Read Aloud the Text Content

This audio was created by Woord's Text to Speech service by content creators from all around the world.

Text Content or SSML code:

Jobs with multiple speakers come equipped with a featured called “Load Speaker Labels”. This feature will automatically detect who is speaking to help assist in transcription so that you’re not reliant on being able to distinguish voices by ear. Let’s take a look at how it works. When you’re assigned a multi speaker job, you will see a feature in the top bar of your word screen that says “Load Speaker Labels”. When you click this button, the audio will be process through a system that will identify unique speakers as they appear in the audio stream. This can take a few minutes (depending on the length of the job) and you will see the button display “loading” while it is processing. But you can still stream the audio and begin typing while it is processing so you’re not waiting. When the job is done processing, you will see a pop up that says “Speaker labels loaded”. The next time you click “play” in the audio stream, you will see an indicator pop up near the top of the screen that distinguishes that speaker. The system will name the speakers “Speaker A, Speaker B, Speaker C”….and so on… by default. But if the names of the speakers are given during the audio, you can go back through and rename the speakers so that you can properly label them for transcription. In this example, the first speaker is labeled Speaker A by default. But in the audio, the speaker identifies himself a Jesse Suttner. So now that we know that Speaker A is Jesse Suttner, we can click “Rename Speaker”. In the first box we input the default name for that speaker (in this case it is Speaker A). And in the second box we input their actual name (in this case Jesse Suttner). Now, when we once again hit play to stream the audio, his name will appear each time he speaks in the audio stream. As we continue listening, a second speaker is featured in the recording. As she begins speaking, the speaker is automatically identified as Speaker B by default. But as we continue listening, the speaker provides her name as Marie Blasgin. Again, we just click on the “Rename Speaker” button, and indicate that Speaker B is Marie Blasgin. When you play the audio again, it will now display each of the speaker names as they speak. You will be able to add this to multiple speakers in the audio. The quality of detection of unique speakers may begin to deteriorate after 8 or 9 speakers, but we rare get audios with that many speakers, so it is unlikely to become an issue. This tool will help you to be able to identify speakers in multi speaker work and is especially useful for long recordings, partial jobs, and jobs that feature a lot of speakers, or voices that sound similar. While this tool is typically very accurate, it is not perfect. If the program indicates one thing, but you- as a person listening to the audio- think that a different person is speaking- your judgement overrides the software. This program uses voice frequencies and probabilities to determine which voice is speaking. But if two voices are similar, for example, the program can be thrown off. So if you have a scenario where the program is indicating that an officer if confessing a crime to a suspect- you can probably correctly deduce that the program has mis-identified the two speakers. This program has been tested across a multitude of recordings and we selected the one that we are most confident provided the best identification. But always use your best judgement as the final call in labeling speakers. Partials If you get a partially completed multi speaker job that the previous typists has loaded speaker labels for, those labels will automatically populate when you pull up the job. To use our earlier example, if a second typists were to get the partial job of the one we started before, they would automatically have Jesse Suttner and Marie Blasgin identified as they stream the audio. If a third speaker were to appear later in the audio, the second typist can add in the name for Speaker C if it is indicated.