On Fri, Jan 14, 2005 at 10:08:20AM +0100, Arnold Krille wrote: > On Friday 14 January 2005 03:03, Jon B wrote: > > For some reason I was under the impression that W64 was used because > > it saved 32-bit floating point data and therefore didn't clip; not > > Every soundformat can clip, the reason is not within soundformats but within > the analog-digital-converters. These little nice boxes convert some > continuous signal ranging from min to max into discret digital values. > Everything above max and below min can not be resolved and gets the values of > max/min and therefor sounds clipped. Well, theres a difference between the format clipping and the DA converters clipping, the DA converts will always (I guess ;) clip at 0dB, but the format may or not, fixed point and floating point formats can go well above 0dB to capture those transient peaks or sudden load passages that might otherwise be lost. Not an issue if your just recording from external imputs, but if you do inmachine processing it can be very helpful. - Steve
