Phil Leigh;578308 Wrote: > No - aaaargh - start again :-) > > Your TV analogy doesn't apply at all. Twice the pixel density = four > times the file fize = 4x information. Bit depth increase from 16 to 24 > = file size goes up by 50% because each number (sample) being stored is > +50% more precise. > > > > The number of possible values only affects the PRECISION of each > sample. It doesn't alter the amount of information, just its accuracy. > > > The only way in which all bits are equal is that each bit gives you 6dB > of Dynamic Range. When there loud stuff going on you simply can't hear > the loss of the really quiet stuff. On classical music with protracted > passages at -30dB... you might. The other place to look is in the > reverb tails at the very end of tracks. > > Of course it's much MUCH easier for us to hear loud things... and very > hard for us to hear quiet things happening at the same time as loud > things. This is part of the reason why MP3 compression works at all. > > It also explains why 13-bit radio was considered state of the art > (better than any available tape machine!) in the 70's and why some > fairly serious test have shown that you have to reduce bit-depth to > about 10 before MOST people hear a definite degradation in sound > quality. > > ...and why Philips thought 14-bit DACs on early CD players were a good > idea... > > 10 bits would equate to 60dB SNR which is not great - but it's about > on par with good non-Dolby cassette playback... > > So no, the difference between 16 and 24 (21) bit playback is not huge > by any way you want to measure it. > > The real benefits of "24-bit" are: > 1) greater headroom and more accurate DSP in recording / mastering = > less distortion/noise - and these benefits are mostly retained after > dithering down to 16-bit - if done properly. > 2) the most non-linear bits are the lowest ones and they are further > away from audible music in 24-bit DAC's than in 16-bit DAC's > 3) A theroretically lower absolute noise floor, but as I've explained, > in practice this compromised to some extent by the ADC's and other > upstream equipment inthe recording chain. Probably still worthwhile > though > > Really it's item 1 that makes "24-bit" very important. I'm sure you're right about the first part, and my attempted analogy was unwise, and i am very grateful for the correction. However it is correct that the 19 bit binary number has 8 times more possible values as i said. I can see that this is not really analagous to the number of pixels, and in information theory terms is not an increase in the amount of information only its precision. I still can't see how the increase in bit depths can only be relevant to the resolution of noise at low levels as opposed to the resolution of small differences at higher levels.
The question I was addressing (however ineptly) was whether and if so why an increase in DAC resolution of 16 bits to 19 is significant (you can take huge or strictly in this case "not huge" to mean a variety of things in different contexts.) The benefit of "24 bit" recording is not the issue although I can see that the issues are related because if there were no benefit in 24 bit recording vs 16 bit, there couldn't be any benefit in any more than 16 bit playback. I remain intrigued by the feeling that the the maximum signal relative to noise floor is not the issue, which is why many people feel that analogue signals with high noise floor can still have a greater resolution than even properly implemented 16 bit. *Perhaps you can help me on this, Sean: if an analog system has an snr of 14 bit equivalent, does this mean that its resolution of noises of amplitude above the noise floor at best equal to 14 bit?*. The comment "in audiophile terms" is one I would happily stand by as it is a very important qualifier. Audiophiles spend large amount on small incremental improvements. -- adamdea ------------------------------------------------------------------------ adamdea's Profile: http://forums.slimdevices.com/member.php?userid=37603 View this thread: http://forums.slimdevices.com/showthread.php?t=82050 _______________________________________________ audiophiles mailing list [email protected] http://lists.slimdevices.com/mailman/listinfo/audiophiles
