|
所在平台: Udemy |
课程主页: https://www.udemy.com/course/big-data-code-optimization-in-python-numpy-sound-processing/
课程评论：没有评论
课程名称：用Python NumPy编辑声音：提升代码性能1000倍 课程概述：编程是一门极具灵活性的领域，能够以多种方式创建满足特定任务的程序。然而，并非所有的方法都同样优秀，尤其是在处理大数据时。大数据意味着要处理巨大的数组和矩阵。通过本课程，你将学习如何编写实现相同功能但速度快1000倍的程序，这一切取决于你的代码编写及编码实践的选择。 这是一个100%基于项目的课程，我们将使用Python、Numpy和Moviepy库来创建一个功能完整的声音处理程序。该程序将按顺序导入视频，提取音频，自动识别音频中的静音间隔，并切除这些间隔，同时在句子之间保留一定的静默。音频处理自然涉及数以百万计的数组元素，因此编写程序的方式至关重要。我们将通过展示糟糕和良好的编程实践，让你看到两者的区别。 最终，你将看到你用Python Numpy编写的最后版本的代码比第一个版本的速度快1000多倍，从而了解如何正确以及避免错误的编码方式。此外，这些知识是普遍适用的，不仅限于音频处理。在课程的最后部分，我们将进行一个计算机视觉的作业，数字图像实际上是巨大的矩阵，如何在代码中处理它们同样重要。我们将使用Python图像处理库Pillow来处理图像矩阵中的大数据，构建一个可以二值化这些图像的小程序，并同样展示好坏两种实现方式。 完成本课程后，你将掌握从一开始就采用正确编程方法的能力。如果你对我们的免费预览视频感兴趣，赶快注册吧，让我们开始学习吧！期待在课程中见到你。
Programming is one of the most flexible fields I know of. You can create a program that achieves a certain task in so many ways. However, that does not mean that all ways are equal. Some are better than others.That is especially visible when your program has to work with big data. Working with big data means working with gigantic arrays and matrices.You can create a program that achieves the same task like the other one, but it does so 1000 times faster. It all depends on how you code and which coding practices you use.And this is what you will learn here. You will learn the good and the bad coding practices, so that you would learn to code the right way when dealing with big data.In this 100% project based course, we will use Python, the Numpy and the Moviepy library to create a fully functional sound processing program.This program will import your videos in sequence, extract their audio, automatically identify the silent intervals in that audio, and then cut them out while still keeping some silence on the edges to preserve a bit of pause in between sentences.Sound processing naturally deals with millions and millions array elements and so it really matters how we write that program. We will do it in a bad way and in a good way, because I want you to see both sides of the coin.In the end, you will see that the last version of your Python Numpy code will be more than 1000 times faster than the first version, and so, you will see how to code and how definitely not to code.Finally, I really want you to see that this knowledge is universal and can be applied in other fields as well, not only audio processing. And therefore, in the last section, there will be an assignment in computer vision.Digital images are in fact, gigantic matrices, and so, it really matters how you handle them in the code. We will build a small program that can binarize these images and we will also do it in a good and in a bad way.We will use the Python image processing library called Pillow to process all this big data inside the image matrices.After this course, you will know how to approach programming in the right way from the beginning. Take a look at some of my free preview videos and if you like what you see, then, ENROLL NOW and let's get started! I'll see you inside.
