audibility
audibility of phase audibility of resistors
I spent a fair bit of time researching the question of whether EQ produces audible or objectionable phase errors, before deciding to use it for my system. There are 2 issues: are phase errors occurring and how severe; and are phase errors audible. I did not get definitive answers, but here are my conclusions.
1. Home listening room errors in terms of frequency response (FR) are huge, and FR errors are definitely audible, whereas phase error audibility is debatable. I cannot see the merit in leaving definitely audible FR errors in our system, especially when the device used is digital and does not require 31 electromechanical level sliders and a cascade of analog parametric EQ electronics. Also analog EQ of closely adjacent frequency bands creates additive errors so the correction is not like the physical appearance of the sliders, but digital EQ can correct that so the EQ does have the same shape as the "sliders". My Behringer does perform this correction mathematically (digitally).
2. The argument that we can somehow "hear through" the room effects and discern the speaker's direct sound only, i.e. that our mind can do room EQ with no phase errors, is pure wishful thinking. The claim is that the time delay in all reflected sound is detected by the mind and "gated out" psychologically, leaving only the original speaker's emissions. This claim has been psychoacoustically tested using proper scientific procedure, and found not to exist. If it did, an orchestra would sound exactly the same outdoors, indoors, and irrespective of room or hall size!
3. There is a minimalist ethic pervading hi-fi. (I should mention here that there is a valid minimalist philosophy that applies to all walks of life. I too am a minimalist and subscribe to this philosophy. But I like Einstein's version: "Everything should be as simple as possible, BUT NO SIMPLER"). In hi-fi it is taken beyond Einstein's maxim and to dysfunctional extremes, resulting in commercial success for products such as:
Almost all of the above products cost over $10,000 each, except the commonplace crippled "pure" preamp.
My point being that for home audiophiles, a decent low-distortion equaliser is unsellable. I had to go to Allan's Music (musicians' store) to find my Behringer DEQ2496, which is 24-bit, 96kHz, -114dB distortion and noise, built-in RTA, microphone and digital and analog inputs, auto-EQ or manual, 31-band EQ plus 9 parametric EQ's per channel, and under $500 --oh and another $100 for a flat-response omni mic from Behringer made for use with the equaliser. ("too cheap to be any good", say my audiophile friends, who can't understand why I pay $500 for a JBL pro tweeter with 0.03% distortion at 105dB which is no good "because it's pro audio" compared to their $200 Scanspeak tweeter, which is 0.3% distortion at 91dB and costs $20 to make). I connect my CD digital output into the Behringer, so the Behringer is a kind of "equalising DAC" in practice.
4. Phase shift is detectable, but not for normal music listening. From Dr Floyd Toole:
From another statistically corrected, scientific study:
If you want more reading, see http://www.audioholics.com/techtips/roomacoustics/HumanHearingPhaseDistortio.php. I have concluded that there is nothing audible about phase shift introduced to music by a typical equaliser, and several generations of home audiophiles are listening to gross FR errors because of their belief in a myth.
Note: of course phase shift is just time shift in a technicolour raincoat, and if we delay the bass 5 seconds after the treble, the listener would not only detect a change, he would probably not even recognise the result as being music! The relevant question is regarding the levels of phase shift that a typical crossover or equaliser introduce, and these have been researched and proven inaudible to my satisfaction, with music in a listening room.
5. As to whether phase shifts are actually occurring, the answer is yes. I was not sure whether digital EQ avoided the shifts of analog EQ. Turns out it still introduces shift. For more info see http://www.ethanwiner.com/EQPhase.html.
6. My work has lead me to study human perception and auto-suggestion, and I have come to realise that it is quite futile for lone individuals to discover their audio preferences by changing this and changing that and listening to the result and deciding whether it made a difference and whether they like it better. (This is exactly the technique used by 99.99% of audiophiles.). If we know which setup we are listening to, psychological factors will totally swamp our sensitivity to the actual changes in sound waves (assuming they are reasonably subtle changes). Blind listening is absolutely critical if we are to avoid self-delusion. I no longer accept my own impressions using "sighted" listening, because I am aware of how auto-suggestion works and it doesn't disappear just because I know about it.
Even with blind listening that leads to detectable differences, research suggests the subjects are probably hearing the EQ (level) difference that usually accompanies phase changes.
Conclusion 6 also has stern implications for the general work of a sound and recording engineer, whose basic job involves tweaking things and deciding if they prefer it. Blind testing (alone or with a test audience) is totally impractical from a work productivity perspective. I don't know the solution. The only thing I think might help would be a big dose of humility. And having a colleague with a trusted ear who you can keep "blind" and say "which do you like better?", and "what are you hearing?".
Any metal film resistor is absolutely inaudible. If you want resistors so badly made as to be audible, try the expensive and sonically highly regarded carbon composition resistors by, methinks, Allen Bradley. Of course you'll have to look for new old stock, the maker is long gone as there's no market in the electronics manufacturing industry for resistors that distort, however pleasantly.
The cheapest resistors currently available are carbon film types (not the same as carbon composition). I don't use carbon film because they are more variable and less reliable than metal film, and the price differential has eroded to nothing, so the local electronics component stores don't even stock them any more.
There are of course boutique manufacturers for everything, including resistors for audio, but in the case of resistors I regard them as industrial art. Exquisite, yes, desirable, yes (and hence justified IMHO), but functionally superior, no. A boutique metal film resistor is, like the standard item, absolutely inaudible and can be used with confidence, nay, pride. I am an industrial art appreciator (for f*** sake I owned an old Ducati), and hence I have no criticism whatsoever of those who buy and use such components. But I will resist if they rubbish the standard items for their functional performance.
I would love to see, admire, and listen to a system that uses, say, a Rothwell attenuator. The name sounds very posh.
Power resistors are another matter as they tend to be made from different materials, to heat up, some have better allowance for heat than others, and some have a nonzero inductance, so some care is needed in selection.
last edited 21 Dec 2016 (spelling), 06 august 2008 (minor tidy-up)
return to tnargsound here