整数与Q15数据类型的区别

嗨，我正在使用PIC32处理音频处理应用程序。为了保持足够快的速度来实时处理音频，我使用整数数学。我已经通过使用乘法和位移位运算来近似分数整数，如SO：这基本上使得整数作为分数*（n／32768）的值而起作用，例如，如果它们由1个符号位和15个小数位组成。这是很好的，因为我有一个读写16位整数的编***音频均衡器过滤库帮助手册的1.7.2节中的Q15和Q31数据类型的章节时，它表示Q15由1个符号位和15个小数位组成，值符号*（n／32768）和THA。t乘两个Q15数，一个如下：在我看来，这两种方法的作用完全相同。两者以相同的方式表示值（以它们的比特表示的方式），并且乘法是使用相同的方法来完成的。它们的功能不同，我看不到。然后看来，使用Q15格式除了提供与DSP库函数一起工作之外，没有任何优势，因为类型不起作用，只是语法问题。我有什么遗漏了吗？



      以下为原文

    Hi, I am working on an audio processing application using PIC32. In order to keep things fast enough to process audio in real time I am using integer math. I was already approximating fractional integers by using multiply and bitshift operations like so:
int16_t a, b, c;int32_t temp32;temp32 = (a * b);c = temp32 >> 15;which essentially causes the integers to function as fractional types with a value of sign * (N/32768), e.g. as if they consisted of 1 sign bit and 15 fractional bits. This works well since I have a codec which reads and writes 16 bit integers.

When reading the section on the Q15 and Q31 data types in section 1.7.2 of the Audio Equalizer Filtering Library Help manual, it says Q15 consists of 1 sign bit and 15 fractional bits, with a value sign * (N/32768), and that to multiply two Q15 numbers one does the following:
libq_q15_t A16, B16, C16;libq_q31_t Temp32;Temp32 = A16 * B16;C16 = Temp32>>(16-1);It seems to me that these two methods function in exactly the same way. Both represent values in the same manner (in terms of what their bits represent), and multiplication is accomplished using the same method. They literally do not function differently in any way that I can see. It then seems that using the Q15 format provides no advantage other than being built to work with the DSP library functions, which, given that the types do not function differently, is only a matter of syntax.

Is there something I am missing?