On 04/02/10 23:06, Kellen wrote: > > Any comments on this quote from the SimAudio site referenced previously:
Yeah.... > "32-bit Processing > > The digital-to-analog conversion process uses bits of digital > information to produce an analog waveform, represented as a sine wave > (in this example only a part of a sine wave is shown) to produce a music > signal. When more digital information is made available, the result is a > more accurate and detailed music signal. *A higher bit-depth (or a > higher resolution) yields smaller, finer and more accurate steps in the > reconstruction of this sine wave (grey) as seen in the animated figure > below. A 32-bit (blue) data stream contains significantly more > information than a 24-bit (green) or 16-bit (red) data stream.* > > As a general rule, increased processing power is directly proportional > to resolution: For each additional bit of resolution,the number of > availble levels will double, as shown in the following table: > > Bit Depth/Steps > > 16 / 65,536 > 20 / 1,048,576 > 24 /16,777,216 > 32 /4,294,967,206 > > To summarize, the higher the resolution of the data, the smaller the > steps will be. This results in more detail as each individual step > covers a smaller section of the waveform. More detail in the digital > domain leads to a much more analog signal signal at the end of the > conversion process. That bit was rubbish. redbook audio has 16 bits. Higher bit depth is better, but the source material doesn't have it. > Another major benefit of this 32-bit process is smaller data truncation > errors. These errors result from the extensive mathematical calculations > performed during upsampling & sampling rate conversion, prior to the > D-to-A process. These truncation errors - shown in the above figure, > occur when the data sample rises outside (or above) the grey curve - > will be signifcantly smaller as bit-depth increases due to finer and > more accurate calculations. As well, because of the sheer processing > power that's readily available at 32-bits, the errors will not impede > the circuit's ability to accurately recreate the music waveform." That bit is correct. Using 32 bits for calculations increases the accuracy of the calculations and avoids truncation/rounding errors. R. _______________________________________________ audiophiles mailing list audiophiles@lists.slimdevices.com http://lists.slimdevices.com/mailman/listinfo/audiophiles