How to Test Speechmatics API Accuracy Using Upload & Transcribe Feature (Full Transcript)

Speaker 1: Hi there, I'm Paul and I'm Product Marketing Manager here at Speechmatics. In this short video you'll see how to quickly test the accuracy of the Speechmatics API using our Upload & Transcribe feature without needing to do any coding. First, you need to head over to Speechmatics.com and either sign up for a free account or log in to an existing account. Once you're in the portal, you're going to use the Upload & Transcribe tool found in the main menu on the left. Here you can select a media file in the most common formats to use as the source of spoken audio you want to transcribe into text. I'm going to click on the Upload button and select the file I want to transcribe. You'll see that this gets added ready to be uploaded and transcribed. You'll see that there are a few extra options to configure how you want the file to be processed. Language allows you to identify the language of the speech in the media file and how the transcription engine should process it. Separation tells the transcription engine whether you want the spoken audio to be separated out into different speakers, either by channel if the audio contains multiple channels, or by speaker, where the transcription engine will automatically identify different speakers in the audio. Lastly, accuracy allows you to choose between two transcription engines, Standard, which provides world-class accuracy out of the box, and Enhanced, which goes even further to improve accuracy and context awareness in your audio. For now, I'm going to leave the settings as they are by default and click this button to get the transcription.

Speaker 2: You can see the media file getting uploaded and the transcription process starting.

Speaker 1: Now you can see that the transcription has been processed and I have the text output displayed on the screen. You can copy this text ready to be pasted into a different application, or download the JSON or SRT files of the transcription for use in development or subtitle usage. Finally, you can look at this and other jobs that you have processed to download the transcription files again, instantly view the text output, or remove the file completely if it's no longer needed. That was a short introduction to testing the accuracy of Speechmatic's speech-to-text transcription without having to write a line of code. If you have any questions about the accuracy of your audio using our no-code transcription tool, please get in touch.