Skip to main content
Version: v0.3.0

Tutorial 2C: Annotation task: Audio regression

Overview​

This tutorial underlines the steps (process) of annotating and specifying an annotation task rubric for an audio regression annotation task. To highlight the process, we are going to annotate a dataset containing 600 audio samples of spoken digits (0-9) of sixty different speakers. This tutorial also quickly explores how you can download the fully annotated dataset supported in H2O Hydrogen Torch.

Step 1: Explore dataset​

We are going to use the preloaded amnist-demo demo dataset for this tutorial. The dataset contains 600 samples (audio) of spoken digits (0-9) of sixty different speakers. Let's quickly explore the dataset.

  1. On the H2O Label Genie navigation menu, click Datasets.
  2. In the datasets table, click amnist-demo.

Dataset details of amnist-demo dataset audio files

Step 2: Create an annotation task​

Now that we have seen the dataset let's create an annotation task that enables you to annotate the dataset. An annotation task refers to the process of labeling data. For this tutorial, an audio regression annotation task refers to assigning one continuous target label to each input audio. Let's create an annotation task.

  1. Click New annotation task.
  2. In the Task name box, enter tutorial-2c.
  3. In the Task description box, enter Annotate a dataset containing samples of spoken digits (0-9) of sixty different speakers.
  4. In the Select task list, select Regression.
  5. Click Create task.

Step 3: Specify annotation task rubric​

Before we can start annotating our dataset, we need to specify an annotation task rubric. An annotation task rubric refers to the labels (for example, object classes) you want to use when annotating your dataset. For our dataset, let's label each audio clip with a value from 0 to 9, where 0 refers to number 0, 1 refers to number 1, etc.

  1. In the Data minimum value box, enter 0.
    • The Data minimum value value refers to the minimum value in your continuous values (in this case, digits ranging from 0 to 9)
  2. In the Data maximum value box, enter 9.
    • The Data maximum value value refers to the maximum value in your continuous values (in this case, digits ranging from 0 to 9)
  3. In the Data step size (interval) box, enter 1.
    • The Data step size (interval) value refers to the value the label range slider interval takes (the slider its utilize in step 4)
  4. Click Apply

Annotated audio dataset with slider ranging from 1 to 9

Let's utilize the slider, not the picker, to annotate the samples. To enable the slider, consider the following instructions:

  1. In the Annotation selection list, select Slider.

Slider

Step 4: Annotate dataset​

Now that we have specified the annotation task rubric, let's annotate the dataset.

  1. Click Continue to annotate.

In the Annotate tab, you can individually annotate each audio clip in the dataset. Let's annotate the first audio.

  1. In the Label slider, slide to the label you associate with the audio sound.
  2. Click Save and next.
    Note
    • Save and next saves the annotated audio
    • To skip an audio clip to annotate later: Click Skip.
      • Skipped audio clips (samples) reappear after all non-skipped audio clips are annotated
    • To download all annotated samples so far, consider the following instructions:
      1. Click the Export tab.
      2. In the Export approved samples list, select Download ZIP.
        note

        H2O Label Genie downloads a zip file containing the annotated dataset in a format that is supported in H2O Hydrogen Torch. To learn more, see Downloaded dataset formats: Audio regression.

Download annotated dataset​

After annotating all the sound clips, you can download the dataset in a format that H2O Hydrogen Torch supports. Let's download the annotated dataset.

  1. In the Annotate tab, click Download approved samples. Export approved samples
  2. In the Export approved samples list, select Download ZIP.

Summary​

In this tutorial, we learned the process of annotating and specifying an annotation task rubric for an audio regression task. We also learned how to download a fully annotated dataset supported in H2O Hydrogen Torch.

Next​

To learn the process of annotating and specifying an annotation task rubric for other various annotation tasks in computer vision (CV), natural language processing (NLP), and audio, see Tutorials.

Feedback