jeff@tesla.UUCP (09/21/83)
The phase shift vs. frequency characteristics and square-wave distortion reasons are summarized very well in the "Computers & Electronics" article on CD players a couple of months ago. It seems to me that, yes, those phase shift characteristics are completely trivial compared to the phase shift vs. frequency characteristics that characterize a symphony orchestra in a concert hall. Speaker manufacturers tried to make much of this by setting tweeters back from woofers by a few inches, a few years ago. Just another "new speaker design!" gimmick. I doubt that ANY distortion (phase, etc.) in a CD player is audible at all. Certainly, however, most of the software available is not up to the rest of the system. Jeff
pmr@drufl.UUCP (Rastocny) (09/22/83)
It amazes me as to how many people believe what they read and few believe their own ears. If you don't believe that a 180 degree phase shift is audibly confusing, switch the polarity of the speaker wire on one of your speakers. This is slightly different from what is happening in the D/A conversion, but the correlation can be made. Things just don't sound quite right with this much phase shift. The best way to compare analog and digital playback integrity is to listen to violins. If you know what a violin sounds like, analog is closer to reproducing one more correctly. Next, listen to an instrument with little harmonic content above 2KHz like a low-level passages of bass drum to eliminate a dynamic range contest. Digital now sounds more accurate (phase is more linear in this region). Phil Rastocny AT&T Information Systems Laboratories drufl!pmr
michaelk@tekmdp.UUCP (Michael Kersenbrock) (09/23/83)
Whereas a CD player handles both channels the same, shouldn't the experiment have you swap leads on *both* speakers before looking for the effect? Further, just swap leads on *both* tweeters. Mike Kersenbrock Tektronix Microcomputer Development Products Aloha, Oregon P.S.- The above doesn't really have much to do about the "real" problem of group delay distortion (which due to other effects isn't a "real" problem at high frequencies anyway)
dyer@wivax.UUCP (Stephen Dyer) (09/24/83)
Switching the polarity of the wires to one speaker demonstrates a 180 degree phase shift RELATIVE to the other speaker. The filtering necessary during digital recording causes phase shift, but it's applied uniformly to both channels. Very few stereo components maintain "absolute phase", anyway. Many well regarded power amps, for example, invert their inputs. /Steve Dyer decvax!wivax!dyer
howard@metheus.UUCP (09/29/83)
To Phil Rastocny (drufl!pmr): I don't recall ANYONE claiming that you couldn't hear the difference when you phase-shift one of two related signals which are being put onto separate speakers, or even onto one speaker. Consider two identical sine waves. If we add them IN phase we get a louder sine wave, if we add them OUT of phase we get nothing. Clearly these will be distinguishable by the ear if the original sine waves were audible. However, this has NOTHING to do with most of the preceding discussion about phase shifts. It is still true that phase-shifting harmonics with respect to one another IN A SINGLE SIGNAL produces NO (or extremely little) perceptible change in sound. Try reversing the leads to BOTH your speakers at once, thus shifting the phase of ALL the signal, and tell me you can tell the difference THEN. I think you'll find you can't. Next time you want to make snide remarks about people's submissions to the net, try understanding them first. You may find there's no need after all. Howard A. Landman tektronix!ogcvax!metheus!howard
rlr@pyuxn.UUCP (09/30/83)
I believe that there has been research done that states that one cannot tell the difference between a tone (say a square wave) which has a certain harmonic content (a square wave contains all odd harmonics in amounts inversely proportional to their harmonic number) and another tone with equivalent harmonic content where the phases of some of the harmonic components are reversed in the resulting waveform (which in this case, would no longer be a square wave in shape as a result). It is my understanding that THIS is the kind of phase shift that we are talking about here. This does not imply anything about one speaker producing the square wave and the other producing the altered tone (i.e., phase shift in one of the two speakers but not the other). It implies that if a range of frequencies are shifted uniformly in both speakers, the listener could not distinguish this from the original sound. This was from a course in electronic music and acoustics many years ago, so I don't remember where this is documented.
mat@hou5d.UUCP (M Terribile) (10/01/83)
Regarding the question of whether phase distortion i.e. dispersion and group delay can be heard, and whether harmonic content is all that matters: I would not be surprised in the least to hear that in the range of 3kHz and up there is no audible difference. On the other hand, since we can begin to distinguish between sound events that happen a few milliseconds, or a few tens of milliseconds, apart, I would not be surprised if such distortion became audible when the frequencies involved are 300 Hz or below. With CD players, the noticable phase distorion is taking place at frequencies above 10 kHz. I would be more concerned with the fact that a few people have SOME preception of sounds over 20 kHz and that CDs are limited to not much more than that. Mark Terribile hou5d!mat