Exclusives

Digital Nirvana used the breakout session “Accelerating Workflows with AI & Auto-Generated Speech To Text” on March 10 at the Smart Content Summit in Los Angeles to explain how its MetadataIQ platform and Trance application use artificial intelligence (AI) and machine learning (ML) to accelerate workflows.

AI-based speech-to-text tools are “being used to accelerate the delivery of transcriptions, captions, subtitling, metadata generation, tagging and the enhancement of associated media assets,” according to Tom Moniak, director of sales at Digital Nirvana.

The company’s two automated solutions are “accelerating delivery and production timelines,” applying AI and ML to “existing workflows to accurately auto-generate text of live and historical assets,” he explained.

Trance is a Software-as-a-Service (SaaS)-based application that supports “automation for transcriptions, captions and subtitles to pre-defined delivery specs,” he said.

Initially created as an internal tool for Digital Nirvana’s turnkey, transcription, captioning and subtitling services, Trance is now commercially available for content creators and media operators and can easily integrate with existing media workflows, he noted.

Trance provides translations in more than 100 languages and improves efficiencies by 50% over conventional desktop applications, according to Digital Nirvana.

MetadataIQ, meanwhile, is a hybrid offering from Digital Nirvana that automates the generation of speech-to-text and video intelligence metadata, increasing the efficiency of production, pre-production, and live content creation services for Avid production asset management (PAM) and media asset management (MAM) users, according to the company.

MetadataIQ leverages the AI, ML and speech-to-text capabilities of Trance and applies that knowledge to the realm of media production. MetadataIQ auto-generates real-time transcripts and time-indexes this metadata to the media as markers within the Avid Interplay PAM environment.

As a result, MetadataIQ significantly accelerates the identification, search and retrieval of valuable media assets while reducing the time, effort and complexity of producing finished content, according to Digital Nirvana.

Trance Use Cases

Moniak went on to provide examples of three Trance use cases.

The first use case involved a client that Digital Nirvana identified as the “most extensive media enterprise in the Spanish-speaking world.”

The challenge was that client needed a faster, more efficient way to generate subtitles and time-code-accurate sidecar files in English and multi-lingual subtitles for an enormous archive, and newly produced, premium high-demand content natively produced in Spanish conformed to global over-the-top (OTT) platforms.

“Using some preset parameters, they can now produce a typical one-hour program and process it in under two hours,” generating transcripts and closed captioning in Spanish and English, while meeting all OTT delivery requirements, Moniak said.

The second use case involved a client that Digital Nirvana identified as a “major broadcaster of tennis sports content/governing body” for tennis in the U.S.

The challenge in that case was manual captioning was taking several hours, which was an unsatisfactory turnaround time, according to Digital Nirvana. The client needed to speed up the captioning process and free its technical personnel to focus on the creative aspects of their jobs.

For that client, Trance has been able to cut the turnaround time for captioning an average video “from hours to only 30 minutes,” with the captioning task completely offloaded from the technical team and automatically published, Moniak said.

The third use case involved a client that Digital Nirvana identified as a “popular Hollywood video tabloid.” It is a client that is “routinely taking in large volumes” of undocumented content with no context and must use the content and turn it around quickly, explained Ed Hauber, director of business development at Digital Nirvana.

The business challenge that organization faced was that contractual obligations for submitted content had to include closed captions no more than 90 minutes from the time of acquisition to the time the content would be published.

In that use case, the client adapted Digital Nirvana’s tool to expand beyond captioning, using it as a form of metadata generation to help users create context and perspective around field content the organization received, Hauber said.

The organization is specially “leveraging three key Trance capabilities to meet and exceed its requirements for fast and seamless content delivery,” according to Digital Nirvana.

Those capabilities are:

1. Speech-to-Text (STT) transcription applying advanced AI-based algorithms to automatically generate accurate transcripts for raw production footage.
2. Automated caption generation, outputting closed captions based on the style guide and enabling users to confirm compliance with output requirements.
3. An enterprise feature set that “provides unprecedented control and flexibility with maximum efficiencies,” Digital Nirvana said.

More on MetadataIQ

Hauber went on to explain how MetadataIQ accelerates content creation in greater detail, noting it offers:

• A secure, scalable platform to automate the process of metadata generation.
• Process production, pre-production and live content residing in Avid.
• Integrated video intelligence for logo detection, face recognition, object identification, etc.
• Automated ingestion of time-coded STT and video intelligence metadata as markers.
• Converts processed metadata into formats required for Avid MAM or other systems.
• Hybrid application enabling metadata generation from Avid system.
• ML and AI application for content analysis.
• Assets are easily searchable thanks to the automation of metadata ingestion.
• Reviews raw footage for faster content generation.
• Easy access to archived content to enable faster repurposing.

To view the presentation, click here.

To download the presentation deck, click here.

The 2022 Smart Content Summit event was held in conjunction with the EIDR Annual Participant Meeting (EIDR APM), and was presented by Whip Media. The event was produced by MESA, in association with the Smart Content Council and EIDR, with sponsorship by BeBanjo, Signiant, Qumulo, Adio, Alteon, Digital Nirvana, Slalom and Rightsline.