Determining maximum/minimum values on decode

Bob · 23 January 2006, 04:40 pm

Is there a way I can determine what the maximum/minimum values in a stream are before I decode it? I'm getting values exceeding +- 1.0

I'm trying to avoid clipping the output, I'd rather scale it instead.

Shy · 23 January 2006, 07:25 pm

You could apply ReplayGain on the encoded file and view the peak value. It's not possible to avoid internal clipping with any lossy codec. The peak value will never be over 1.0 with the original wave file, only with a lossy file.
The louder and denser (more "compressed") the audio is (example: most modern poorly mastered pop albums) the more clipped the encoded file will be.

Scaling is never recommended, especially if you want to do it to avoid clipping which is useless because that clipping is not audible. Clipping in lossy codecs is not ordinary clipping as you may think, it's very different than say taking a wave file and changing its volume with a wave editor, thus causing (possibly very audible) clipping.

23 January 2006, 04:40 pm	#1
Bob Member Join Date: Dec 2005 Posts: 7	Determining maximum/minimum values on decode Is there a way I can determine what the maximum/minimum values in a stream are before I decode it? I'm getting values exceeding +- 1.0 I'm trying to avoid clipping the output, I'd rather scale it instead.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Switch to Linear Mode Hybrid Mode Switch to Threaded Mode

23 January 2006, 07:25 pm	#2
Shy Admin Join Date: Jul 2004 Posts: 372	You could apply ReplayGain on the encoded file and view the peak value. It's not possible to avoid internal clipping with any lossy codec. The peak value will never be over 1.0 with the original wave file, only with a lossy file. The louder and denser (more "compressed") the audio is (example: most modern poorly mastered pop albums) the more clipped the encoded file will be. Scaling is never recommended, especially if you want to do it to avoid clipping which is useless because that clipping is not audible. Clipping in lossy codecs is not ordinary clipping as you may think, it's very different than say taking a wave file and changing its volume with a wave editor, thus causing (possibly very audible) clipping.