Podcasts are a great way to learn new enriching information, listen to interviews, and get deep insights into various topics - but for the normally abled only. Podcasts are not always inclusive, often leaving the hearing impaired out of the loop. This project is my take on making podcasts an inclusive resource for everyone via Google Podcasts.
PS. this is just a personal project.
Problems & Pain points
I tried to find people that were specially abled and consumed podcasts, but couldn't find any. Hence, I had to resort to the internet with online forums and reviews to figure out what issues currently prevail in the whole experience.
In Google podcasts,
- The hearing impaired cannot consume podcasts in any possible way.
- The visually impaired cannot scan through the podcast content with ease.
- Manually skipping only allows 1-sec skips.
- Jumping exactly 30sec forward loses the immediate context of the podcast.
I went online to get more context on podcasts and how accessibility is achieved in media streaming. I also tested out the interfaces for learning for how they implemented accessibility and what they could improve. Here are a few things I discovered:
W3 accessibility standards for podcasts
Audio-only (e.g., podcast):
- For pre-recorded:
- Transcripts are required at WCAG Level A.
- For live:
- Transcripts are at WCAG Level AAA. Usually, this needs to be a live text stream. If the audio follows a script, you can provide the text script.
Types of podcasts
Based on content, there are two types of podcasts:
- Narrative: Podcasts that are spoken in a narrative style, where the narrator gets in-depth and provides insightful information on their subjects.
- Conversational: Podcasts that have more than one person speaking, often involving 1:1 interviews and group discussions.
Transcripts are usually in different formats according to the type of podcast.
Researching through the platforms where people consume podcasts, I came across various implementations to improve the podcast experience. I'll cover them via the specific features they provide.
Many of the self-hosted podcasts provide a transcript of their episodes along with their episode release.
- Allows the hearing impaired to read through podcast contents.
- Allows scanning of podcast contents easily.
Closed captions - Youtube
Platforms like Youtube provide closed captions (sometimes auto-generated) for their content that provides the live text for what the host is saying.
- Allows the hearing impaired to see what's being spoken.
Chapters - Youtube
Youtube provides an option for creators to add chapters defining the notable sections in their videos.
- Allows easy scanning of media content via jumping to a specific topic.
- Chapters are keyboard-focusable so screen-readers can also access them.
- Usually, chapters are also mentioned in the description for easy access.
Transcripts with time-stamps
TED is a platform that provides transcripts with time-stamps. Since the content on TED talks is usually narrative, this proves to be a good implementation of transcripts.
- Time-stamps allow jumping through chunks of content.
- Time-stamps are keyboard-focusable hence they work well for the visually impaired.
I went ahead with using Transcripts because they proved to be better for scanning content and are more reliable than auto-generated text. The technology for captioning could be used to pre-generate transcripts for podcasts.
How do podcasts affect transcripts?
Since podcasts have different formats, having a single transcript format doesn't allow for effective consumption. For conversational podcasts, the emphasis should be on "who" is speaking. And for narrative podcasts, the emphasis should be on "what" is being spoken.
How to make transcripts scannable?
The most common form of facilitating scanning in a large area is providing a table of contents, eg. books. The table of contents lists out the important topics the book will cover and shows the page from where each specific topic starts. This helps in jumping to the specific topic you want to access.
Currently, podcasts are a single stream of audio, more often than not, providing no means to skip to sections of relevant content or scan through what the podcast covers.
Transcripts for Google podcasts
To facilitate the discovery of a new option, Transcripts were kept beside the most-used Play button. This keeps the Transcripts format on the same level as the Audio format.
Similar to the podcast media player, Transcripts open up in a bottom-sheet that can be scrolled. To enable easy jumping through the podcast contents by screen-readers, time-stamps are included before each sensible section.
Sections make the contents of the podcast more scannable for screen readers. Each section consists of a sensible chunk of information along with a time-stamp to allow screen readers to quickly jump to that specific section instead of scouring through the whole transcript for relevant information.
Transcripts for Conversational podcasts
Conversational podcasts focus on who is speaking, occasionally dividing the whole narration into acts or chapters that loosely change the scene or environment. Transcripts for conversational podcasts use sections for defining chapters (or acts) and the rest of the transcript follows according to who is speaking when.
Live text highlighting
The advent of transcripts allow the abled as well as the hard of hearing to follow the contents of the podcast along with listening to it. This helps them keep track where they are at that time and what section they are present in.
I feel like accessibility is considered to be a nice-to-have feature but instead it should be engrained in the concept of design itself. I do agree that new features that make the product accessible may often get less priority due to business needs but the narrative changes when they realise how much of the customer base need those features to use the product.
This was an insightful project and it led me to deep dive into the technologies used by the specially-abled folks to understand their patterns better. I hope I'll be able to make better solutions next time from the learnings I had from this project.
Thanks for reading ✌️