Jump to content

Auto generate subtitles not working


generalakaba
 Share

Recommended Posts

So, i have Videopad 10.64 and i wanted to insert subtitles in a 10 minutes clip, why is the auto generate subtitles not working? 

I have selected Microsoft English recognizer 5.1 as Speech to text engine, but it doesn't seem to be working. 

Anyone knows how this function works?

Thanks.

Link to comment
Share on other sites

Hi

Can't get this to work either. As I have a French version of Vista and presumably a ST engine that works in French it still seems to produce gibberish. For example a clear audio of Maison....Maison.....Maison  ... produces a non related text.....

zzz.jpg

At a stretch of the imagination there might be SOME sound similarity but in practice it's useless! Other tests were equally fruitless, often seem to be random text...often hilarious. So IMHO this feature is best left alone. Similar sounding English and French words didn't work either.

For example Table...Table...Table (repeated)    came out as...

 La Thailande.. Thailande...Thailande.... and Tremblement de terre   Tremble....

Obviously a very shaky table (Tremblement de terre is french for an earthquake.

Sorry no good (but funny) 

Go here  https://speechnotes.co/  and say what you want then save the result as a text file. Use this in VP it seems to work OK.

Nat

Link to comment
Share on other sites

Hi

"........Anyone knows how this function works?........"

In theory at least you..........

  • Open VP and place a video clip with sound or audio clip on the timeline and select it.  
  • Click the Audio tab on the top toolbar to open the Audio toolbar. The Speech to Text tab is now activated.
  • Click the Speech to Text tab. This will open the video with a default text visible (Usually "Sample text")
  • Set the Font; Colour etc. for the required speech text
  • Click the Apply button. VP will now generate a text file based on the recorded audio. The text file is automatically placed on the overlay track at the Cursor position.

That's how it should work. Obviously the audio file should be a speaking VOICE (not music or background sound.)  You can, of course create audio voice files directly using the Audio Toolbar's Record Audio feature.

The text generator engine of your PC should be capable of recognizing the language being used. e.g. According to the Control Panel on my French Vista machine (OK it's old!) this is the Microsoft 8.0 Voice recognition utility (francais France)  So expecting text to come out in recognizable English if the audio is in english is, in this case a bit over-optimistic.

The result of an audio clip saying 0ne  two  three  four  five  six  seven  eight  nine  ten   was....

z.jpg

Which was rubbish (although the words were all french.)

The result of an audio clip saying un deux  trois  quatre  cing  six  sept  huit  neuf  dix  was...

zz.jpg

Which was reasonable (although other attempts were less successful (to say the least :lol:)  RK might be for an unrecognizable word/sound. ..Un might be "heard" as err. 

I am called Nat came out as......

zzz.jpg

Which is actually french for  I remember the act!    :o:o.....Ooops...What did I do??

 (It should have been je m'appelle Nat )

I would be interested to see some simple English examples like this. :lol:  All in all this feature needs some refining.

Nat

Link to comment
Share on other sites

In addition to the odd translation, which is purportedly the fault of the MS speech engine, subtitle generation is one long line - not linked to the spoken video.

Perhaps one could generate it piecemeal, but that's more trouble than it's worth.😆

Link to comment
Share on other sites

Hi B

I agree. It's not VP. :). I remember years ago a free disc with some magazine  that had Philips Speech Recognition Program. After almost an hour of reading lines of text into my PC, correcting and repeating, the program had reportedly learnt my voice and was competent to write what was spoken. ....or what it thought it had heard. The output was hilarious. If you ummed or ahed, coughed, sniffed or rustled a paper it it heard it via the mike and wrote gibberish (which occasionally made sense) and if you sniggered or laughed (one couldn't help it when it was appearing in real time)... it simply made it worse. There were hours of fun to be had....It was a Philips Chinese Whispers generator! :lol:

Nat

Link to comment
Share on other sites

The specialists seem to have it nailed.  Every test made here with an on-line subtitling tool has been 100% accurate.

Explored the capabilities of another prosumer NLE that doesn't do auto-generate caps... their FAQ points elsewhere.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...