Have you tried the Azure Speech Studio? I wonder how your custom model compares to this solution.
I played around with python scripts for the same purpose. The AI gives feedback that can be transformed to a percentage of correctness. One annoyance is that for Mandarin, the percentage is calculated at the character level, whereas with English, it gives you a more granular score at the phoneme level.
> One annoyance is that for Mandarin, the percentage is calculated at the character level, whereas with English, it gives you a more granular score at the phoneme level.
This is the case for most solutions you'd find for this task. Probably because of the 1 character -> 1 syllable property. It's pretty straightforward to split the detected pinyin into initial+final and build a score from that though.
Qt Widgets is fantastic, but now dated since it has not been updated for the modern world. Plus it is C++, which a lot of devs dislike.
QML feels like a refresh with great ideas, bringing declarative UI and reactive programming. Where it falls short for me is it does not have feature parity with Qt Widgets, so you end up having to roll up your own components, wasting a ton of time. Dealing with layouts in QML is also an exercise in frustration.
Not having an extra language to deal with and so many features being just a flag away is why I decided to go with QtWidgets for the GUI of a project I am working on at work. And it is so nice to use despite being very old. For the graphically intensive parts I am just using Vulkan. I understand this might not be enough for all types of GUIs though and just wish QtWidgets had some sort of GPU acceleration.
If anything, that’s their excuse for keeping it locked.
“Mr hacker man can trick you to download an app and take all of grandmas money from the bank”, “North Korean hackers can tap into your baby cameras unless we gate the app installation process and charge 30%.”, etc.
Whatever product manager team decided to jump into React Native with both feet for the Windows experience needs to be ejected from the industry permanently. Think of how many thousands of human hours per day are now spent waiting on React Native jank, all in order to save the Windows developers from having to program in Windows using Microsoft products.
Most likely saving Windows developers from learning programming. They are just javascript monkeys because that is what Microsoft is hiring these days and you can't trust monkeys with native C++. The thing would leak memory and explode immediately if those idiots tried to write native code. So in the name of hiring cheap idiots that can't program we all have to put up with this slow, bloated garbage.
There is no universe in which I should see lag between a click and a menu appearing in File explorer, taskbar or anywhere in the OS. Not on a machine with 8 idling cores.
MS could/should have just made other XAML/MAUI options a better experience in general over the React Native thing... It might be different if they actually embrace web as a whole and at least gave a consistent UX, more like say WebOS or ChromeOS, but that's not what they're doing here.
What MS really needs to do is create a really long checklist of all the UI defained configurations and options, along with a connected list of all the relevant API interfaces they connect to... then come up with a consistent, complete and competent component library to do a ground up re-implementation of all the things in a consistent way.
This would, of course mean stabilizing the released version of windows to mostly bug fixes for a couple years while frantically generating and dog-fooding the new UX... starting with a re-revamped task manager, and launcher/file-picker. Just a bare desktop and a hotkey that opens task manager as the first and only UI elements then working out from there.
You do have to wonder what the hell the people in charge of the Windows UI are thinking. They seem to have got it so badly wrong. But so has Apple in macos.
I'm not sure the cost would be significantly worse than all the half-assed abandoned efforts so far... and it would result in the first consistent UX in Windows since Win2k.
Actually in the latest MacOS release it constantly trains local models for all kinds of junk without asking you at all.
I had to disable all that AI garbage in the settings after researching a bunch of local inference and training processes that were grinding my MacBook to a halt every time it woke up from sleep.
> The ones hit hardest are usually younger generations
Reminds me of prop 13. If you challenge grandma having a $3M house paying peanuts for property taxes you are a monster.
If you defend young people that are ready to start a family, "they can kick rocks and move to Bumfuck, Middle-Of-Nowhere, no one is entitled to live in the Bay Area".
CA Prop 13 was an unfortunate, short-term bandaid in 1978 that didn't address excessive property taxes for elderly, disabled, and poor people who came after them. It truly was another boomer selfishness moment. The solution is to expand Prop 13 to all who meet low income requirements to make property taxes progressive rather than unreasonable "flat" taxes that punish the poor far more than the rich and moderately rich.
PS: I grew up in south San Jose, graduated from Leland, but can't afford a home anywhere near where I grew up because rich people from all over the world gentrified the Bay Area and boomers went full NIMBY on new developments.
boomers were just coming of age politically when prop 13 passed in 1978. the main culprits were actually the silent generation and older greatest gen homeowners—think postwar suburbanites who had bought in cheap and were now watching their property taxes spike in a period of wild inflation + ballooning home values. boomers were still mostly renters or too young to own, especially in california’s pricey metros.
Incorrect. They were buying homes like my parents (born in 47 and 48) did in San Jose, who benefitted greatly from Prop 13. It was the grandparents of boomers (Jarvis and the older bits of the Greatest Gen) who were initially impacted by rapidly rising property values and property taxes who pushed for it.
I played around with python scripts for the same purpose. The AI gives feedback that can be transformed to a percentage of correctness. One annoyance is that for Mandarin, the percentage is calculated at the character level, whereas with English, it gives you a more granular score at the phoneme level.
reply