我们可以从WP7中的录音中获得最高频率吗? [英] Can we get the highest frequency level from a recording in WP7?

问题描述

我正在Windows Phone上做一个应用程序，以确定使用手机本身录制的录音的频率级别.
是否有确定频率水平的算法?
到目前为止，我只能测量/确定声音的最大音量.
谢谢！

Hi,

I am doing an app on the Windows Phone to determine frequency level of a recording that is recorded using the phone itself.
Is there and algorithms to determine the frequency level??
So far i can only measure/determine the loudest level of the sound.
Thanks!

推荐答案

请查看我对问题的所有评论.仔细阅读.

这是我的建议:放弃.您将节省大量时间，并避免感到沮丧.音频比较的所有任务都非常困难.很少有产品可以相对接近此类任务，但是只能在非常狭窄的应用领域中使用.而且，您没有背景，只是了解您真正想要的是什么，不去解释它.真的，做些更现实的事情.

如果您有兴趣了解该领域，请阅读有关光谱分析，傅立叶变换的基本信息:

​​ http://en.wikipedia.org/wiki/Audio_spectrum#Sound [ http://en.wikipedia.org/wiki/Spectrum [ http://en.wikipedia.org/wiki/Fourier_analysis [ http://en.wikipedia.org/wiki/Harmonic_analysis [ http://en.wikipedia.org/wiki/Fourier_transform [ http://en.wikipedia.org/wiki/Fft [
没有声音记录是一个频率.没有频率水平".记录的所有可测量特征都应仔细定义以使其具有某种意义.而且，不同但相似的记录的比较更加复杂.尝试严格定义它(否则，您甚至不能谈论编程，一个正式的系统)，并且您可能会发现它有多难.

—SA

Please see all my comments to the question. Read thoroughly.

Here is my advice: give up. You will save a lot of time and avoid great frustration. All tasks of audio comparison are extremely difficult. There are very few products which gets relatively close to such tasks, but only in a very narrow field of application. And you have no background just to understand what do you really want, set aside explaining it. Really, do something more realistic.

If you are interested to get basic understanding of the field, read the basic information on spectral analysis, Fourier transform:

http://en.wikipedia.org/wiki/Audio_spectrum#Sound[^],
http://en.wikipedia.org/wiki/Spectrum[^],
http://en.wikipedia.org/wiki/Fourier_analysis[^],
http://en.wikipedia.org/wiki/Harmonic_analysis[^],
http://en.wikipedia.org/wiki/Fourier_transform[^],
http://en.wikipedia.org/wiki/Fft[^].

Every fragment of sound has a spectrum which can be considered as a discreet spectrum consisting of a series with, in general case, infinite number of periodic members (like sine functions), but even this is only an approximation, which is strictly the case only for the sound of infinite duration. If the sound sample is limited in time, it can only be represented as an continuous spectrum. In functional analysis, the set of the components which gives a general-case continuous function is not just infinite, it is considered as a point in a space of infinite number of dimensions (!). Of course, in "technical" spectral analysis, some reasonable approximations are used (which does not make the theory simple then with the ideal analysis; it makes it only more difficult). For example, there is no a practical need to consider frequencies not perceived but the human organs of hearing. Besides, it should be taken into account that the digital record has limited finite number of samples, but this is again, a complication of the theory, not simplification.

No sound record is one frequency. There is no a "frequency level". All measurable characteristics of a record should be carefully defined to make some sense. And the comparison of different but similar records is even more complex. Try to define it strictly (otherwise you cannot even talk in term of programming, a formal system), and you probably will be able to see how hard is that.

—SA

这篇关于我们可以从WP7中的录音中获得最高频率吗?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！