Samsung Android Accessibility Framework 35.0 Overview: The Good and the Bad

Last updated on 8 August 2025

For the past couple of years, I’ve been reviewing Samsung Accessibility Framework evolutions, starting with Framework 33, and then we saw together how it came along to being very stable, at least with the third-party Jieshuo screen reader. 34.1.1, for example, was an evolution I’ve been waiting for since 33.0. Finally, the notification shade and quick settings started to not close anymore if you were too fast or if you tapped on an empty place.

Accessibility Framework 35.0 doesn’t have much new, to be exact. Personally, I am not sure how the experience is for TalkBack users, but for Jieshuo, at least, everything seems to be stable.

Accessibility Framework 35.0: What’s Really Changed?

1. Changes to the Accessibility Volume Streaming Channel

   Android 15 completely changed the accessibility volume stream when it comes to the audio effects applied and equalization. Fortunately, there’s only good news about this—news that can only make me happy.

   • The accessibility volume got bassier. Not as bassy as the media volume, but it did get a little bassier.
   • There’s a kind of dynamic range compression that actually makes screen reader sounds, somehow, better and fuller. It’s more optimized for speech than ever before. Those with better ears will notice the difference in the compression itself. It’s a welcome change, more so knowing that the old equalization settings for the accessibility volume stream on Android 14 and below bothered some people.
   • SoundBooster does still matter, but Android15 completely changed the audio logic and digital signal processing, only available on this year’s flagships, where Artificial Intelligence based limiting, compression, gates and more, are used on the phone’s speakers to improve their sound. Doesn’t apply to headphones by default.
     Works on any custom ROM port from this year’s flagships, as it’s in the core jar logic of Android.

2. Sound Feedback as Lists Auto-Scroll

   If you are in a chat, have read the last message, and a new one comes in, screen readers—at least Jieshuo—will make a sound, the end-of-scrolling sound, to tell you that elements on the screen have auto-scrolled. It’s one of those small things that count, further reinforcing my stance: Jieshuo gets better and better as the framework updates.

3. Fingerprint Hint Has Been Fixed

   I’m sure you all remember the fingerprint message: “The fingerprint sensor is centered on the lower part of the screen. You’ll hear feedback when you touch and hold it.”

   It seems I was predicting its functionality correctly. Now, when you change windows from the lockscreen to the notification shade, then to quick settings, for example, it’s not announced anymore like it was the last time. It now works properly, and the fingerprint message is only spoken when you return to the lockscreen. I’m happy this issue got solved; it was an annoying speech feedback interruption on Android 14.

But, of course, nothing comes without its issues, and the Samsung accessibility framework is no exception. It, too, has some issues—one direct, one indirect to it—but they’re still issues, and the second one is still handled by it, most likely. If not, it’s still a serious issue, and we’ll discuss them both below:

1. Move up, move up, move up: I’ve already finished my screen’s surface! Where should I put my finger now?

   For those already on Android 15, I’m sure you know what I’m talking about. I lauded Samsung for what they did with the other fingerprint message, but this? This speaks of yet another half-baked feature that, most likely, will get fixed with new huge updates to the framework, if they properly implement it. Until we see the next big framework update, however, there’s still some time left, and we won’t see a solution to this issue in the near future.

   The idea is: No matter where you touch on the screen, it still tells you to move up, as if the sensor was above your finger, even though you’re already past it. For advanced users who remember exactly where their fingerprint is, it’s no issue, and they’ll easily ignore it and tap where they should. But for novice users who are trying to learn the position of their sensor, it’s an issue. The fact that some users are advanced, however, doesn’t mean we should dismiss this mishap and let it go. It’s still a half-baked feature that, again I say, could’ve been good, but it seems it isn’t. Let’s hope Samsung addresses this issue with framework 35.1, if there’ll be any, or at least in 36.0 in Q3 2025.

2. A Rain of Useless Neural TTS Drops

   Samsung Text-to-Speech Engine was the first TTS engine to slowly introduce more and more neural voices on Android. Those voices sounded realistic from the beginning.

   Even though their spot is now taken by Google TTS with its neural voices that are used in Pixel AI, Samsung TTS was, and most likely still is, loved, even if the voices aren’t really that great when it comes to speech dictionaries.

   The TTS voices have quite a low dictionary. It’s not just one time I’ve seen them mispronounce some names, apps, and even basic Android features.

   • Samsung TTS mispronounces Manamon, a game for the blind, Nekogram, a Telegram client, and more such apps.
   • But what I didn’t expect to hear from it is the voices pronouncing ADB as “Ad B.”

   Regardless, Samsung TTS is loved and used by some people.

   Android 15 users, however, will have an issue: Samsung TTS can only be used with Samsung TalkBack. Multi-language automatic switching helpers like Tech Freedom, AutoTTS, and Multilingual TTS can’t use the Samsung voices; it switches to Google TTS. Reader apps like Legere Reader, @Voice Aloud Reader, and any other apps can’t use the voices.

   Screen readers like Jieshuo and Prudence can’t use the voices. Not even Google TalkBack can take advantage of the voices.

   Unfortunately, there’s nothing one can do about it. Because if using TalkBack 15.1 and the latest version of Samsung TTS on Android 14 and OneUI 6, you can use the voices just fine by using the same versions of software that are used on Android 15. This behavior, unfortunately, can’t be overwritten through special settings or floating feature tweaks, so it’s stuck this way.

   The only solutions you’d have are either not upgrading to Android 15 or building a Samsung TTS player and using the voices in a third-party adaptation.

   In my honest opinion, this shouldn’t have happened. I don’t see any reason for Samsung not to remove this limitation, even from Google TalkBack itself. It is infuriating to see this happen, especially now that we have neural voices with an expressivity that’s better than it’s ever been. Even though the neural voices are quite slow for use with a screen reader, that’s all the more infuriating, as we can’t make use of them, not even in reader apps.

   I hope Samsung sees this article and makes a change, as Samsung TTS has some potential. And with this newest limitation? It’s all wasted.

Conclusion

In conclusion, Samsung Accessibility 35.0 doesn’t have much more to fix. The fingerprint location message fix was welcome, as it addressed a real speech interruption bug.

However, the fact that we are wrongly guided on where the fingerprint is, and, even more so, that it’s communicating with Samsung TTS to shut it down if not used with Samsung TalkBack, is a real “no-no.”

On the other hand, it really shows that, unintentionally, the Samsung accessibility framework supports a screen reader it wasn’t originally intended to support. Jieshuo is getting better and better on Samsung devices as time goes by. And, as a user of Jieshuo, this makes me prouder the more I see things improving and being supported.