Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

OneCore rate slider seems to have very limited actual rate range on 2019.2 #10097

Open
ultrasound1372 opened this issue Aug 14, 2019 · 8 comments

Comments

@ultrasound1372
Copy link

A friend (@garretk18) has brought up to me that upon upgrading to 2019.2 from 2019.1, Microsoft Mark is slower. He previously had the windows and NVDA rates set to 100% and was getting blazingly fast speech. Upon upgrading to 2019.2, he noticed it was slower, and enabled rate boost to compensate. Still noticing it was slower, it came up in conversation that this was occurring. I tested with the beginning of a readme file and noticed that at least from 50 to 100 is rather little rate change, on other synths it might be the equivalent of 15. The same can be said with rate boost on, though with rate boost on I honestly cannot tell the difference between speeds 50 and 100. He was running on 1809, I'm running on 1903.

Steps to reproduce:

Select Microsoft Mark in the windows OneCore voices driver, and set the rate to 50. Read a long piece of text. Then set the rate to 100 and read the same piece of text. Do the same with rate boost on. Notice how narrow the rate change actually is. This new API change from #7498 was supposed to increase it's range. It may have actually narrowed it, or constrained it to if the windows rate had been 50. He runs his OneCore at approximately the same rate as espeak NG with rate boost on set to 35%.

System configuration

NVDA installed/portable/running from source:

installed

NVDA version:

2019.2

Windows version:

observed on one system running windows 10 1809 and another running 1903

Name and version of other software in use when reproducing the issue:

Microsoft Mark - English (US) Windows OneCore voice.

Other information about your system:

The windows rate was set to 100% in the 2019.1 speed he used.

Other questions

Does the issue still occur after restarting your PC?

yes.

Have you tried any other versions of NVDA? If so, please report their behaviors.

See above.

@garrettk18
Copy link

I'm the friend that @ultrasound1372 mentioned, he just spelled my handle wrong. If you need more info/testing, just let me know.

@trodick
Copy link

trodick commented Oct 3, 2019

I can't reproduce this, following the steps to reproduce. For me, running Windows 10 1903 (build 18362.356), NVDA 2019.2.1 results in a noticeable change in the speech rate with Windows OneCore set to 100% with and without rate boost enabled.

@lukaszgo1
Copy link
Contributor

@LeonarddeR Have you had time to look into this?

@LeonarddeR
Copy link
Collaborator

I've just looked into this briefly.

The OneCore docs say that rate may range from 0.5 to 6.0. This is exactly the range we allow with rate boost, so with rate boost on, 0% is equal to 0.5, 100% is equal to 6.0 and 50% is equal to 3.25. This means that from 50 to 100%, there theoretically is a significant rate change. However, it looks like in practise, this works differently.

It is not clear to me how Microsoft expects us to increase the rate range, nor have I heard proof of the synth speaking faster than it can speak now. It would help if someone could make audio samples of that, i.e. make OneCore speak at maximum rate in NVDA 2019.2 and NVDA 2019.1 with the same text, and than compare the rates.

@garrettk18
Copy link

@LeonarddeR If you can point me to a link to 2019.1 (I have 2018.3RC3, go figure!), I can produce this in about 5 minutes.

@trodick
Copy link

trodick commented May 2, 2020

@garrettk18 I'm not @LeonarddeR, but perhaps grab the 2019.1 tag and build it from source? https://github.com/nvaccess/nvda/releases/tag/release-2019.1

@Adriani90
Copy link
Collaborator

@garrettk18, @trodick please if you are testing, do this in the last NVDA alpha version because this might have changed alot since 2019.2 or 2019.1 version. Thanks.

@Adriani90
Copy link
Collaborator

The alpha snapshots are here:
https://www.nvaccess.org/files/nvda/snapshots/

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

6 participants