在dsPIC30F上的xc16上的Q1.15数学运算？

你好！我有点困惑，事实上，许多文档似乎引用C30编译器没有帮助，但简单地说：我的代码运行在一个DSPIC30F2011上。它是基于Cee11AdCddSPLIbl滤波器样本。它在8K数字化，并通过2个FIR滤波器运行数据。其目的是取结果的包络并输出一个数字值，其中滤波器输出中的一个是最大的。现在，来自ADC的数据并因此在滤波器之后被签名Q1.15。基本上我想这样做：我如何告诉编译器把操作当作Q1.15？我需要为此调用特殊函数吗？比如DSP、μBug或Q15AVS（）？还是有编译器指令，只要变量都被声明为分数，编译器就可以得出结论？提前感谢！



      以下为原文

    Hi!

I am a bit confused here and the fact that a lot of the docs seem to reference the C30 compiler does not help, but briefly:

I have code running on a dsPIC30f2011. It is based on the CE001_ADC_DSP_lib_Filter sample. It digitises at 8k and runs the data through 2 FIR filters. The intention is to take the envelope of the result and output a digital value based on which one of the filter outputs is the largest.

Now, the data from the ADC and thus after the filter is signed Q1.15. Basically I want to do this:

                FIR(1, &outputSignal[0], &inputSignal[0], &filter1);
                FIR(1, &outputSignal2[0], &inputSignal[0], &filter2);

                hi_ave = (hi_ave / 5)*4 + (abs(outputSignal[0])/5); // Moving average
                lo_ave = (lo_ave / 5)*4 + (abs(outputSignal2[0])/5);
                //
                if(hi_ave > lo_ave) LATDbits.LATD0 = 1; else LATDbits.LATD0 = 0;

How do I tell the compiler to treat the operations as Q1.15? Do I need to call special functions for that? Such as the DSP __builtins or Q15avs()? Or are there complier directives that lets the compiler figure that out as long as the variables are all declared as fractional?

Thanks in advance!