|
|
Posted by Don Pearce on 10/16/65 11:52
On 10 Jul 2006 10:56:07 -0400, kludge@panix.com (Scott Dorsey) wrote:
>Don Pearce <nospam@nospam.com> wrote:
>>On Mon, 10 Jul 2006 11:10:55 GMT, "dadiOH" <dadiOH@guesswhere.com>
>>wrote:
>>
>>>David Morgan (MAMS) wrote:
>>>
>>>> I had a look, and I can't see how the device works without modifying
>>>> dynamic range
>>>
>>>Simply put...
>>>
>>>1. Find max volume among all songs
>>>2. "Turn up the volume" in all others so their max is the same as the
>>>loudest.
>>
>>Won't work - can't work. If you have a mixture of music, then peak
>>levels as related to average loudness will vary wildly. The loudest
>>sounding will have highly compressed dynamics, with most of the tune
>>crammed against the limit. If you try to increase the levels of all
>>the others until they sound as loud, they will all be clipped to hell.
>
>What you're forgetting here is that the MP3 is not just PCM data with
>absolute scalar values for every sample. There is a "volume" field in
>the header which tells the playback application how loud to play the
>thing. By tweaking THAT, you can turn down the level of a track without
>affecting the actual resolution (except of course the resolution is still
>limited by the converters at the final playback, and if the gain is lowered
>digitally, that will be reduced... still, that's a marginal problem at
>best).
>
>The thing is, that field is ALWAYS maxed out whenever anyone hands you
>an MP3, because all the encoders want their products to play back as loudly
>as possible. So you can usually turn it down, but seldom can you turn it
>up.
>--scott
So.........?
d
--
Pearce Consulting
http://www.pearce.uk.com
Navigation:
[Reply to this message]
|