Visual Guide: Audio To Text
This post covers how to transcribe audio content. We will use a podcast as our sample audio content, and assume you are on a free trial. This post covers:
- Sign into the application.
- Create an audio specific template.
- Create a new item, with our audio specific template as parent.
- Upload our podcast, and then use an Artificial Intelligence (AI) to transcribe our content.
On reviewing this visual guide, you will be able to
upload your own audio content, and transcribe using an AI.
A quick summary of the process below is, to start with we create a new template
audioTemplate which we setup specifically for podcasts as a once off. Then, we will use this to create an audio item,
myFirstPodcast of type audioTemplate, for transcription.
Steps - Create An Audio Template
Please open your browser and go to videotranslator.ai, and then click on the
Loginbutton. Assuming you just registered for the free trial, the below is what you should see when you sign into the application.
Normally, we could just use
myTemplate. However, as this is a visual guide, we will make a new template, and do this the long winded way. Click on the
New Templatebutton, and use
audioTemplatefor the name as shown in the below image and
This is a new template, and we are going to add an
Audiocomponent. Click the
Add Fieldsbutton, then select
Audio, as shown below.
expand the audio component using the highlighted + button. The other highlight is what the audio component will look like once we create a new item.
Use the action bar to
Publishthe template. This makes the template usable to create new items. We have also changed the
Please upload the podcast?, this change is purely cosmetic.
Exit the template editor. Your
Rootwill now look like the below image. The green is a visual indicator indicating that this template was just created. You have successfully created your ‘audioTemplate’.
Why am I doing this?
template editor provides you the flexibility to tailor your template to your specific use case. Specifically, in addition to the ‘audio component’ we added into our template, you could add:
Image Fields: Assuming you were uploading your content into a third party audio sharing service, you can
add a poster or banner image, which should boost engagement.
Text Fields: Assuming you were uploading your content into a third party audio sharing service, you can add a title, description and other metadata as additional text fields. Uploading these
increases SEO substantially, and this effect is magnified if we were to also translate the audio content.
Should I do this every time?
No - absolutely not. The entire idea of a template is to
create a data structure which meets your specific use case and reuse it every time.
Steps - Create An Audio Item, and Transcribe
audioTemplate, and then click
Add New Itemto create an item of type
audioTemplateas shown below.
Today, we will use a podcast from the Creative Commons team, known as Plays Well With Others. We are big fans of the work Creative Commons does, and you should check them out if you are not familiar with the contribution of CC to modern open source software.
*.mp3, and you should see something like the below image.
Now, trigger the transcription. We used
American Englishin this case. Accept to start the process. The item will exit, and lock itself until the AI completes the transcription process.
You are done! Click on the Captions to see the results of the AI. We recommend checking/editing to make sure the AI has transcribed your podcast properly.
Remember, check grammar and capitalisation. In the below, 'plays well with others' is not capitalised. The AI did not pick this up, because it was spoken normally, not especially pronounced. People do not speak like the rules of English require us to write unfortunately.
In this visual guide, we used the application to transcribe an English podcast with AI assistance.
We are very grateful for your support!