[mythtv] [mythtv-commits] Ticket #4775: Change defaut audioupmix setting; add help text

Mark Spieth mark at digivation.com.au
Mon Feb 25 12:55:59 UTC 2008


>>>   * Changes the default upmix setting to "Passive," which requires the
>>>   least amount of CPU hoursepower and is thus the safest choice.
>>>  * Adds to the help text for the setting to somewhat explain each
>>>   setting's CPU needs.
>
>> 19 +                "'Passive' is the least demanding on the CPU. "
>> 20 +                "'Active Simple' is more demanding and 'Active 
>> Linear'
>> 21 +                "is the most demanding (but highest quality)."));
>
>Could you guys explain the difference between the different upmix
>settings? This question comes up quite frequently on IRC and everyone
>just wants to know how an upmixed 2ch sound can be lower or higher
>quality?

its all a matter of separation of the channels.
passive has poor/no separation of L/R from center and SR=SL
active "simple" uses a nonlinear source steering algo but uses less CPU than 
the active linear method
both simple/nonlinear and linear both have excellent separation.
perhaps better than DPL2 the author suggests from tests/measurements he has 
done.

if you picture a circle where the source of the audio lies at some angle wrt 
the observer (you),
simple will distort the apparent location of the source whereas linear will 
more accurately present its location.

>Is it possible to approximately tell the CPU needed for each of the 
>settings?

I cant remember extactly but passive was about +2-5% , simple was 10-15% and 
linear about 2-4%  more than simple. this is on a XP2000+ playing the same 
bit of an AVI

hope this helps
mark 



More information about the mythtv-dev mailing list