Objectivist or Subjectivist? Give Me a Break
- Thread starter Jeff Fritz
- Start date
Tim, are you being abtruse now? Let's look at what he says about the 50+ year old null tests "So the notion that some audiophiles can hear aspects of fidelity that can't be measured is easily disproved." What level of nulling are you suggesting is he talking about here? i.e what null level would you suggest is required to make the above statement?
Let me help you - we have already been through these arguments before on other threads
http://www.whatsbestforum.com/showt...o-transparency&p=109621&viewfull=1#post109621
which I believe then became modified to 80dB down.
Are you saying that this 80dB figure is yet again a moveable one because it's unachievable in reality or is it just a mis-representation of what is possible when testing the playback configurations that we listen to i.e including speakers?
Uh...John
for the record I am not of the advice that all amps sound the same or that if they measure the same under usual tests (THD, Power , Fr, etc) they sound the same... and so far several measurements among them THD don't tell enough
is not the same thing as stating -80 dB null between two amplifiers ... that could be a strawman argument
for the record I am not of the advice that all amps sound the same or that if they measure the same under usual tests (THD, Power , Fr, etc) they sound the same... and so far several measurements among them THD don't tell enough
Don, any idea why that large thread on Gearslutz I linked to showed nulls from AudioDiffmaker no better than 36dB when testing AD/DA loops? There was another larger thread also reporting similar tests (which seems to have been deleted for some reason) & this also showed no better than 36dB. Problems with test procedure or AudioDiffmaker or ....?
Frantz, would not a null of 80dB imply that both amplifiers are identical sounding? Would this not mean that they are either both distorting in "exactly" the same way or that they are both transparent? Would this not be the same as stating that they are free of artifacts to a level that defines transparency - in this case 100dB defined by Ethan? What would you reckon could be the causes for differences between two amplifiers which would give a null less deep than 80dB?
I agree about your view on "usual" measurements but are you then stating, as Ethan states, that a null test trumps all measurements & has done so for 50+ years & this is the gold-standard measurement?
Nope, not trying to be obtuse, John, just didn't see anything about 80 dB, which seemed to be what the fuss was about.
Tim
Just to set an 80dB null in perspective it would require amplitude flatness to better than 0.001dB (0.01%).
Take a typical speaker with a 4ohm resistive component, it might well increase to 16ohms or more at the lower resonance. A 500 microohm output impedance would be roughly consistent with an 80dB null meaning a damping factor of 16,000 is called for.
Take a typical speaker with a 4ohm resistive component, it might well increase to 16ohms or more at the lower resonance. A 500 microohm output impedance would be roughly consistent with an 80dB null meaning a damping factor of 16,000 is called for.
Tim, I trust my post that you just quoted gives you the context for the 80dB null & how it relates to the other statements about transparency that have been made!
BTW, I would never call you obtuse (i.e stupid) but I misspelled abstruse (difficult to understand)
Well, I'm with Tim here, that's why I made my previous post. Ethan has mentioned the -80 dB null before, but if he actually tested two different amplifiers (operated into any approximation of a real world load) and nulled them to that degree, I don't remember reading it, and i probably wouldn't have believed it if I had. He has talked about comparing a signal run through A>D>A with the original and that nulling to less than -80 dB; that's more reasonable although still a little questionable.
I notice that AudioDiffMaker may be able to compensate for that, if it's nothing more than a minor frequency response change (which may, or may not be considered acceptable in an amplifier):
I'd not rely on Diffmaker if I were you - on the same GearSlutz thread that jkeny mentioned it was exposed that it has trouble compensating for phase vs frequency and therefore the nulls it gives aren't as deep as can be achieved by manual (but very timeconsuming) means.
I'm also skeptical of removing frequency response variations unless it can be shown they're completely predicted by some model - for example where its clear they're due to the variation of the load impedance. For this though we'd need to model the load accurately so it could be taken out of the picture. Nonchalantly removing frequency response variations of unknown origin suggests we may well throw out the baby along with the bathwater if the purpose is really to uncover audible differences.
I'm also skeptical of removing frequency response variations unless it can be shown they're completely predicted by some model - for example where its clear they're due to the variation of the load impedance. For this though we'd need to model the load accurately so it could be taken out of the picture. Nonchalantly removing frequency response variations of unknown origin suggests we may well throw out the baby along with the bathwater if the purpose is really to uncover audible differences.
No idea, didn't read it yet (rehearsal last night, kid's performance tonight, work, etc. make me a fly-by WBF reader for now). Since I do not know the test setup and measurement equipment I cannot comment. I can say you have to do a lot more than just match amplitude to get a good null; phase matching has always been one of the biggest challenges when I have tried it. It is hard to align the phase precisely without introducing other artifacts (e.g. from the phase-adjust circuit). Bandwidth is also a consideration; if the amplifier limits the bandwidth, often the case in RF systems, then the null test may read false (poor results) due to out-of-band content that does not really reflect the amplifier’s in-band ability. I have found null tests very cool but often very hard to implement. I do not have a lot of experience with them at audio frequencies, and what I do have was decades ago, thus my reluctance to enter into this discussion (I do not need help to show my ignorance).
Yes, though this is possible with a 10-turn trim-pot. As I recall (was almost 40 years ago) that's what the old HP analyzers used for nulling.
Well, if you're measuring at the amplifiers output terminals, and not exceeding what the amp can drive, the speaker's varying impedance shouldn't have an affect.
I've obtained nulls down to total silence using DAW recording software and Wave files. I also have a software volume control that adjusts in steps of 0.01 dB.
Phase shift can prevent a complete null, but in most gear the phase shift is at the frequency extremes - below 20 Hz due to coupling capacitors, and above 20 KHz due to digital filtering. So depending on the program material phase shift can easily be out of range.
My comments about 80 dB (or 100 dB) are more theoretical than practical. When people argue about stuff like jitter being audible, even though it's typically 120+ dB down, I know they're imagining things.
The best null I was able to get with an acoustic test going through loudspeakers was about -50 dB. You can download the original and nulled files here:
The Truth About Record Levels
--Ethan
http://www.britannica.com/EBchecked/topic/494815/reductio-ad-absurdum
reductio- ad-absurdum
Edit; This post has no relationship with the one above. It is replay to to the "Strawman Argument."
reductio- ad-absurdum
Edit; This post has no relationship with the one above. It is replay to to the "Strawman Argument."
It's funny how things you swear you heard, you know you heard, not only disappear during blind listening, but after blind listening. When you no longer expect to hear it, you don't.
Tim
That's not really the question. It supposes we passed the "blind test." Then we heard it but still can't measure it. What then?
Greg
Then we go along trying to find ways to measure it and if we fail and it is consistently heard, we work with that .. Eventually we will find a way to measure if, since it exists ...
Similar threads
 | Steve Williams Site Founder | Site Owner | Administrator |  | Ron Resnick Site Owner | Administrator |  | Julian (The Fixer) Website Build | Marketing Managersing |