This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(179.20000000000002, 63%, 27%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EJ%3C/text%3E%3C/svg%3E)
does Microsoft Speech Platform -- ISpRecoContext [Microsoft Speech Platform Microsoft Speech Platform Native Code API] perform better?
i am using the windows crate with rust -- for the following. it seems like there is some lag with that api for Windows.Media.SpeechRecognition
the "voice access" app has similar lag too during dictation. but it is NOT known what API it's using: Windows.Media.SpeechRecognition???
F.Y.I. i am NOT interested in the 100% cloud api but would possibly try out the win32 -- ISpRecoContext
does another local/offline 'api' have less lag than: Windows.Media.SpeechRecognition? am open to an MacOS Apple solution as i've already done some research and investigation into that but the "cost of entry" for current hardware is like 4 times higher!!!
i have limited experience with -- win32 api. however was learning Swift for MacOs
ISpRecoContext is not a win32 api. There is no winapi to realize speech recognition. And now the latest supported speech solutions are in the Azure Cognitive Services for Speech.
okay. i'm gonna try out the free local/offline open-ai 'whisper' speech recognition as well as possibly 'vosk' all for free. you do realize that the Web Speech API inside Google Chrome does seem to have "less lag" -- than my experience with Microsoft "voice access" as well as what i threw together with the Rust language 'windows' crate that's on Windows.Media.SpeechRecognition. b.t.w. some of these cloud api's charge $500 a month. have had issues before with internet access
ISpRecoContext is not a win32 api. There is no winapi to realize speech recognition.
It is SAPI : https://learn.microsoft.com/en-us/previous-versions/windows/desktop/ms720151(v=vs.85)#api-for-speech-recognition
I tested MS SAPI samples like DictationPad and I don't see particular lag
using Windows.Media.SpeechRecognition the test i had was just getting a "speech recogonition" result from a SpeechContinuousRecognitionSession. using that test i had noticed more 'lag' than using speech to text dictation in the Google Chrome browser [which is apparently called soda]. the microsoft tool called "voice access" has a similar 'lag' on my computer
so that made me think that maybe this - which you say is SAPI- is faster. but i still DON'T know if it is indeed faster;
according to a microsoft webpage, SAPI is ---
Legacy features
Legacy, deprecated, and/or unsupported versions of Microsoft speech and conversation technology.
###################################
however is there less lag using SAPI than getting a "speech recognition" result from a SpeechContinuousRecognitionSession of Windows.Media.SpeechRecognition?
the example link you provide (about dictationpad) would be helpful if i were to set up a 2nd test at some point to compare dictation.
however i have NOT had time -- at this point to see if SAPI does have less lag compared to what i've already tried (using rust)