CUNY Academic Commons

Thanks for this, Laurie! I am adding a few watchers here, and perhaps we can touch on this during the team meeting given other recent accessibility changes and the emphasis there.

Hi All,

Following up on this to provide a bit more context since I have to miss Friday's team meeting....

I think offering some mechanism for text to audio on the Commons would be great for accessibility. More often now you might be a seeing "make this website talk" across the web, and it would be great if the commons offered this feature, too.

What I suggested above is a browser add on that anyone could use but requires the user to install.

Thinking about this a bit more, I am wondering if we want to instead offer text to speech through the Commons via plugin that works across the network (and does not rely on the user to install)

There a a couple plugins that might do this:

1. https://wordpress.org/plugins/text-to-audio/ - this plugin appears to be widely used and update often. You can use shortcode to embed a text to audio button on any page or post. Maybe we could add this to the footer across the network? We could also make the button available in a widget area on teaching sites.

2. https://www.trinityaudio.ai/the-trinity-audio-wordpress-plugin-implementation-guide -- this may not work as well because it requires an API Key and not sure if that restricts us from network enabling. Though I like the option to add audio on a post-by post basis and the audio player at the top if the post is great - it is similar to websites that now that the "listen to this article" feature.

3. https://wordpress.org/plugins/responsivevoice-text-to-speech/ - this is another plugin that uses shortcodes to add the text to speech button,but also requires API key. Based on install numbers, it looks like the most popular text to speech plugin

Recapping our discussion during the monthly meeting today, Boone briefed us on some work he's been doing with the Open Lab on a text-to-speech tool that they asked for as a way to help students in commuting and other scenarios where audio content might be helpful, outside of standard accessibility needs.

Boone researched the browser and the Wordpress plugin tools that integrate with text-to-speech engines, and he found them all to be severely lacking technically. TThese services have large computational needs, so they need to integrate with cloud computing like Amazon, and it gets messy quickly.

On the other hand, computers themselves (and in particular mobile devices) have their own "client side" text to speech systems built in already that are more useful and reliable. It is good that the mobile ones are stronger and makes sense, because the portability aspect is what's driven a lot of this.

Boone built a proof-of-concept client side text-to-speech tool for the Open Lab. Free, no external services, favors mobile which is better for use cases envisioned anyway. He will look into bringing this to the Commons once it's farther along. It's likely a better investment of our resources than the alternatives, even though it will have limited use cases (can't be used for things like OCR on PDFs), at least at first.

Also, regarding the Commons homepage having gaps in how it's read by the other text-to-speech tools, Boone is not concerned about this given the semantic markup that we use. It's likely more of a failure of these tools themselves.

Boone built a proof-of-concept client side text-to-speech tool for the Open Lab. Free, no external services, favors mobile which is better for use cases envisioned anyway. He will look into bringing this to the Commons once it's farther along. It's likely a better investment of our resources than the alternatives, even though it will have limited use cases (can't be used for things like OCR on PDFs), at least at first.

Colin, thanks for this. To add a bit more context: Speech synthesis on desktop OSes is generally pretty bad, and on some systems it's missing altogether. It's likely that a first version of this tool will only be enabled for mobile devices, where the speech tools are known to be usable. I'll continue to keep the group updated as we make progress on the tool.