How to show spectrum of the System.Speech?

Question

4.50/5 (2 votes)

See more:

, +

Hello!

I have a little problem in my win forms application.
The main idea was to show the spectrum/oscillator, of the out going sound of the System.Speech.
It supposed to look like this:
http://img196.imageshack.us/i/spectrump.png[^]

I've found a lot of projects but 99% of them was about playing a .wav, or recording from microphone.
This one was perfect:
Sound Activated Recorder with Spectrogram in C#[^]
It was drawing fast enough, and had exatcly what I needed, but it used the microphone as a source of audio, and I just couldn't find a way to change the source...

Now can some one explain me how to draw a spectogram of System.Speech, or even just the global playback of windows?

Thanks for any reply!
QwertyXoid.

Posted 8-Apr-11 8:11am

QwertyXoid

Add a Solution

Comments

Sergey Alexandrovich Kryukov 8-Apr-11 14:18pm

I voted 5 for this Questions, but not for the Question itself, but because it should be an interesting project.
--SA

1 solution

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Sergey Alexandrovich Kryukov · Accepted Answer · 2011-04-08T08:17:00

Solution 1

You cannot use anything from System.Speech to access the raw source of the audio. There is nothing like that in the engine. Instead, you still need to use the same source of audio data to feed to the System.Speech engine (I guess, the recognizer) and to your spectrum analysis unit. What is you requirements in general terms? What are the ultimate goals of your project?

[FOLLOW-UP:]

Here is what you can do: collect the raw input voice data in the close-circle buffer with appropriate time stamps in real time. Handle the events SpeechDetected, SpeechRecognitionRejected and SpeechRecognized. Have two time stamps (of the type System.DataTime), such as StartTime and EndTime. Take a first time stamp on SpeechDetected, second time stamp on SpeechRecognitionRejected and SpeechRecognized. You can use the property AudioPosition for the timing. When the time stamp EndTime is set, use both time stamp and you audio buffer to extract the appropriate fragment of your raw data and feed it the spectrum analysis unit, present the result to the UI. That's it!

But… Just for decoration? I'm somewhat frustrated… :-)

—SA

Posted 8-Apr-11 8:17am

Sergey Alexandrovich Kryukov

Updated 8-Apr-11 8:57am

v4

Comments

QwertyXoid 8-Apr-11 14:21pm

Its a small helper, that uses Text-to-Speech to notify the user, and uses recognizer to listen to the user, and his commands.
So I hoped that its possible to show the spectrum for the part of the Text-to-Speech... To make the helper look a little more cool/techy...

Sergey Alexandrovich Kryukov 8-Apr-11 14:46pm

Yes, of course. Is "more cool" is the main reason. I would advice you to avoid anything just for decoration and show only the functional things. I thought it's more serious.
Did you see the application which teach correct pronunciation and use the spectrum for feedback, point out comparison with perfect model spectrum, etc. That is serious!
--SA

Sergey Alexandrovich Kryukov 8-Apr-11 14:54pm

OK, see my update.
Isn't it about the time to formally accept my Answer?.. :-)
--SA

Sergey Alexandrovich Kryukov 8-Apr-11 15:16pm

Thank you...
Will it work for you?
Feel free to ask other questions -- I've done speech recognition. Did you face thread apartment model problem -- it needs to be resolved in some cases.
--SA

QwertyXoid 8-Apr-11 15:15pm

I'm sorry I didnt answer your expectations :(
And also... I'm sorry that I didnt actually understood your answer :(
I understood the global logic, but I'm pretty much still basic-developer, I've started exploring C# about 6 month ago. Mostly learning from examples.
So if you got any example for me, I would be VERY glad :)

Sergey Alexandrovich Kryukov 8-Apr-11 16:08pm

No problem :-) and thank you for accepting this Answer.
But... it look like you still need help.
What's wrong with the follow-up section of my Answer?
I don't have anything which can serve you as an essential sample right now.
If you need some other help, please explain what you did not understand. I would be helpful if: 1) you explain the ultimate purpose of your project; 2) what do you use for UI (essential item)? 3) does speech recognition already works for you, if not, what's the problem? 4) how do you input data and feed it to the recognizer so far? 5) which recognizer class do you use right now and why (there are two, they work on different apartment models, one of them has extended capabilies (another works only with default engine, is it a limitation for you?)? 6) did you run the spectrum analysis parts, are there any problems?
--SA

QwertyXoid 9-Apr-11 7:28am

Wow :) a lot of questions...
I didnt exactly understood how am I going to code that. I've never touched audio issues before. I'm used to find examples and then check each part of it, to understand it.
1) You know what this project supposed to do(if you ment the technical purpose), Im developing it to train my self in diffrent areas(this is the purpose for me).
2) You mean the control I want to draw the spectrum on? If yes, then its a pictureBox.
3) Speech recognition works fine for me, the problem is with the spectrum of the TTS.
4) Didnt understand the question :(
5) I think that I work with the default engine class, but I'm not sure. How can I check this?
6) I did run the spectrum analysis parts, and they did work fine for the InLine audio.
Thanks, QwerrtyXoid.

Sergey Alexandrovich Kryukov 9-Apr-11 20:24pm

1) Not quite. I'm asking what is the use of this project? Hands-free commanding (of what commands, for what application?) I would not recommend to do something for pure exercise if you have a change to do something useful. What the project is supposed to do.
2) You should never write anything dynamic on pictureBox; this is completely useless. You should draw in control via OnPaint or in Paint event. Misuse of PictureBox is one of the typical mistakes here, CodeProject experts explained it millions of times.
3) Good. What's the name of the class. There are two. (Q5)
4) Do you use just microphone or something else.
5) Tell me the name of the class. There are two similar classes in System.Speech.Recognition. Which one? Why this one. (None of them is default; there are two for your choice).
6) Good, so you have recognition, you have spectrum. You need to do them both. The only one problems was the timing: what part of audio to show in spectrum form? I suggested how you do it. What's still unclear?

--SA