Within software I've written, I'm currently testing every single sample to get the peak volume of music files, but that is slow for a large file and I don't think necessary. Can I cheat and only check every 100th sample (for example) to get a close approximation of the peak volume. 100% accuracy is not required. My thinking is a loud tone that is audible to humans would be longer than 100 samples at 48000 samples per second. Am I correct?