Published in · 5 min read · Jan 30, 2019
--
It’s an introduction familiar to many: “From The New York Times, I’m Michael Barbaro. This is ‘The Daily.’” But a large number of Times readers have never heard Barbaro’s greeting. That is because “The Daily” has only been available in audio form, which means its report is inaccessible to many of our readers. What makes podcasts so special, also renders them inconvenient: they can only be heard.
Over the past few months, we’ve experimented with ways to make our audio content more accessible to readers. We’re releasing audio transcripts for each new episode of “The Daily,” with transcripts rolling out for other Times podcasts in the future. It’s important to us that all of our readers can engage with our report, whether they choose to listen, share or read, and audio transcripts are a big step forward in our commitment to make Times content accessible to all users.
What are audio transcripts?
An audio transcript is a word-for-word text document of an audio file. Each line of the transcript is printed alongside its corresponding timestamp, making it easy to navigate the audio file. Reporters often use transcripts to find and reference quotes from interviews, but transcripts can be useful to readers, too. The script-like layout is great for understanding what is mentioned in an audio clip, finding a line that covers a particular subject or copying a salient quote to share on social media.
The process of transforming audio into text can be tedious. While we could have a Times producer listen to every episode and type out what they hear, that would be too time consuming. For podcasts that publish frequently, like “The Daily,” doing our own transcription wasn’t a feasible option for us. We instead opted to use 3Play, a third-party service, to transcribe the audio. Being able to hand off the task of the transcription is key in a workflow that is already packed and very fast-paced.
The document that 3Play produces not only captures the words of the show, but also includes labels that indicate when different people are speaking. Before the transcript is published, a Times editor reads over the document to review its accuracy and replace the generic labels with the names of those speaking.
Although we could have simply pasted the full transcript on the page, we decided to introduce more interactivity into our transcript pages. We took feedback from Times readers and listeners of “The Daily,” and we designed for accessibility, shareability and reference.
The design process
When we approached this design project, we knew that transcripts needed to be easily findable, and we knew that they had to be closely tethered to the original audio piece. We decided to build the feature into our podcast pages and our embeddable audio player, but what that actually looked like took some experimentation.
The audio player on “The Daily’s” podcast pages is often followed by supplemental content related to the episode. With that in mind, we wanted to give readers the option of toggling between the transcript and the supplemental content, without sending them to a new page.
Some of our initial designs included text that appeared, karaoke-style, as an overlay on the audio player. While this gave weight to what was being said in the audio at a given moment, it didn’t allow for searching through the episode. Much like closed captions on video, how quickly text appeared was dependent on how fast words were said in the audio. Since many people often process audible words faster than they read printed words, we didn’t see this as an optimal experience for our users.
A reading experience that was so tied to the audio was problematic because it prioritized audio in a way that wasn’t great for our readers who are hard of hearing.
This led to us explore other options, like the one below. In this version, we considered revealing more text in an overlay on the audio player. As the audio advanced, the corresponding line would be highlighted. Though this exposed more of the text, it was limiting and still too closely tied to the audio. Anyone who wanted to search through the transcript would have to navigate through the audio.
Our initial designs presented the transcripts tightly bound to the audio, meaning readers would have to engage with the player to read the text. We wanted readers to be able to skim the transcript without playing the audio, but still be able to control the audio if they chose to listen and read at the same time.
Creating a page with its own URL was important for shareability reasons, so we opted for an overlay with a dedicated URL that could be accessed from the podcast page. As a nice side effect of dedicated pages, the content of Daily episodes now surfaces in search results. Let’s say, for example, that “The Daily” aired an episode about climate change, and in that episode coral reefs were discussed. With transcripts attached to the metadata of an episode, anyone searching ‘coral reefs’ can discover that episode of “The Daily.”
What’s next
In future releases, we are planning to build in functionality that allows readers to click on a line in the transcript and be taken to that segment of the audio. We are also exploring ways that our readers might share specific clips of the podcast on social media. Transcripts will also be rolling out in the coming weeks for “The Argument” and for the next season of “Still Processing.” But for now, with this launch of audio transcripts for “The Daily,” we’re excited to provide an additional way for readers to experience The Times’s journalism.
As with all of our products, we are receptive to feedback from our readers and listeners and are eager to make improvements that would better the overall experience.
