How 3Play Media Utilizes AI Technology for Captioning

Artificial intelligence (AI technology) makes it possible for machines to learn from experience, adjust to new inputs, and perform human-like tasks.

At 3Play, AI technology plays an integral part in improving the efficiency and accuracy of the captioning process. Our 3-step process combines artificial intelligence and human editing to maximize efficiency and optimize our process.

Regardless of audio difficulty or file size, we guarantee a 99% accuracy rate on all files.

Download the State of Automatic Speech Recognition Report ✨

3Play Media & AI Technology

Since the beginning, 3Play has been utilizing AI technology through the use of Automated Speech Recognition (ASR). ASR is the first step in our patent-pending 3-step captioning process. ASR converts the spoken words in a video or audio into computerized text.

As you upload more content into our account system for captioning, the ASR improves through machine learning. Machine learning is a system that allows machines to learn from outputs from previous experience.

Through AI technology we are able to speed up the editing process to offer faster turnaround options, decrease costs for high-quality captioning, and precisely synchronize timestamps to spoken audio.

3Play Media’s Highly Accurate ASR Solution

In a comprehensive analysis between ASR technologies in the industry, 3Play Media’s engine performed significantly higher than ASR engines like IBM Watson, Rev’s Temi, Google, and Trint.

Despite audio that was filled with low volume, accents, and complicated terminology, 3Play’s engine averaged an accuracy rate of 90.91% for ASR only. In contrast, engines like Watson, Google, and Temi averaged between 80% – 89% accuracy in ASR only.

Utilizing ASR and Humans for Higher Accuracy

When it comes to ASR in captioning, there are limitations to the quality of the output of spoken text due to the long-form nature of the content. Therefore, humans are an essential part of a highly-accurate, efficient, and cost-effective captioning process.

Artificial intelligence can generate a transcript with an accuracy rate between 80%-90%. An accuracy rate of 99% is imperative to the comprehension of the closed captions. ASR often misses speaker ids, punctuation and grammar, non-speech elements, [INAUDIBLE] tags, and many more critical components of accurate captions.

At 3Play Media, we guarantee at least a 99% accuracy rate on all your files. Our process directly impacts the accuracy of your files. After ASR, every file is put through two rounds of human editing and review.

How is 3Play Different from the Competition?

At 3Play, we are constantly innovating our features and services. Technology has allowed us to remain the best video accessibility service on the market.

1. Best in-class AI technology: Our ASR engine outperforms other engines on the market, even in cases of low quality audio.
2. Customer-specific learning: Our ASR engine learns from the files you upload and continuously improves.
3. Our three step process: Our three-step process was created to ensure maximum efficiency and quality. We use a mix of automatic speech recognition (ASR) and human editing to speed up the process and ensure our editors focus on the quality of the transcript.
4. Word-for-Word Synchronization: To ensure maximum accuracy and completeness, we timecode you files based on each spoken word. No other vendor offers this.
5. Ability to upload wordlists: For our live automatic captioning solution, you can upload wordlist to rapidly improve the quality of the ASR.
6. Consistent Accuracy with Every File: Regardless of how long your file is, or even how many files you upload, we always guarantee a 99% accuracy. With other captioning vendors, accuracy rates tend to crumble the longer the file.