posted
Great and informative string. As stated there is a lot more information to cover, some of which relate to SQ. From what I understand of the digital domain, the type of DAC's or sampling rate has no, or little effect on SQ compared to error correction issues.
-------------------- E.C. Wuz here Posts: 1057 | From: Vancouver, British Columbia, Canada 'eh | Registered: May 1999
| IP: Logged |
posted
Yes I would agree that higher-order DAC's(20 bit as opposed to 1 bit) have little or no effect on the sound quality. Of coarse, the quality of construction and the quality of the DAC's components will have a lot to do with things. A cheaply built 20-bit DAC will sound worse than a well built 1-bit DAC. OK here's an interesting point...another thing that really matters a lot is the TYPE of interpolation the DAC does to "connect" the bits of data during the signal reconstruction process. For instance, imagine that you are "looking" at a very small digital sample of a sine wave. What you would see is a bunch of small dots that somewhat look like a sine wave. Now, imagine that it's your job to turn those digital dots into an analogue sine wave. One way you could do it is to just connect all of the dots by a straight line. Now, the result of this would look a bit like a sine wave, but it would be composed of a bunch of short, but straight lines. Now, imagine that instead of just connecting the digital "dots(samples)" by straight lines, you instead drew curves to connect the lines. Now, the result of that would look more like a sine wave than when you had just used straight lines, right? This is an example of two differnt methods of D-A conversion. You might ask, is there a way to connect all of the "dots" so that the result looks EXACLTY like a sine wave? Well, in theory, there does exist a mathematical model which, when applied to the digital "dots," would create the exact sine wave. However, we do not yet have the mathematical tools to derive that model. Ash
-------------------- ****************************** ****************************** Posts: 1130 | From: Los Angeles, CA | Registered: Aug 2001
| IP: Logged |
posted
I suppose the dillema of reproducing an analogue waveform starts at the begining. That is when the initial Conversion and sampling is done. When an analog signal is sampled (at a sampling frequency of 44.1KHz), it is being sliced up into vertical pieces. Each vertical piece is then estimated as to its amplitude (How large is the audio signal at this instant?). This estimation process is the data converters job. It compares the original signal against its best estimate and chooses the closest answer, and that's where the problem is. The more bits, the more choices the data converter has to choose from, but it's still a "best guess" deal.
-------------------- E.C. Wuz here Posts: 1057 | From: Vancouver, British Columbia, Canada 'eh | Registered: May 1999
| IP: Logged |
posted
Well, I sit here and talk about the problems with D/A and A/D convertors, but really I don't think there are any major problems. I mean sure the post-conversion waveform isn't an exact match to the original, but do we really hear the difference? I mean as long as the Nyquest rate or higher was used in the sampling process, and the converters are built well, I think the post-conversion waveform will be a fine representation of the original. Or, I guess it could be that I've never heard a truely "spectacular" convertor and don't have a good reference as to what a very good converter sounds like in comparison to an average one. What do ya'll think?
-------------------- ****************************** ****************************** Posts: 1130 | From: Los Angeles, CA | Registered: Aug 2001
| IP: Logged |
posted
Well, the age old question is back, do we really hear the difference? I have to say a resounding YES! Having tried a demo of vinyl and CD, same music, same album, and I cued the cd so that when switched, not a syllable was missing, The only thing I couldn't do instantly and absolut perfectly was volume level control, but I had 5 people in the sound room, and we all agreed on a level for the phono. There was no question about it, the LP sounded more open, spacious, and had better detail on strings, female vocal, and piano. The CD tended to be more upfront and a tad closed in. The equipment used was a $2000. cd player(Luxman hybrid) and a Thorens/Stanton combo on an Angstrom mat (that's a flat heavy rubber mat) Keeping in mind that the human ear is tuned to even order harmonics, and the digital domain can only reproduce odd order harmonics, it's no wonder manufacturers are trying to come up with ways to "better" something that has a specific design.
-------------------- E.C. Wuz here Posts: 1057 | From: Vancouver, British Columbia, Canada 'eh | Registered: May 1999
| IP: Logged |
posted
Well, I agree with you on the point about even vs. odd harmonics. But, here's my selling point. This is a car audio forum. No one around here is carting LP's around in their car. Any slight audible advantage an analogue system has over a digital system(even then its really a matter of opinion) would be completely destroyed by the fact that driving around town makes a lot of noise.
-------------------- ****************************** ****************************** Posts: 1130 | From: Los Angeles, CA | Registered: Aug 2001
| IP: Logged |
posted
I agree, the issue was the digital domain and how D/A's work. I see the digital issue as a practical way to hear music, but being practical sometimes means having to make a compromise. As I was saying manufacturers are constantly trying to improve SQ, who knows maybe someone might come up with some new way.
-------------------- E.C. Wuz here Posts: 1057 | From: Vancouver, British Columbia, Canada 'eh | Registered: May 1999
| IP: Logged |
Printer-friendly view of this topic