[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference cookie::notes$archive:cd_v1

Title:	Welcome to the CD Notes Conference
Notice:	Welcome to COOKIE
Moderator:	COOKIE::ROLLOW

Created:	Mon Feb 17 1986
Last Modified:	Fri Mar 03 1989
Last Successful Update:	Fri Jun 06 1997
Number of topics:	1517
Total number of notes:	13349

441.0. "Will the real phase shift please stand up?" by PYRITE::WEAVER (Dave - Laboratory Data Products) Tue Aug 26 1986 14:06

    Prompted by note 430.9 on A/D converters.
    
    Mark, you just made me think of an interesting point.  Most
    multichannel A/D converters use multiplexed inputs.  Up until
    recently, most CD players used multiplexed outputs.  Now we have
    players with dual D/A's that prevent phase shifts.  But if the
    original A/D was multiplexed, and didn't utilize simultaneous
    sample and hold amplifiers, then there will be an inherent phase
    shift of the original source material.  If the channels were
    recorded in the same order as single D/A CD players play them
    back, then any inherent phase shift may be cancelled.  And dual
    D/A players would then be producing phase shift.
    
    What a can of worms!  I think I will hold off buying a dual-D/A
    CD player until the source becomes as good as the sink.
    
    Does anyone have actual knowledge as to how the different digital
    recorders handle the A/D conversion?
    
    							-Dave

T.R	Title	User	Personal Name	Date	Lines
441.1		PSW::WINALSKI	Paul S. Winalski	`Tue Aug 26 1986 16:57`	6
	If you tilt your head while listening to music, you also get phase shift. Or if your speakers are not exactly aligned properly. Or..... Personally, I don't think it's worth worrying about. --PSW
441.2		NSSG::KAEPPLEIN		`Tue Aug 26 1986 17:21`	11
	I think just about all professional CD recorders use dual analog-to- digital converters. Now, some of the semi-professional and portables don't, and nearly all the home-use PCM machines are single ADC and DAC -- which is fine because they cancel out the delay! So, ADD, AAD disks will probably be ok, but a small budget remote digital recording with something like a Sony F1 will have the phase shift. The DAT recorders will also only have one ADC and DAC, but again that will be ok since they will cancel on playback.
441.3		PYRITE::WEAVER	Dave - Laboratory Data Products	`Tue Aug 26 1986 19:25`	25
	You really don't need 2 A/D's, just simultaneous sample and hold amplifiers. But I tend to agree with .1 (which is why I posted this to begin with). However, a lot of the companies duing dual-D/A's are also putting in better isolation between the digital and analog circuitry. This is where I feel a lot of truly audible differences in CD players will be evident (yes, I also agree that better quality components in the analog section will also make an audible difference to some people, myself probably included.). Dual D/A's also allow the D/A to sample twice as fast to allow more effective filtering as the sampling normally done for 2 channels, only need be done on one. Thus if you have a D/A converter capable of sampling at 196.8KHz, and it samples 2 channels, each channel is limited to 88.4KHZ (2X oversampling). Using the same exact D/A but one on each channel will allow the single channel to be sampled at 196.8KHz (4X oversampling). I would trust a 196.8KHz D/A much more than the 293.6KHz D/A for 4X oversampling simply because the faster the D/A, the more it is pushing technology. I got off on a tangent; the bottom line is that I really think the phase shift is irrelevent. -Dave
441.4		ENGINE::ROTH		`Wed Aug 27 1986 09:03`	22
	One point about phase shift that's often misunderstood. The ear is very sensitive to changing interaural time delays, but much less sensitive to time delays that are completely stationary. If you are listening to a player with mulitiplexed single DAC you get a static time delay - and the effect is completely minimal. However, if you are listening to a player with passive elliptic anti-image filters, and the L/R filters have a mismatched group delay then it is possible to detect this. You're hearing the variation of left vs right delay with frequency, not the delay itself. However, it is more likely that small (.2 db or so) amplitude response variations will accompany such mismatch, and these are much more detectable. In many cases, time delays are blamed for probems that are really just frequency response problems. For example, time-aligned speakers are really time aligned to eliminate comb filtering effects in the first signal arrivals - which is a frequency response problem, not really a time delay one. - Jim
441.5		NSSG::KAEPPLEIN		`Wed Aug 27 1986 11:57`	6
	There was an interesting letter in this month's HFN&RR. I think it was from Stanley Lipschitz (sp?). Anyway, he was writing in support of a previous letter (John Curl?) which noted that capacitor effects are "linear". The bottom line is that you don't measure bad capacitors by looking at harmonic distortion, but by examining the frequency response or phase response.