When Blind Professionals Make Audio Description, Everyone Sees a Better Story

From editing rooms to training programs to Netflix’s global workflows, blind and low vision experts are shaping how silent visuals are described in the US and the UK.

In 2021, Nefertiti Matos Olivares, a New York–based blind leader in audio description whose work spans writing, narration, quality control, consulting, advocacy, and training, was hired to write the audio description for the documentary Say His Name: Five Days for George Floyd.

“It was an all-blind cast,” says Matos Olivares, “with a blind writer, quality control specialist, narrator, mixer, and producer.”

To write the script, Matos Olivares enlisted her sighted sister’s help not to make editorial decisions but to say what she saw on screen.

They watched the documentary together three times.

After an initial viewing to take in the content, Matos Olivares carefully noted time gaps and jotted down clarifying questions as her sister paid closer attention to the visuals. On the final viewing, her sister answered Nefertiti’s questions and either corrected or added specific details to what she surmised was happening based on context and audio cues.

“How many officers are holding George down?”

“Three.”

“Can you see that George is black and the officers are white?”

“Yes, they are.”

“Anything else stand out to you?”

“There’s a car in the background and some trash in the upper left of the frame.”

At this level of clarification, Matos Olivares decided what was and wasn’t relevant and tonally significant to the scene, and crafted her script accordingly.

Matos Olivares’s process challenges a long-held assumption that blind professionals – who are the real consumers of audio descriptions – are not suited to creating and quality-controlling them.

But blind professionals tend to approach description differently from their sighted counterparts, drawing on their lived experience to determine what information blind audiences need and how best to convey it in ways that are meaningful and make sense. “We are the ones who most benefit from audio description,” Matos Olivares says. “Who would be better than us to know what works for us?”

Matos Olivares has built a career in blind-led audio description over the past five years and is part of a small but growing community of professionals working to establish blind-led writing, narration, and quality control as standard practice in the field.

A craft built around sighted expertise treats accessibility as an afterthought.

Audio description (AD) is the narration added during pauses in dialogue to describe visual details for blind and low-vision audiences. According to the American Council of the Blind, AD has been part of film, television, theatre, and museums since the late 1970s.

Unsurprisingly, audio descriptions around the world are largely written by sighted professionals. For many in the industry, the idea of a blind person writing AD still seems counterintuitive. “It’s been accepted that we can do all the other roles,” Matos Olivares says. “Not so much the writing.”

AD is created and added after a film, television program or exhibition is largely complete, treating accessibility as the final stage of production rather than part of the creative process.

Cheryl Green is a US-based, non-blind, disabled, and chronically ill audio description writer, narrator, producer, consultant, and trainer who works collaboratively with blind professionals. She says that by the time describers are brought in, many decisions that shape good AD have already been made. For instance, editors may leave too little space between dialogue for narration to fit naturally, or transitions may be too brief to convey essential visual information. When Green and her longtime collaborator and blind AD professional, Thomas Reid, consult together on under-production films, they give them feedback on pacing and timing to support better AD.

The involvement of an audio describer at the beginning of production gives them the liberty to shape how accessibility is built into a piece of work rather than forcing it around a finished edit.

What blind professionals bring

In a blind-led production, blind quality control experts may work with sighted collaborators or use tools such as AI-powered image recognition and visual assistance services like SeeingAI, AIRA, and BeMyEyes to access visual information. They then apply their own expertise and judgment to determine which details are meaningful and appropriate for audio description. AI can identify images and generate descriptions of what it sees,” says Matos Olivares. “It can be a useful tool for accessing visual information, but it cannot replace the interpretation, context, and editorial judgment required to create meaningful audio description.

While comparative studies have yet to measure whether blind-led audio description produces better scripts, blind audiences say that the descriptions are blind-friendly and richer when informed by lived experience.

“I find that when there is blind writing or quality control, audio descriptions are inherently better,” says US-based blind lawyer Qudsiya Naqui. “Blind professionals know what we want and what sighted people miss.”

Since the blind community is not monolithic, there is always a debate about how much description is good description. Some want more, some like less. Naqui says that’s where the involvement of blind and low-vision professionals at every stage, such as writing, quality control, and review, might help in negotiating such tensions around good AD.

Ramya Amuthan, a blind Toronto-based audio description narrator and consultant who works across film, television, video games and live performance, says good AD begins with understanding how blind audiences experience the world.

Descriptions that rely on visual conventions are not the best. Saying a stove dial is blue or that a table is 1.5 meters long may mean little to someone who has never had usable vision. Instead, she prefers descriptions grounded in familiar physical references: a table “about the size of a queen mattress” or a person whose chest reaches the height of a kitchen counter.

“AD should be a translation of visual perception,” she says.

Some non-negotiables for her are details sighted audiences absorb immediately, such as a character’s apparent race, the time of day, or a person’s posture or bearing.

“These are things, if not described, we don’t get,” Amuthan says, recalling films where she never discovered the actors’ race. “I find that to be extremely disconcerting.”

Matos Olivares says that audio description should not soften or censor the original work.

“Don’t censor me. If there’s a lot of cursing in the show, it’s okay to have that in the description,” she says. “Don’t infantilize me either. If there’s a sexy, steamy scene, I want to know all about it.”

Collaboration changes the work.

At SoundScribe, an accessibility consultancy in London, writing is done collaboratively.

Elaine Punia-Joseph, a non-blind AD professional, and Kirin Saeed, who has worked in audio description for nearly three decades and is blind, develop scripts together from the outset, rather than waiting until the end for quality control.

The collaboration changed how Punia-Joseph thinks about description.

While describing the London Transport Museum’s Mind the Gap exhibition, Punia-Joseph realized her spatial descriptions weren’t working. Saeed explored the exhibition by touch and reshaped them from a blind visitor’s perspective. “I think about my audio description in a completely different way now,” Punia-Joseph says.

“Art could really easily stagnate when sighted people gate-keep the standards, and the ways that it can be experimented with,” Punia-Joseph says. When the audience is part of the culture shaping how their own experience is described, she said, “it can then evolve and go to new places.”

For example, when Saeed writes descriptions of theatre productions for her organization, Visually Impaired Creators Scotland, the approach is to weave them into the dialogue. For example, “I get the cast to say, ‘I love your bright red dress, and it matches your suitcase, it’s red too.’ So it becomes part of the show, and the audience doesn’t realize it’s being described.”

That approach has also influenced filmmakers. Jenny Brady, whose last five projects have involved SoundScribe’s AD services, now treats accessibility as a creative consideration from the get-go. “It really is the first thing she thinks about when making a film: how am I gonna do the audio description for this,” Punia-Joseph says.

Scaling the industry

Blind-led audio description is still far from standard practice, but there are signs that parts of the industry are beginning to move in that direction.

Netflix, that added more than 13,000 hours of audio-described programming across 34 languages in 2025, says it encourages its audio-description partners to include blind and low-vision professionals throughout the production process. In an email, Elisa Beniero, senior manager of strategy and innovation at Netflix, said the company’s new authoring platform, Originator Studio, was designed with full screen-reader compatibility so blind professionals can participate fully in audio description workflows. Netflix has also funded the Audio Description Training Retreat for two years, and its recent graduates now work with production companies, including the Canadian audio description company Descriptive Video Works (DVW).

Rhys Lloyd, the studio head at DVW, works with several leading streaming platforms. “There are no streaming services that are actively requesting this,” he says of blind-led production specifically. “However, there are streaming services who do award more work to companies that engage the services of blind talent.”

One reason adoption of blind-led AD remains slow is that meaningful collaboration costs more than simply generating descriptions as cheaply as possible.

In the US, accessibility laws such as the Americans with Disabilities Act and the 21st Century Communications and Video Accessibility Act require digital content to be accessible. Matos Olivares says that has encouraged some companies to treat audio description as a compliance exercise rather than a creative one.

“Companies and even individual content creators aren’t doing this out of the kindness of their hearts,” she says. “A lot of them are just checking a box. They’ll give the work to whoever does it for the cheapest.”

Some blind viewers would prefer synthetic narration to no audio description at all. Others avoid it entirely. Whatever the preference, Matos Olivares says, the stakes go beyond convenience. “I put a lot of trust into the idea that what I’m being told is true.”

Lloyd says streaming platforms are experimenting with AI, including synthetic voices, but, based on the demonstrations he has seen, the adoption of AI-generated description itself “has a long way to go.”

Building the workforce

The biggest obstacle to blind-led production is not proving that it works, but expanding the number of people who can do it.

AD work makes it difficult for practitioners to build sustainable careers, as it is “inconsistent and not well-paid,” says Amuthan, a freelancer.

Training opportunities are equally limited.

Green and her collaborator, Reid, the producer of the podcast Reid My Mind Radio, run free 9- to 12-hour workshops in Boston, New York, and St. Louis, prioritizing students from the global majority and training blind and sighted students side by side. But without outside funding, she says, the program can’t expand beyond the three cities where local organizations sponsor it.

In the UK, Saeed says that formal training for blind describers barely exists. The country’s main accreditation bodies train sighted describers almost exclusively.

“We learn description by watching how sighted describers narrate and by talking to, listening to, and teaching each other.”

A new accreditation policy now in development in the UK by the Audio Description Association worries Saeed. Formal certification, she says, could lock out experienced, talented blind describers who can’t afford or access the assessment process. “Do I have to pay for it? Is it gonna be accessible for me?” are some of the concerns it raises.

In the US, Matos Olivares says that many blind people do not realize audio description as a viable career option. Others lack access to assistive technology, recording equipment, or the training needed to read scripts fluently. Some others, including herself, have faced bias and lack of trust in their competency because of their blindness and other identities such as race, which is a huge deterrent in picking it up as a career.

“A person might really want it,” she says, “but it’s not as simple as, ‘Hey, I’m blind and watch lots of described content. Let me in.’”

The mainstreaming of blind-led AD may be slow, but blind and low vision professionals are ensuring to bring blind culture into accessibility.  

Saeed actively works to establish quality control processes involving blind and low-vision people and to advocate for fair compensation for their assessments and feedback. She also helps artists find ways to make their work collaborative and accessible without compromising their artistic vision.

Matos Olivares recently received internal funding at DVW for a blind co-authorship initiative to expand opportunities for blind and low-vision professionals to become audio description writers. In her role as workflow manager for quality & inclusion at DVW, she works to advance accessibility and inclusion within audio description workflows. She is also a project lead, quality control specialist, and bilingual voice talent with SADC, which is developing training and resources to support blind professionals in quality control and other roles within the field. When she first entered the industry, Reid mentored her, and she hopes to create similar opportunities for others. ‘I want to be to others what Thomas has been to me,’ says Matos Olivares.