Trimming Dead Space

Less data saved and less data streaming for a couple.

MBGanttLess data saved and less data streaming for a couple.

Don't understand. For a couple of what? The size of a mixed down song is not determined by the complexity nor number of files making up the mix but on its length and resolution.
 
I agree that it more than likely has to do with getting rid of any noise, rumble, pops, etc. 

An extreme example, but I I record my self at the beginning of a song going '1 ah 2, and 1234' and at the end, with 'Yeah!', you wouldn't want 4 mins of silence between, you's cut it out and the the 2 tiny files would replace the one big one, plus it would be cleaner as said.

BlueTubes Expander/Gate VST, but it will just deal with the volume, not split the file.

rampage2
Is there an automatic way to reduce to -∞ in a audio file all noise under , i.e. , 50db ?

 
Process / Apply Effect / Remove silence.  This will delete all the portions of the waveform where the sound level is lower than a specified amount.  If there is background noise during the singing however, that will still be present when the singing is present. 

Probably the biggest advantage already mentioned is the FX processing. If FX are heavy-handed, each sample requires processing (whether it is silent/low volume or not), so having discrete clips will remove unnecessary CPU overhead on playback.
 
The visual use is another big one, since it is quick to see/find discrete components within a track.
 
The streaming part is a double-edged sword, since a clip is really a window into the underlying wav file. If the associated wav files are all over the place, it can still be giving some unnecessary disk I/O (unless the underlying wav was bounced at some point).
 
Bottom line is that with the capability of modern machines, FX processing is not a massive issue, but each one has to do "some" work as the transport runs. If an effect is "set" there is no reason to keep processing it each pass, so judicious use of bouncing can also alleviate this overhead quite a bit (depending on the effect used).
 
To that end, this question -
 

rampage2
 
Is there an automatic way to reduce to -∞ in a audio file all noise under , i.e. , 50db ?
I mean , if I record a voice of an entire track is there a way to delete background noises leaving only the real singing ?
 

 
Is one where I do a destructive noise reduction edit religiously (with Audition). Although gates, etc. will deal with this, there is no reason to "process" that wav repeatedly. I remove it, bounce (overwrite actually) the wav, and forget about it. Destructive edits done within SONAR will still leave the original audio on disk, but doing that NR edit via the Utilities menu actually overwrites the wav, so it also cuts down on disk space used. A gate inserted as a clip FX, then bouncing the clip will achieve the same thing within SONAR, only leave the original wav file in the audio folder.