Audio and Video Best Practices

Audio description help individuals with low vision or blindness.  These descriptions provide essential information that these users can't otherwise access or understand.

At the same time, audio description can also help users listening to content, including:

• Users looking away from screens to write notes while listening
• Students seated in the back of lecture halls
• Commuters who can't look at visual information but who can listen to it being read aloud
• Learners who benefit from seeing and hearing at the same time
• Content creators, who wish to highlight and explain visual aspects of a video that might otherwise be missed or misunderstood

Please consider the following guidelines as you create audio descriptions:

• Record videos with narration included, so adding additional audio description isn't necessary.
• If you are recording a lecture video, describe what you're doing or seeing as you lecture.
• Describe the scene as the action unfolds. 
• Do not speak over the dialogue unless it is necessary. 
• Speak in complete sentences, and use clear and concise language.
• Match your tone, style, and pace to the video you are describing; it should feel cohesive. 
• Do not use vague terminology such as "he" or "she", "this" or "that," and "here" or "there."

If you are unable to provide audio description, one possible alternative is creating a text transcript.  This transcript should provide audio and visual information.  By accessing these types of transcripts, users with both hearing and visual impairments should be able to understand the full content of the video. 

Some possible ways to create this alternative transcript are as follows:

• Start with a transcript based on available captions.  If your video doesn't have captions, request them and edit them if necessary.  Many video platforms will also have a transcript available.  If that's the case, you can use that instead.
• Add a narrative of visual elements to the existing captions/transcript.
• Read the transcript independently of the video.  If it makes sense, then it's done.  If not, go back and revise the transcript until it makes sense with referencing the video.
• Provide the transcript to users in an accessible format, such as Word or PDF file.

Captions provide an alternative to audio that helps all types of users, including:

• Deaf or hard of hearing
• English language learners
• Learning disabilities
• Visual or multi-modal learners

Additionally, captions or subtitles also help users who are:

• Watching videos in noisy environment
• Watching videos in quiet environment
• Scrolling on social media
• Deciphering unfamiliar accents

• Imagine that you can’t hear and include whatever sounds you need to understand the meaning of the video.
• Include all sounds, including stutters, pauses, etc. So, if a speaker pauses and makes a sound, you can use an “erm,” or a “hmmm” in the caption text.
• Include background sounds whenever necessary to understand the meaning of the video.
• Punctuation matters. For example, the following two sentences have the same words but wildly different meanings.\
  • “let’s eat grandma!”
  • “Let’s eat, Grandma!”
• Whenever possible, synchronize the caption with the action on the screen.
• Watch your color contrast. If a white caption appears on a white or light-colored section of the video background, it will be difficult, if not impossible, to read.

Captions benefit a wide variety of users, but they can only do so if they are accurate. Many video platforms, including YuJa, provide automated captioning. While automatic captions can help cut down the time needed to provide accurate captioning, be aware that auto-captioning relies on machine-generated algorithms that are never 100% accurate. For example, videos captioned in YuJa are supposed to be 90% accurate, but they often fall considerably short of this because of variances in recording environment, accents, and other complicating factors.

Auto-captions must be accurate to be usable. Oftentimes, that requires manual editing.

Audio Transcripts

A transcript is the only method of providing a text alternative for audio-only files. This can either be a downloadable document or text can be displayed alongside the audio file.

When audio files are embedded in another document, such as a PowerPoint file, a separate transcript is still needed.

Video Transcripts

Transcripts differ from captions in that the text is not displayed in sync with the video. Oftentimes, a transcript is viewed separately from its video file. Given this separation, transcripts need to include logical formatting (including headings). Transcripts also need to make sense on their own.

To provide an equitable experience of videos, synchronized captions are required; on their own, transcripts are not usually sufficient. However, there are cases where transcripts may be preferred or required. Having both available is recommended.

Transcripts can benefit deaf or hard-of-hearing users.

They can also support many other learning strategies, including:

• Locating a specific point in a video
• Note-taking
• Learning new vocabulary
• Learning a new language
• Focusing
• Reviewing
• Searching for terms

Consider the following guidelines when generating transcripts for audio and video files:

• Textually describe the scene.
• Include relevant, non-speech sounds (e.g., laughter, applause, or background noise), as these elements are essential for a complete understanding of the content.
• Decide if you want a verbatim or edited transcript
  • Verbatim Transcript: includes all sounds, including stutters, pauses, and filler words (e.g., “hmmm” or “erm”).
  • Edited Transcript: cleaned-up versions prioritizing grammar and punctuation, while maintaining the original context.
• Add labels or names to differentiate between multiple speakers. This organizes the transcript, making it easier to follow.
• Review your completed transcript for accuracy and correct any errors.
• Break the content down into defined sections and topics to make it easier for readers to navigate and comprehend.
• Link the transcript file to its accompanying audio or video.

Automatically generated transcripts rely on speech recognition software. The accuracy of transcription may vary based on several factors: program software, audio quality, speaker dialect, and content complexity.  To ensure accuracy, some post-editing may be required.