kmr wrote: > pippin;299889 Wrote: > >> No. I have the same thing in my car radio. This is usually not a >> correlation thing, because you don't hear correlations, as soon as the >> first track is different you are on a completely new trail and don't >> tell me that you can hear a correlation like in >> 1 20 4 80 5 vs 2 40 8 160 10 >> >> What you describe usually is a bad randseed issue, that is: the same >> start value is used on consecutive rand sequences instead of using >> something like time() as a start value. >> > > I think you're mistaken here; if the RNG seed were the same, you'd get > the SAME random number sequence. What I get with my iPod and dislike > is similar but not identical random number sequences - a classic sign > of correlation with the RNG (i.e., different seeds produce similar > sequences). And yes, I can tell that, out of several thousand songs, I > tend to hear the same ones in the first 100 or so. >
Well, the software is open source so pry out the Fisher-Yates shuffling algorithm and the RNG and create a benchmark setup to measure the correlation you mention. Tell us about the result and I'm sure people will start to work on a better algorithm especially with a method to measure the effect. Me, I'm sceptical. The human mind is so eager to find patterns, it often mistakes random chance for destiny. I can't tell you how many times my player plays a song that applies to the situation I'm in. I swear it's psychic! Regards, Peter _______________________________________________ discuss mailing list [email protected] http://lists.slimdevices.com/lists/listinfo/discuss
