Distributed audio processing
Most likely audio transcription will be the most time-consuming part of the process, this is because it tends to be done more or less in real time by Apple’s servers.
This means, for example, that an audio file 10 minutes long, is processed in about 10-11 minutes.
However, when there are many audio files, or when they have a long duration, waiting for several tens of minutes or even hours could be unnerving and represent a bottleneck.
For this reason, we have implemented (at the moment in an experimental phase, but giving us excellent results) the possibility of transcribing audio in a shared way.
How does it work?
The principle is simple: divide the processing over several computers.
If a 10-minute file takes 10 minutes to process, more or less 10 computers can process this file in just over a minute.
Each application on each computer can thus become a processing node not only for its own files, but also for those of other people around the world.
Enabling distributed processing is very simple: just go to the preferences, in the audio control panel, and enable this check:
At this point the computer can use the computing resources made available by all other users who have enabled this option.
Ulti.Media servers are always available, so there are at least 4-6 computers available to process for you in the cloud.
Once this option is enabled whenever an audio file is processed, the work will be “split”, and each segment sent to the various computers available in the world for processing.
Keeping an eye on the processing status
By going to the Window > Show Shared Execution Status…
You can watch the segmentation process and see which computers in the world are collaborating on the process:
You can also name our computer by clicking the icon near “This Node” in the lower right corner:
