Leo Ma

That means the CRUD is all right. You can define your own content that you want to store on NVM or flash storage.

`有一次` should be close to the segmentation of `一次` (not `有一` because it is not a word in Chinese) and `第一次` should be close to the segmentation of `第一`.

Well it is really confusing when you first learn Chinese on `一`, `不` and double 3rd tone. Let me show you the rough rule on `一` base on sentences above....

That is right for what you have learned.

`一线希望` can be regarded as a regular word in such case while `一线城市` should be segmented as `一`, `线` and `城市`. That is why Chinese always drives you mad.

I do not think one can do Chinese Pinyin conversion totally correct. There are no rules but conventions. A enoumous pinyin dictionary is indisensable in such issue. That is what...

I have found a well designed Chinese pinyin dictionary from [espeak](http://espeak.sourceforge.net/data/index.html) with 21567 single characters plus 36098 compound exceptions (includes 332 added 'yi' and 10720 added 'bu' exceptions, and 9713...

I am afraid I am not sure whether shall I pick your commit into this project since there are other users using it. Maybe it is better you keep this...

可以[SrsEncoder.java](https://github.com/begeekmyfriend/yasea/blob/master/library/src/main/java/net/ossrs/yasea/SrsEncoder.java#L37)里面调节编码参数，比如GOP象征是FLV缓存队列的大小，但现有代码不支持动态调节这些参数。

你确认你获取到了吗？RGBA中的alpha分量一般都是255，你自己检查一下数据是否真实的RGBA，或者自己写一段代码dump出来。hwRgbaFrame调用的是native方法，具体是在libenc.c里面的libyuv接口