- "Media" means any content which features video-only, audio-only, or video and audio together.
- "Vendor" means the supplier with whom we have made a purchase for deliverables which include media.
- "Customer" means [COMPANY NAME] including any and all sub-organizations [COMPANY NAME]
- "Deliverable" means any media produced as a result of a project that is intended to be delivered to the customer.
- "Captions" means text displayed on screen when video or audio-video media is playing, intended to provide interpretive or additional information.
- "Subtitles" means captions used to translate the content in another language. Subtitles are not generally direct translation of captions as they do not include actions, sounds, etc.
- "Transcript" means a written version of the audio portion of Media.
- "Audio Description" means a descriptive narration of all the visual elements of a video. Audio description is written and recorded so that it falls in the gaps between dialogue (and other important audio elements).
This addendum describes the requirements for acceptance of any deliverables which feature media. Deliverables which do not meet these requirements will not be accepted and invoices submitted for those deliverables will not be paid until these requirements have been met in their entirety and acknowledged as accepted by the customer. Vendor understands that this addendum is part of the overall contract of services purchased and that any changes required to deliverables so that they conform to these requirements will not incur any extra charges beyond that which is defined in [Reference to purchase price elsewhere in contract]
This addendum is written in conformance with RFC 2119, which defines keywords used to indicate conformance. This addendum follows the "Must/Shall", "Must Not/Shall Not", "Should/May", and "Should Not" convention of defining requirements at the following levels:
- "Must" and "Shall" indicate requirements that deliverables are required to meet in order to be accepted by the customer.
- "Must not" and "Shall not" indicate characteristics that will be deemed to have violated these requirements and the deliverable will be rejected by Customer
- "Should" and "May" indicate optional characteristics and hints for a higher quality deliverable. These characteristics are not alternatives to Customer's requirements but simply suggestions.
- "Should Not" indicates characteristics that will erode quality of the Deliverable and avoiding such characteristics is advisable.
This addendum applies to all media purchased by Customer and delivered by Vendor. This includes any and all media content purchased directly or supplied as part of training/ support materials.
Captions are intended to serve as an alternative to audio content and as such must convey the same meaning as the audio. This means that captions must cover more than just verbal dialog but anything conveyed through sound that is conveys meaning or context to the listener.
- All audio-visual media content must have captions.
- Captions must be accurate with accurate, country specific spelling, e.g. US-English if the media was produced in the United States.
- Captions must be accurately synchronized to the content. In other words, as sounds are made or words are spoken, the captions must be presented at the same time
- Captions must not correct poor grammar used by speakers.
- Captions must not deviate from or editorialize the words used by speakers.
- Captions must be presented in high-contrast against the video content. White text on a black background is advised. Adherence to WCAG 2.0 Level AAA contrast ratios is required.
- Captions must identify each speaker each time the speaker changes. The identification of the speaker must be consistent throughout the content.
- Captions must identify meaningful non-spoken sounds including background sound. Care should be taken to avoid making the captions for non-spoken content too verbose, concentrating on conciseness.
- Captions must not block other meaningful text content on screen, e.g. Speaker's job title when displayed.
- Captions must not be displayed on more than 2 lines.
- Captions length, while displayed, should have a maximum length of 68% of the width of a 16:9 video and 90% of the width of a 4:3 video. The number of characters this represent depends on the font type and size used to display captions.
- Captions should use UPPER CASE for speech, and lower case for actions or identifying speakers, e.g. Man: I THINK THE DOG IS UPSET.
- Captions should include actions or sound effects when they add meaning within parenthesis or brackets, e.g. (dog barking)
- Captions should utilize punctuation to clarify meaning and intent on the part of the speaker(s)
- When a sentence must be divided into two or more captions, it should break at a logical phrase rather than at a random point
Subtitles are captions displayed at the bottom of video content that translates the dialog.
- Subtitles are not required
- If subtitles are displayed, the following apply:
- Subtitles must be accurate with accurate, country specific spelling, e.g. US-English if the media was produced in the United States.
- Subtitles must be accurately synchronized to the content.
- Subtitles must not correct poor grammar used by speakers.
- Subtitles must not deviate from or editorialize the words used by speakers unless such changes are warranted to accurately reflect the intent of the speaker to ensure such word choice is accurate in context of the language that the subtitles are translated to.
- Subtitles must be presented in high-contrast against the video content. White text on a black background is advised. Adherence to WCAG 2.0 Level AAA contrast ratios is required.
- Subtitles must identify each speaker each time the speaker changes. This identification must be consistent throughout the content.
- Subtitles must not be displayed on more than 2 lines
- Subtitles length, while displayed, should have a maximum length of 68% of the width of a 16:9 video and 90% of the width of a 4:3 video. The number of characters this represent depends on the font type and size used to display subtitles.
- Subtitles do not generally include actions, sounds/music, and other non-verbal cues.
- When a sentence must be divided into two or more captions, it should break at a logical phrase rather than at a random point
A transcript is a separate text document that contains the information conveyed through sound.
- All audio-only, video-only and audio-visual media content must have a transcript.
- Transcripts must be accurate with accurate, country specific spelling, e.g. US-English if the media was produced in the United States.
- Transcripts must not correct poor grammar used by speakers.
- Transcripts must not deviate from or editorialize the words used by speakers.
- Transcripts may skip filler words, such as "ah", "huh", and other similar words or sounds.
- Transcripts may skip false starts, e.g. "St... St... St... Start".
- Transcripts must identify each speaker each time the speaker changes. This identification must be consistent throughout the content.
- Transcripts must identify meaningful non-spoken sounds including background sound. Care should be taken to avoid making the transcripts for non-spoken content too verbose, concentrating on conciseness.
- Transcripts must identify meaningful visual content displayed on audio-visual or video-only content. This includes meaningful motion and must include detailed step-by-step descriptions for any and all training content.
Audio description, also referred to as a video description, described video, or more precisely called a visual description, is an additional narration track intended primarily for blind and visually impaired consumers of visual media.
- All video-only and audio-visual media content must have audio description.
- Due to the technical challenges of accurately delivering described content, the vendor may supply a separate deliverable that features audio description.
- Audio descriptions will be time constrained by the soundtrack.
- Audio descriptions volume must be 20% higher than the volume of primary audio content in the deliverable.
- Audio descriptions must accurately describe any relevant visual content.
- Audio descriptions should coincide with the action being described.
- Audio descriptions should be written in present tense.
- Audio descriptions must identify new speakers.
- Audio descriptions should use a vocal style compatible with the target audience of the media, e.g. children programming requires a different style than scientific programming aimed at adults.
- Audio descriptions should include all information important to the plot, e.g. clothing, location, facial expressions, settings, etc.
- Audio descriptions should indicate the passage of time between scenes.
- Audio descriptions should be neutral and avoid interpretative descriptions.
- Audio descriptions should describe emotional states without interpreting them. It is better to describe facial gesture rather than guess at emotion.
- Audio descriptions should use complete sentences whenever possible.
- Audio descriptions should be delivered during pauses in dialog.
- Audio descriptions should not fill all gaps in the audio track. Music and silences can tell the story as well.
- Audio descriptions must read opening and end credits, if present.
The following resources can be used to help understand how to create captions and audio descriptions. These resources are informative in nature. In any instance where the guidance in the following resources disagree with anything above, our requirements prevail.
- Media Access Group
- American Council of the Blind Audio Description Project
- BBC Subtitle Guidelines (Note the word used is 'subtitle', but that is equivalent to 'caption' in this case).