As I gathered from other threads here, the read aloud-function in RL simply triggers the browser-native text-to-speech function.
As such, I don’t think switching between the different browser-native voices should be a premium feature. It’s like artificially restricting a feature of the browser and then returning it to the user for a fee.
For Edge users, the situation appears to be more peculiar still. Microsoft has great natural-sounding voices for many languages which can be used on any website for free with Edge’s read-aloud feature. However, I found that, at least for Russian, they are not available for selection in the Readlang reader. Instead, I am offered a far lower quality voice for free (Irina) and a second one only for premium users (Pavel). I have not yet bought premium, but I suspect Pavel is no better because it is not one of the voice names I can see in Edge browser under natural voices.
Sorry you feel that way. Readlang offers a lot for free and I plan to keep it that way, so that free users get a decent experience. As with many online services it is a bit arbitrary what is in the premium vs free plan and doesn’t necessarily correlate with marginal cost of providing the service. A year ago, neither the free plan nor the premium plan had Read Aloud. Now they both do! It’s just that the premium version is a bit better in that it gives choice.
I agree that the choice of voices at the moment on many devices is not great, and falls far short of state of the art TTS in 2024. Hopefully on device local TTS voices will continue to improve over the years. And if not, maybe I’ll look into using server generated TTS to provide a better and more consistent experience across all browsers. But no promises. I have limited bandwidth and it’s hard to prioritize.
I think I may have misunderstood something. You wrote elsewhere that the availability of voices depends on the browser used. I gathered from that that whatever voices are available in browser-native read-aloud-features should also be available in RL and so when I did not see those in Edge, I was disappointed.
Do you have to manually code for each voice to appear as an option? If so, I would still think it’s worth it to enable support for Microsoft’s natural voices in Edge and just straight ought recommend that browser to users who care for TTS and not even bother with supporting TTS in other browsers. The difference can hardly be overstated. Especially for non-english languages, it’s like going from something you want to turn off immediately after hitting play, to a voice you could listen to for hours.
I don’t manually code each voice. I only filter out the voices which are generated remotely as opposed to on the device for reasons I mentioned on the other thread. Maybe those Microsoft natural voices are server generated and not local? A very quick search indicates that this could well be the case: Redirecting
Could you perhaps partner with AwesomeTTS (add-on for Anki) somehow? I pay to subscribe to their service for my Anki flashcards in order to access a greater variety of accents, and the subscription cost is based on how much TTS you generate per month. I pay like $5 USD monthly for thousands of monthly card reviews.
I don’t actually use the TTS in Readlang except when I highlight words, but having a third-party TTS option could be a value-add for users who want that feature, and you wouldn’t have to build your own solution from scratch.
Readlang doesn’t have the ability to create 3rd party plugins the way that Anki does so I’m not sure how partnering with AwesomeTTS would work exactly. (Someone could make a Chrome extension which augments Readlang’s features e.g. adding TTS but I wouldn’t recommend it since it’s likely to break as I make changes to the Readlang webapp.)
Readlang already uses an external TTS provider (MS Azure) for individual words and phrases in Speaking Mode. This is higher quality than the Read Aloud voices on most browsers, and I may look into using it or something like it for Read Aloud in future.
I appreciate the move, although I found that it makes no difference for Russian. There is still only poor quality Milena on MacOS and iOS and poor quality Irina and Pavel on Windows. I saw that MacOS and iOS also have Milena for Russian so it appears to be an OS voice. It sounds the same as well. I tried downloading the enhanced version of Milena to these OSes. It sounds much better. Unfortunately, Readlang continues to use the basic version. Would this be something you can adjust on your end?
For English content, there is now a choice of 25 voices, but most of them are joke voices which no one would actually use. The most usable one is Daniel which is a British english voice. It is good, but not as good as Microsoft’s natural voices.
Unfortunately on macOS Safari you are correct, but Chrome and Brave browsers give a much better selection of voices. I tried installing “Milena (Enhanced)” and “Katya (Enhanced)” via macOS Settings → Accessibility → Spoken Content → System Voice → Manage Voices… and they in Chrome and Brave they do appear as options within Readlang. (They don’t say “Enhanced” but they sound like they are the enhanced versions)