He is and it's all kind of funny in its own way. Some people forgot about the thread on WBF where someone asked Ethan what his favorite CDs were. Ethan replied that he really didn't have a list and that since he improved his system there were only actually 2-3 CDs that he could listen to and enjoy.
Why Tube Amps Sound Different (and better) Than SS Amps
- Thread starter mep
- Start date
- Status
He is and it's all kind of funny in its own way. Some people forgot about the thread on WBF where someone asked Ethan what his favorite CDs were. Ethan replied that he really didn't have a list and that since he improved his system there were only actually 2-3 CDs that he could listen to and enjoy.
But when they can't refute things with logic and evidence, all that's left is personal attacks. I wouldn't even mind if they quoted me accurately!
Thanks Amir, your post was badly needed.
--Ethan
I have explained here countless times that nulling reveals all differences between (for example) the input and output of an amplifier, including things you might not think to look for or know how to measure. Yet here we are with the same old tired argument that "science" doesn't know enough about fidelity.
--Ethan
zzzzzzzzzzzzzzzzzzzzzzzzzzzzz.
Mr. Winer,
With all due respect, it appears there are many knowledgeable, experienced folks who disagree with your assertions.
Why can't you just accept that fact and move on?
Just because you state something countless times doesn't make it "true".
Politicians do it all the time.
GG
Mr. Winer,
With all due respect, it appears there are many knowledgeable, experienced folks who disagree with your assertions.
Why can't you just accept that fact and move on?
Just because you state something countless times doesn't make it "true".
Politicians do it all the time.
GG
I'm not sure people dispute the idea of nulling being useful, but there is definitely disagreement about what the difference signal means and when or if it becomes insignificant (which disagreement should be clear to you by now, it's been stated and discussed enough times here).
Maybe because there are knowledgeable experienced people here who do agree with him. And also have done some of those experiments which is one of the reasons they agree. And what he says is very reasonable.
Well when the result becomes there is nothing left to hear, as in play the difference result and hear nothing well the result has become insignificant.
The rule of thumb would be if you cannot hear a differenced result at normal playback levels (and there still may be a difference that is measurable), then one can have complete confidence that when any remaining difference is among music being played back it will surely be masked. There is a gray area where a low level difference would be perceived when replaying a differenced signal that also would be so low it will be fully masked in normal playback. The exact dividing line there is open to some debate.
And those that obviously don't.
Seems like this discussion is the equivalent of a dog chasing its tail. No?
Woof!
Seems like this discussion is the equivalent of a dog chasing its tail. No?
Woof!
In the digital realm you can achieve a perfect infinite null. Once you hit the analog realm there is no perfect null. There is always thermal noise even with no signal. So nulls are describe by how deep the null is versus the signal being nulled. If you take a 1 khz tone at full zero level, subtract one from the other and the remaining residual is -80 db, then you have an 80 db null. If you play that signal difference back over your system you almost surely will not hear it. Now a 60 db null you likely will hear faintly (though likely a difference that small would not be heard among the full signal). A mere 30 db null you will hear and that level of difference is fairly likely to be heard between two signals even played back at normal loudness. That is why there is no sharp dividing line. Some say 60 is enough for audible equivalence, some more even 80 db. But if the difference by itself is too low to be heard you can be sure it is even less possible to hear it when the signal is played back at normal levels.
Don't know. Haven't seen anyone claim absolute perfection that way. Now if we are talking audible perfection it isn't a straw man at all. In things like the Carver challenge the claim wasn't that either amp was perfect. Only that his amp could sound audibly identical to the other amp. It was to the satisfaction of those listening, and if I recall his measurements in a nulling procedure showed a null of -70 db. Something interesting would be how deep a null you get between two amps of the same kind. I believe some fair number wouldn't match that closely to each other. I know quite a few amps don't do any better than that from one channel to the other in stereo versions. Not many people complaining, "oh this amp is wonderful on the right channel, but something is sub-standard in the left." I believe 70 db nulls channel to channel are probably audibly indistinguishable.
If you look at the measurements that JA does on preamps and power amps both SS and tube, there are usually differences between the channels. The SP Carver challenge is somewhat confusing. After the first round was over, SP declared Carver won. After that something happened and they claimed they could hear a difference between Carver's amp and the CJ Premier 4.
I think the real problem is that some people try to oversimplify the design, manufacturing, and testing of audio gear. According to one of our resident experts, everything you need to know about any audio component is contained in 4 measurements. The reality is that audio gear is built using components that all come with a tolerance attached. The more expensive the gear, the more you would like to think that the manufacturers are closely matching each resistor, capacitor, and transistor. In the case of gear built on circuit boards, the layout between channels may not be symmetrical which means the traces between the channels may not be equal. There is a reason that gear usually has differences in measurements from one channel to another. If you buy cheap gear that has great specs on paper, how much money do you think they are really spending on tight tolerance parts and then matching them?
I think the real problem is that some people try to oversimplify the design, manufacturing, and testing of audio gear. According to one of our resident experts, everything you need to know about any audio component is contained in 4 measurements. The reality is that audio gear is built using components that all come with a tolerance attached. The more expensive the gear, the more you would like to think that the manufacturers are closely matching each resistor, capacitor, and transistor. In the case of gear built on circuit boards, the layout between channels may not be symmetrical which means the traces between the channels may not be equal. There is a reason that gear usually has differences in measurements from one channel to another. If you buy cheap gear that has great specs on paper, how much money do you think they are really spending on tight tolerance parts and then matching them?
Dear elsdude,
Just read your definition of an Audiophile.
How about this.
Someone who enjoys listening to and emotionally connecting with the music.
Fear?
Nah. Sounds like the mantra of a certain political party. Or the basis for many religions.
GG
Just read your definition of an Audiophile.
How about this.
Someone who enjoys listening to and emotionally connecting with the music.
Fear?
Nah. Sounds like the mantra of a certain political party. Or the basis for many religions.
GG
The null test is a very useful development test but IMHO is of no help for the OP subject under debate in this thread. Most of the time we are not able to correlate direct measurements with with sound quality in a straightforward way, except for a few very basic properties that can not by any means explain our preferences. Should we dream that just measuring the residuals of a null test will explain how the amplifier sounds?
There are few things in life that are as simple as some people try and make them out to be. Haven't there been plenty of amplifiers designed and built that look 'perfect' on the test bench using the standard tool chest of tests and yet they don't sound good? So how can something measure great and yet not sound great? Once you get past some basic measurements that do tell us something (like measuring for power supply noise that is riding on your signal, or poor frequency response that you will obviously hear, or a poor S/N ratio that you will hear, various types of high-level distortion, etc.) and you pass those basic measurements, all you have proven is that you have something that was competently designed and it works. It doesn't mean you have something that is going to sound great to lots of ears.
Well they were satisfied the Carver matched the CJ Premier during the challenge. There was no change of heart in some round two. Later they auditioned Carver's production amp supposedly based upon the CJ for that challenge. It was then that they said it was not quite a match. In production units I think Carver claimed a -43 db average null with the original, and the particular sample Stereophile's JGH auditioned turned out to have a -38 db null.
I heard a couple of the first model Carvers in question, and at the time owned the smaller non-Premier CJ. CJ used the same topology in all their amps. Not saying the CJ I had matched the Premier, but it would have been very similar. In my opinion that first Carver tube simulation indeed sounded tube like in the old school tube sense. Much like un-modded Dynaco 70's. It didn't seem to match the CJ I owned.
As for 4 measurements being enough, I don't know if I would quite go there. But almost for everything except power amps. If the 4 measures were also done using a speaker loads for power amps, even a general simulated one, and two devices were below certain thresholds on the 4 measures I bet there would be little or no difference. JA measures FR into a simulated load. Would be nice if he duplicated all the measurements into one.
- Status
Similar threads
