisations have realised how they have not focussed on accessibility in their work. A number of audio and video content out there have lacked subtitles or captions to make them accessible. It changed to a certain degree when Google began experimenting with auto-captions on YouTube.
This immediately made creators and professionals think about a way to add subtitles to their content in a cost effective manner. Based in Amsterdam and founded in 2017, Amberscript does just that with its mission to “make audio accessible.”
It offers a simple service that allows individuals or organisations to create text and subtitles from audio or video. Amberscript, like many others, follows a methodology where users can either create text out of their audio or video file automatically and then perfect it or use language experts and professional subtitlers.
Amberscript and the work it is doing is essentially a perfect example of how AI systems and humans will complement each other. In the case of automatic transcription, the text file from audio or video is created by AI. Amberscript has built powerful language models that can understand the nuance of the speech and turn that into text.
How does Amberscript work?
Amberscript is a simple online tool that can be accessed by visiting this link. On the website, all one needs to do is upload their audio or video file. Once uploaded, the speech recognition engine built by Amberscript or professional transcribers will handle your request.
Once the file is processed, Amberscript presents it in the form of a text file in an online text editor. This is meant to help users revise, highlight, and search through text with ease.
After an audio or video file is converted into text, Amberscript users can quickly export their transcript in text, SRT, VTT, or any other format. The platform also offers an option to add timestamps and speaker distinction to these files.
Behind the scenes, Amberscript combines domain-specific AI speech recognition engine with its ‘human layer’ of transcribers. The Dutch startup, founded by Peter-Paul de Leeuw, Thomas Dieste and Timo Behrens, claims that its AI algorithm produces subtitles with the highest accuracy in the market.
The startup also claims that its AI speech engine is capable of delivering results eight times faster than traditional manual methods of transcription. It offers this service tailored to businesses, academic research, communications and media professionals, and professionals in healthcare, psychology, journalism, and law.
A look at how Amberscript is used in the industry
Amberscript is helping Sporters, a sports education platform from Milan, Italy. The problem facing Sporters is that their internal team has to spend hundreds of hours manually captioning the video in Adobe Premiere to ensure the videos are accessible for its entire customer base.
Amberscript is solving this problem for Sporters with its subtitling solution, which reduces the time spent on manual transcription. Sporters said earlier one person would spend two days of manual post-production work on one hour of video content. “With Amberscript, the time saved is a game-changer. The workload is reduced by over 40 per cent,” says Andrea Barosi, CEO of RedFir, the video production agency helping Sporters.
It is not just in the sports field where Amberscript is making a difference. The platform is also being used by a research team at Amsterdam’s University of Applied Sciences (HVA). The research team CAREM conducts many kinds of research, which also involves recorded interviews that need to be transcribed.
The manual process was time consuming and Amsterdam’s University of Applied Sciences says it has reduced the time needed to transcribe documents from 3 weeks to just 5 to 7 days with Amberscript. It has managed to do that by ensuring the hourly rate of transcription remains the same.
Power of captions and subtitles
Amberscript is showing the power of subtitles, captions and transcription to an industry that has stayed away from the idea of accessibility. Today, the success of Netflix and its content like Squid Game or Money Heist is purely on the strength of accessible captions.
This accessibility is possible because of services like Amberscript, which take speech and turn them into text. By doing this conversation while accounting for language and accent makes them stand out in the crowd.
What to read next?
- 📖11 best books on AI, its impact in marketing and technology you must read right now
- 🤖Cleaning is a repetitive task and Dutch AI startup Loop Robots is automating it with smart mobile robots
- 🤗Hugging Face wants to be GitHub of machine learning: a look at its history, key members, and major achievements