Transcription
Can you also deliver transcriptions for other media formats?
We deliver data annotation for speech-to-text solutions. However, if you have a special request, please contact our sales team here.
Do you deliver timestamps?
Yes, using our software and different tools, our annotators work hard to make sure the timestamps are accurate and any utterences are mitigated.
In which file formats can the transcripts be delivered?
For data annotation, the recommended formats are JSON and XML, which we can both provide. However, should you wish to work with another format, our software also supports Text, Word, SRT, EBU-STL and VTT files.
Which kind of specifications do you use for data annotation?
Depending on your needs, we can provide different acoustic models or different linguistic models. To find out more about this, please contact our sales team here.
How do you ensure the confidentiality of personal data?
Amberscript’s IT infrastructure is built on the server infrastructure of Amazon Web Services located in Frankfurt, Germany. All data that is processed by Amberscript will be stored and processed on highly secured servers with regular back-ups on the same infrastructure.
Which languages are supported?
We support many languages for our data annotation projects, please reach out here to learn more.
How do you ensure timely delivery of results?
Should you wish to make use of our data annotation services, we will assign a project planner to your project, who will be in close contact to discuss the details and timeline.
How do you ensure high quality?
We work with a vast network of professional annotators, who will be trained to your annotation guidelines. All annotations go through rigorous quality checks using our sophisticated data annotation AI.
How does data annotation work?
Data Annotation is the process of labeling data, which could be in various forms such as images, video, audio or text. Basically data annotation is done using various tools like bounding, semantic segmentation etc. Data labeling is usually done to train various computer models.