"You are responsible for the accessibility of your own content. You can't wait around for somebody else to fix the problem for you."
Joe Clark, 27 November 2004, commenting on "Google is a deaf user", BestKungFu Weblog
Podcasts are inaccessible to members of the deaf community.
Linking to a transcript of your podcast will make it much more accessible.
Peter Batchelor has created an
example of good podcasting
practice. It combines a video and a transcript into a
single, captioned presentation. Thanks to
NASA for the video and the transcript.
Podcasts are an increasingly popular way to present content on the Web. Because they are audio files, they generally only work for people who can hear. In addition, they are often quite large files, which can present accessibility problems for people with slower connections to the Internet. Podcasts that include video, often referred to as vodcasts, generally have larger file sizes than a podcast of the same duration, so the speed of the user's Internet access is even more crucial.
Podcasters are aware of this problem, and would like to find solutions.
"I've had a couple of requests to offer the podcast in an accessible format for hard of hearing and deaf persons. Specifically, I've been asked for transcripts of the show. Short of sitting down and writing them up I'm unsure of how to do this." Larry, 14 June 2006, Podcast Accessibility for Deaf and Hard of Hearing, Disability Nation
One way to make your podcast accessible is to provide a written transcript.
As usual, making accessible podcasts turns out to have benefits for a bunch of other reasons, too. It will make it easier for search engines to discover the information, it will make it easier for people to quote you, and it will help people who can't play audio (because the baby is asleep, for example).
Please note that we assume that you know how to make a podcast. We do not provide any advice on creating or publishing your podcast. Also, we are not expert podcasters ('novice' would be a better description). So if you have other suggestions, please let us know
A podcast is an audio file that people can download from a Web site. Generally, it is wrapped in some additional RSS (Really Simple Syndication or Rich Site Summary) information, so that people can easily discover your podcast and subscribe to it and other podcasts you might make. Podcasts are often thought of as files that are downloaded to be played on personal audio devices such as iPods, but they are often used on conventional computers as well.
Sometimes, podcasts include still images or video.
To meet accessibility requirements, you should provide a textual equivalent such as a transcript. This is what NASA do with their podcasts. Each podcast is accompanied by a transcript. They point to the transcript from the RSS information.
Ideally, and particularly if you were creating a vodcast, you could also provide closed captioning, preferably using Synchronized Multimedia Integration Language (SMIL) or perhaps Microsoft's Synchronized Accessible Media Interchange (SAMI). These standards allow text, images, audio and video to appear on a Web site in a structured, timed way, and only if the user requires it. See the Web Accessibility in Mind thread on Accessible podcasts for suggestions. Some vodcasters include open (also referred to as burned in) captions - text as part of the video file, which avoids the need to use SMIL, but this side-steps the idea of device independence. If the video is displayed on a small screen - such as an iPod, the text may be too small to be of much use, and there may be other viewers who would prefer to not see the text.
As with most things, it is easier to make a podcast accessible if you build accessibility in from the start.
Write a script before you begin.
This is not just a good idea from an accessibility point of view, it is a good idea from a podcasting point of view.
"If you are interviewing, prepare the questions ahead of time. ... If you aren't interviewing, take the time to prepare exactly what you'll be talking about. Write down an agenda with talking point notes." Carson McComas, 2 February 2006, 20 Ideas for a Great Podcast, Frogbody
If you plan to include images in your podcast, include a description of them in your script.
If you plan to include video in your podcast, you might want to caption it. Software like Magpie will help you to create captions for your video.
The more planning that you can do, the better your resulting podcast will be. And you will end up with text that will help make your podcast more accessible.
Put your script on the Web.
You could put it up as a Web page, or as a blog entry. Even a text file would be more accessible than nothing at all. The main thing is that you put it somewhere that has a permanent Web address.
Link to the Web page from your podcast.
In particular, link to the Web page from the RSS information that you include with your podcast. To do this, add the address of your Web page in the <link> information that you include in your RSS information.
If your blog supports podcasts, it will probably help you to create the RSS feed. If it doesn't include a link, or if you are uploading the podcast to your own server, you might need to create your own RSS. If you are not sure how to create RSS for your podcast, you can use several different services to help you.
"But what about the thousands of podcasts that I have already made?", I hear you ask. If you already have made your podcast, you should try to provide a transcript to accompany it.
There are three possible approaches to creating a transcript. Speech recognition and transcription services cost money, while transcribing the audio yourself costs time.
Convert the audio to text.
If you are the only voice on the podcast, you can use speech recognition software like Dragon Naturally Speaking to convert the audio to text. Speech recognition software generally needs to be trained to recognise a voice, so it will work best if there is only one voice on the podcast and you can train the software to recognise that voice.
To use Dragon Naturally Speaking to convert your podcast to text, you will probably need to convert your MP3 audio file into WAV format, probably 16 bit mono. A Web search for "MP3 to WAV converter" will turn up a number of different converters.
Please note that you probably need to carefully check your text for errors. While speech recognition software can do remarkable things, it can also make some horrible mistakes.
Dragon Naturally Speaking costs approximately $A200 - $A300, depending on the version you buy. We do not know of any free (or even cheap) speech recognition software that will import an audio file and output a text file. If anyone has any suggestions, please let us know.
Pay a transcription service.
A transcription service will transcribe your podcast for a fee. This is particularly handy if you are interviewing people and so cannot create a script.
Casting Words, for example, charges $US0.42 per minute to transcribe podcasts. All transcribing is done by people, not machines. Transcriptions are delivered in plain text, HTML and RTF formats. You get an RSS feed of all of your transcripts.
Transcribe the audio yourself.
If you are going to transcribe the audio yourself, you might want to use an audio player that allows you to match the speed of playback to your typing speed. This will make things much easier.
Express Scribe is a free transcription playback program that will play most audio file formats at a variable rate of playback.
If you are doing a lot of transcription, you might wish to invest in a pedal control, too. This will allow you to control playback without using your hands (which will be busy typing).
Often, you will want to make someone else's material accessible, so that you can have access to it.
Ask them to make it accessible.
Most people are thrilled to have an audience. They may not realise that their material is inaccessible, or that it is their responsibility to make it accessible.
Ask another listener to transcribe it, or summarise it for you.
Use one of the transcription services, as described above.
Discoverability
Search engines will pick up your transcript, so more people will find it. Also, the transcript gives people something tangible to link to. This will increase your ranking with search engines, too.
Audience preference
When asked, students using podcasts say that they prefer both a podcast and a text version. The podcast makes it easier to hear a lecture when they are travelling, jogging, etc. But they prefer to use the transcript for revision and for quoting in essays. You can't highlight a sentence in a podcast.
Also, some sorts of material, such as step by step instructions, are easier to use when they are read. Imagine trying to cook a recipe via podcast, for example.
Bandwidth
A one hour podcast (of a university lecture, for example) might be over 10 Mbytes. Even 10 minute podcasts can by 5-6 Mbytes. The accompanying transcript will probably be about will be 200 Kbytes.
Bandwidth can be a major problem for people in rural areas. In Australia, about 50% of people outside of the major metropolitan areas are still on dial-up modems. In addition, electric fences on farms will often disrupt the telephone signal, making large downloads almost impossible.
No audio
There are a range of environments where a podcast is not usable. Assistive technology may not have speakers attached. Computers in libraries also may not have speakers and may not allow people to plug-in external devices (like headphones).
If it is done right, we think that there might be a working business model in this idea. With the agreement of the podcaster, you could transcribe popular podcasts and host the transcriptions on your site. Your revenue would either come from subscription fees for the transcripts or advertising revenue from advertisements embedded in the Web pages.
The podcaster can point to your transcription from their RSS information. Other people would discover the pages via a Web search.
This means that the podcaster would now have transcripts of all podcasts, and would be satisfying accessibility laws in their area. A lot of people would love you to transcribe their podcast for them, but do not want to do it themselves.
When we were researching this article, Peter and I came across two other ideas that are tangental to the main topic, but still worth talking about.
Turn Blog entries into podcasts
Audiolicious is Microsoft Windows software that uses text-to-speech software to turn any RSS feed (like a blog entry) into MP3 files.
Pod catching software for screen readers
Accessible Podcatcher is Microsoft Windows software that provides a screen-reader friendly way to grab and listen to podcasts.