Absolute phase, why does it matter?

This question has bugged me for a while and I was reminded of that by the DAC64 phase question.

Looking through boards such as AA, it seems to be taken as gospel that absolute phase matters but I have never found an explanation as to why it matters.

Can someone please enlighten me.

thanks

reg

 

I was speaking with Derek@ overkill Audio yesterday, and He says his speakers are designed with this abosolute phase coherience in mind, with tonal accuracy and nautraliness, lower down the design criteria. Wm

 

But Tone, phase coherence is completely different to absolute phase AFAIK. I know all about phase coherence - it's one of the major strengths of my speakers

reg

 

As I stated on the DAC64 thread, I can't hear the difference between a normal and phase inverted signal. I really can't see how it could matter or make a difference. I tried switching the speaker cables around to invert phase when I got my DAC64 but when it made absolutely no difference at all I switched them back.

Take a sine wave as an example - phase inversion merely has the effect of turning the sine wave "upside down" which results in exactly the same sound as you would get if you moved your head fractionally backwards or forwards (the amount depending on the frequency of course).

Michael.

 

Okay Michael,

I can see that, experimentally, it doesn't seem to matter. So why the big song and dance about it? Is there any theoretical justification for obsessing about it all? IIRC, early CD players used to have a 'de-emphasis' button to invert the phase on recordings which had been mastered with the phase inverted. In common with your experience, I can't say I noticed any difference between de-emphasis on or off.

reg

 

In general there is no such thing as absolute phase, only phase relative to some reference point. If you're talking in terms of the phase inversion option on some equipment, I must confess I never understood how this could make the slightest sonic difference.

 

reg, in answer to your question I've no idea why anyone makes a big song and dance about it. As MartinC said, there's only ever phase relative to a reference point, never absolute phase.

As for de-emphasis, that has nothing to do with phase. It was an attempt at a kind of "Dolby noise reduction" for CDs (why, I have no idea since CDs already have an incredible s/n ratio). Recordings were made with an HF boost (emphasis) and then players would apply a corresponding HF rolloff (de-emphasis) which is more or less how Dolby NR for tapes worked aswell.

I believe that most CDPs and DACs these days can detect the emphasis flag and do de-emphasis processing accordingly without requiring user intervention (my DAC64 and previously my Tag DAC20 are like that). The Tag had a little LED that came on to tell you that de-emphasis was being applied. I only noticed it on a couple of classical CDs I have.

In the end it didn't work terribly well and these days hardly any CDs (if indeed any) use emphasis.

Michael.

 

having also swapped my speaker leads round and hearing zero difference i've come to the conclusion that it's another one of those things that those truly afflicted with audioboxaphillia can witter and faff about with to make 'jaw dropping' improvements in sound quality that are inaudiable to those without bionic hearing.
in other words, and in true zero gain stylee, it's all a load of bollox.
cheers


julian

 

Audio signals aren't symmetrical, look at a rip in an editor sometime, I don't think the symmetry of a sine wave is a good argument for absolute phase not mattering.

For the record I think it matters a little, and whether you can hear the difference is very source dependent. Much more worth playing with than snakey mains cables or digital interconnects.

FWIW when a kick drum is kicked I want to be pushed back rather than sucked in... Unless the studio engineer wired the cable up wrong of course.

Paul

 

Thanks michael, for setting me straight on de-emphasis. I should know better than to rely on memory alone without checking the facts first!

<Going away to think about what Paul's just written before commenting any further!>

reg

 

Oh Paul your such a damp squib
Reg, just repeating what the guy was stating, Phase inverstion on the wadia does diddly squat (I do believe some early recordings were recorded inverted and this option is for these)
Phase abolsute? to what would you use a reference?, A/C wave forms does it matter? (input side), speaker signals? digital signals?

Thermal noise in electronic equipment is typically created by the excitation of phonons, which are resonances in conductors caused by the movement of electrons through the quantum well. Phonons are a type of noise in themselves; they are caused by a certain class of electrons colliding with the crystal lattice of conductors (silver, copper, etc.) through which they are being transmitted. Such phonons travel, not at the speed of light, but of sound, and therefore are instantly out of phase with the signal. Shot noise, which is generated by the operation of transistors, is similar in nature to thermal or white noise. Low-frequency (1/f) noise is always present, and increases as an inverse power of the frequency.

Quantum physicists have learned that electrons have an intrinsic angular characteristic expressed in terms of spin (either up or down), which describes their orbital behavior around the nucleus of an atom. When subjected to a high-temperature near-superconductive material, electrons tend to join in a beneficial manner, increasing the velocity of propagation (VP) by forming what are known as Cooper's Pairs (one spin-up electron joined with a spin-down). Coopers' Pairs have the unique ability to tunnel through the crystal lattice of the conductor (such as a copper wire) essentially unimpeded, therefore eliminating virtually all quantum noise phenomena. To understand this effect, imagine a football game in which the player receiving the kick off could run straight down field to the goal line without being touched by any defenders.


These qualities are beneficial in numerous ways. When placed between an amplifier's power transformer and diode bridge, for example, undesirable impedance mismatches, In an amplifier-to-speaker connection, the absence of reactance creates an optimal signal transfer and presents an easier load to the amplifier. When transmitting digital information, this then eliminates the overshoot and ringing that can occur in the leading edge of the square wave. This type of distortion is a major contributor to the harshness and glare often associated with digital sound.


Some more snake oil for you to play around with Paul

 

Tone - I don't know what you've been smoking - but what does any of that have to do with phase inversion?

Paul - I, like reg, will have to think about what you said before I respond :bookworm:

Michael.

 

Mike, just thought you could use this is some phase inversion technique, thats all Wm

 

WM, impressive stuff. Please put me down for some of whatever you're on, too.

 

:respect: :kneel: omg wm!!

apart from that, I have had 2 dacs with digital phase inversion and can tell diddly squat, too..

what I am curious about is when people talk about phase, the very fact of putting something tho' electronics alters the phase...capacitors shift phase 90 degrees, so as soon as something is recorded, going through all that electronics will affect the phase, then playing it back, you can't not affect it somehow...amplifiers, chips, speaker crossovers, so I am sceptical about what people mean when they say..phase coherent, phase accurate, etc....

 

Right - I have thought about it a bit more and I disagree with you Paul I don't think the kick drum example you used any more representative than my sine wave.

If you saw a kick drum sound on a .wav file you would see a more or less symmetrical spike so the idea about being "sucked in" instead of "pushed back" is really a fallacy. The "thud" of a kick drum is actually a short burst of a high amplitude, low frequency (with many harmonics) waveform. I doubt there are many real world sounds that consist of a single positive or negative impulse.

Michael.

 

Ok, I'm gunna really put the cat in the pigeons, I use phase re-alignment devices (compensators if you like), in my system, and not just on the pre bridge rectifier side either, in various places, and seems you guys like what they do, you've heard em If fact, I put some in Timpys system today.
BTW Ian, not all caps Phase shift, Black gates don't (so I'm Informed), does anyone know/has measured this?, to whether it actually makes/does contisute a difference?
what about repolarising the shift before final exit?, just a thought?
How about impedence matching?, does this tie in with this configuration of phase shift?, Questions?, answers?, aint got round to doing that one yet but maybe soon, permitting the surgery on the Wadia is completed this week.
Mind you which cable company all quotes VP in the sales blurb
Is it all bollox, do we really care humm, well if it gives an advancement on what we've got then yep, I'm in

Lifted from the Rubycon site (Blackgates)

"Successfully eliminated ion distortion and also non linear distortion which is inevitable with conventional polarised electrolyte capacitors because of the non-reversibleness of the electrodes. Keeps the total amount of distortion down to –150db or less."

 

I think you've been misinformed, Tone. I think it's the nature of capacitors that they shift the phase. That's why some valve amps, like the Soul and the Steinharts (to name two) use trannies instead of coupling transformers between stages.

 

Steve, I'm not going to argue here (I only the 3rd party, however they work damn well), Some guys even use rechargeable batteries between them, so as the valve is discharging? (help me with this one) the battery charges up, on the valves 'off cycle?' and only needs a 2 year maintainace programme, as you throw the valve away at the same time as the battery. (One of Andys little tricks)

A formation voltage for cathode electrode oxide film of Black Gate can be selected to be an arbitrary value that is different from that of anode electrode. Therefore, the voltages for cathode electrode formation of every Black Gate product are selected to be appropriate values for each of the whole products as shown in Figure 4, within which a non-polarised operation with very low distortion is secured. Moreover, all of the cathode electrode oxide films are permanent. For BG-WKZ 500V used for a tube amplifier, the voltage is selected as 250V. With this voltage, the amplifier maintains a complete non-distortion operation to the full power of A-class voltage amplification. This result is indeed a revolution.



An electrolytic capacitor, consisting a rolled pair of aluminium electrodes set face to face, forms a specific resonance frequency between the inductance L element of the electrodes and its self capacitance. In this type of capacitor, it is impossible to eliminate resonance. Normally the resonance frequency is about 200KHz at 100 µF, about 70KHz at I000µF, and about 35KHz at 2200µF. These values apply almost equally to capacitors of the same capacitance per area, whether polarised or non-polarised.
As the impedance and E.S.R. characteristic curves D and E in Figure 5 shows, the BG-N (100v 2200µF) has a resonance point around 35KHz. With this as the turning point, the impedance and E.S.R. Increase as the frequency increases, causing the function of capacitor to stop. Also, at the-resonance point, the phase of the internal current changes drastically from negative to positive. That Is to say, the band above the resonance point is an imaginary part in any electronics product, if signals enter in this band, the impedance and phase change. Engineers design and make products without giving much consideration to this fact. Even if they are aware of this, there has been no effective remedial means of remedying this.



Cancelling resonance frequency by using the characteristics of Black Gate-N.
Jelmax has developed a new system which uses a pair of two BG-Ns of the same voltage and capacity in such a way that their internal inductance's (L) cancelled each other by utilising the mechanical characteristic of the BG-N using a rolled pair of electrode foils of the same area and same mechanism set face to face, and thereby, allows the pair to function as a single ideal capacitor called Super E-Caps".
Figure 6 shows how this works. Two non-polarised capacitors BG-N C, and C: have their respective inner capacities CN1 and CN2 equivalent series resistance E.S.R elements RS1 and RS2 very small inductance's L1 and L2 resulting from the internal electrode foil roll, and resonance frequencies each other Inductance's L1 and L2 generate magnetic fluxes m1 and m2 Since both electrodes of BG-Ns of the same specifications are the same and symmetric in mechanism and size, with their rolls start and end at the same positions. Suppose signal o is applied to a pair of the same BG-Ns corrected in parallel. If magnetic fluxes m1 and m2 are in such a relation that they are cancelled with each other as shown in the figure, the total resonance frequency disappears in load ZL. The pair of the BG-Ns leaves nothing but CN1 and CN2, serving as an ideal capacitor allowing the Impedance and E.S.R to lower infinitely as the frequency increases. This can be attained only by the BG-N Pair. Hence it's naming of "Super E-Caps".

This is also shown in Figure 5 In the BG-N Pair, the capacity is twice as large as and the E.S.R. is half as small as those of a single BG-N. In addition, as a result of the elimination of resonance frequency, the impedance and E.S.R. lower infinitely as the frequency increases.



Marking to show the starting and ending positions of a foil roll.
The BG - N and BG - NX are always manufactured in the same manner as shown in Figure 7, each product shows the toil roll starting and ending positions by means of lead lines of different lengths or by the marking of O and × on the terminal rivets. These allow users to connect these capacitors in parallel or in series as needed to cancel the resonance frequency due to residual inductance.



Sample power supply circuits using Super E-Caps.
Figure 8 gives a sample + 2power supply "Super E-Caps" L-Cancel Pair circuit widely used for IC power supplies it is a power supply of 12V and 5V. C1 to C6 are BG-Ns. Correcting them as shown in the figure based on the rivet marks and/or lead line lengths forms an ideal power supply for very high-speed, high-density information processing meeting the requirements for a very low level of noise and a very high speed.
Figure 9 is a parallel L-Cancel Pair power supply circuit diagram just shown in Figure 6 'principle of BG-N pair Super E-Caps'. The alternating voltage or the secondary side is rectified, causing DC output to be generated at BG-N C1 and C2 and also 12V DC output to be generated at BG-N C3 and C4 on the output side of The stabilisation circuit T1. When an alternating supply on the secondary side is to be given as intermittent high frequency pulse voltage it is a switching power supply and its applications are numerous. This type of circuit can keep the level of residual ripple-noise down to several mV. The BG-N permits making the world's first completely noiseless switching power supply.







In summary above, the main features of Super-E Caps are as follows:

An electron transfer

A completely non-polarised capacitor

No limitation to operation frequency

No ripple noise a Long life

No digital phase change

Small size, light weight

Cost saving

Providing a power source completely free from any restriction.

Sorry Steve shold have mentioned NO Digital phase change Tone

However good all of the above is, it's still got to pass through chips and coils and I/C's I know, hence why I use, those other devices.

 

Tone - reading your post is like drinking out of a fire hose

Michael.