Release Notes

21.1 Product Update

Stephanie Rosas
April 6, 2021

We’ve gone a long way since the first versions of Limecraft Edge. Back then, it was a great way to offload rushes while preserving their folder structure while at the same time creating and transferring proxies for review and pre-cut.

Now, with the release series of Edge 6.x, we support remote post-production better than ever before. Expect reliable ingest and download vast collections, extended file format support, and a state-of-the-art encoding pipeline with built-in functions for image manipulation such as scaling and cropping, watermarking and many more.

In the area of transcription and translation, we’ve made several improvements as well. Capitalising on the results of the MeMAD project, we’ve drastically improved the speaker segmentation. Also, we’ve included support for several languages and language flavours, so you can now translate transcripts and subtitles back and forth Brazilian Portuguese, Hungarian, Czech, Romanian, Slovenian and many more.

Finally, when you look at Limecraft Flow, you’ll notice we’re in the process of reworking the way we present metadata and workflows in Limecraft Flow, as well as the way you search for the right fragments. Expect more to come.

Limecraft Edge – the platform of choice for many DIT’s and Assistant Editors

Limecraft edge series 6 has been used for a while by early adopters, and it is now publicly available. It has been rebuilt from the ground up for maximum performance and reliability. While earlier versions of Limecraft Edge were suited for processing and uploading a shooting day or -say- a few 100’s of files, you can now use Edge to safely process an entire stock, including several 1000’s of files.

Integrated backup and import

The most visible change on the surface is the integration of the backup and import function. Recall that backup is the process of transferring media from the original disc or card to a backup drive. In contrast, import referred to the indexing process, making sure Limecraft Edge knows the files before playout, manipulation or ingest.

Limecraft Edge Series 6

Central Management of Settings

The configuration of Limecraft Edge is now accessible through the production settings in Limecraft Flow. This configuration is because video encoding settings are usually specific for a particular production and because many customers are using different instances of Edge for various productions. To ensure overall consistency, we’ve made sure you can manage edge settings on the production level in Limecraft Flow. These are automatically distributed to the different instances of Limecraft Edge upon login.

A new encoding Pipeline

The most functional change is that you can now control each encoding format’s settings (raw, edit proxy, review proxy, archived) in great detail, including scaling and cropping, audio mapping, watermarking, etc.

Central Management of the encoding profile of Limecraft Edge

Remote control of Limecraft Edge via Flow

To achieve maximum functionality for teams working remotely, we’ve further optimised the symbiosis between Limecraft Flow and Edge. You can now instruct Edge to perform specific actions using Limecraft Flow as a remote control.

For example, you can instruct Limecraft Edge to execute a download in a specific format to a particular destination using Limecraft Flow. By delegating this task to a stand-alone application, you don’t need to keep the browser window open.

Limecraft Flow to command a download function in Limecraft Edge

Transcription and Translation Improvements

We pride ourselves on offering you the best possible automated transcription and translation services available on the market.
Consequently, we are continuously improving our services, optimising them for the shortest time required for post-editing.

By systematically adding words to the lexicon and updating language models, and making available custom dictionaries, the recognition of words now comes close to perfection. Yet, we took the extra mile, and we’ve improved the results even more.

Thanks to your invaluable feedback, we’ve been able to track down the three main sources of frustration in post-editing. These are concentrated around wrong speaker segmentation, repeated poor recognition of proper names and specialised terms, and the consequent use of local language flavours.

Improved Speaker Segmentation

In case you need an accurate transcript of an interview, or if you want to use the transcription for creating subtitles, it is essential the speakers are properly segmented.

Proper automatic segmentation of speakers, sometimes referred to as ‘diarisation’, is sometimes a challenge. Specially in sitcoms or reality production, or in other cases where there are little or no pauses between the speakers.

Improved speaker segmentation – images by courtesy of De Vijver Media

The speaker diarisation has been extensively investigated in the MeMAD project, a joint R&D initiative funded by the Horizon 2020 programme of the European Commission in which Limecraft collaborated.

The results are stunning. In several cases where other software fails to detect any speaker changes, we correctly identified all. This identification is very beneficial for subtitle professionals, as it is best practice to avoid multiple speakers in the same subtitle as much as possible.

‘Custom Dictionaries’ to improve Recognition of Proper Names and Specialised Terms

Besides correcting speaker changes, statistics show that 60% of the time for post-editing is spent replacing words that are not properly recognised. This task is particularly frustrating in use cases where specific terminology or proper names should be properly recognised, like in sports or corporate video production using brand-specific terms.

To mitigate the problem of specific terms, proper names and homonyms in general, we now allow you the ability to manage a custom dictionary, i.e. a list of words to be recognised by preference. If you think custom dictionaries are helpful in your case, please reach out to our service desk. They will help you set up as it is still under investigation how we will make this available via the standard configuration and settings. 

Support for Language Locale

Besides supporting all common languages when it comes to transcribing audio or translating transcripts or subtitles, we now allow you to select the specific language locale.

Selecting Language Locale

Selecting the specific language flavour will impose the correct spelling during transcription and affect vocabulary and spelling when using machine translation.

Further improvement of Limecraft Flow

To improve the usability of Limecraft Flow, we reworked the item detail. There’s a lot of information to show. To make sure we can properly display the metadata, the process information, and the detail of every instance of the file, we’ve modified the way they are displayed.

Courtesy by De Mensen

Sneak Preview – What’s Up

In the coming weeks, we expect a couple of other significant improvements to be brought to life. These are already available in private beta. If you have any questions about them, feel free to reach out.

Supporting Multi-channel Audio

We’ve been managing multi-track audio files for a long time, but allowing you to play out the right audio tracks in the browser was challenging. This feature is now available, though it requires careful preparation of the proxies.

Supporting Multi-channel Audio Proxies

Enhanced Security through Multi-factor Authentication

With the increased demand for remote collaboration, we experienced more and more questions for increased security. Currently available in private beta and due to be released in production as part of the 21.2 release.

We hope you enjoy our services and to see you again soon on Limecraft Flow. In the meantime, stay tuned. As always, we greatly appreciate your suggestions and comments.

To learn more about collaborative video production and how to accelerate your workflow, subscribe to our monthly newsletter.

Subscribe
Receive a monthly update on new features, tips & tricks, use cases & workflows.