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

Reply via email to