ABX testing was mentioned and i have to say, that even tho abx testing works in logical level, and it is nice to test things once in a while. But ABX testing is
1) Just pain in the *** to set up correctly and also implement.
2) Convincing someone with ABX test is useless (you can just lie, or the "someone" is a skeptic). They won't believe it if the result differs from what they expected or they will just ignore it or they will say you didn't do 1) right.
3) Doing ABX other places but at home is useless. At least my brain is negating the acoustics of my listening room much differently than rooms unknown to me.
4) Playing same piece over and over again does not produce very good results. Try to listen to Diana Krall and then just a-b a part of of the song. It gets really annoying really fast. Well, that already tells the difference is small, but small difference can matter and some of the differences are be lost in the irritation.
Only thing that really matters is that the difference that is heard in normal A/B testing does not fade away. Just testing the equiment at your home with the previous device present is much more productive than doing ABX. I have found that if i have to really find something nice about the device, maybe its not for me and when changing back to my own device, if that produces just a good feeling and i don't miss the tested device, then i can live without it.
But i understand this is different to a hifi engineer than to hifi consumer. But i am no hifi engineer. I know the basic priciples, but nothin more. For me, if it sounds better, its better.
1) Just pain in the *** to set up correctly and also implement.
2) Convincing someone with ABX test is useless (you can just lie, or the "someone" is a skeptic). They won't believe it if the result differs from what they expected or they will just ignore it or they will say you didn't do 1) right.
3) Doing ABX other places but at home is useless. At least my brain is negating the acoustics of my listening room much differently than rooms unknown to me.
4) Playing same piece over and over again does not produce very good results. Try to listen to Diana Krall and then just a-b a part of of the song. It gets really annoying really fast. Well, that already tells the difference is small, but small difference can matter and some of the differences are be lost in the irritation.
Only thing that really matters is that the difference that is heard in normal A/B testing does not fade away. Just testing the equiment at your home with the previous device present is much more productive than doing ABX. I have found that if i have to really find something nice about the device, maybe its not for me and when changing back to my own device, if that produces just a good feeling and i don't miss the tested device, then i can live without it.
But i understand this is different to a hifi engineer than to hifi consumer. But i am no hifi engineer. I know the basic priciples, but nothin more. For me, if it sounds better, its better.
Last edited:
Depends on what you are testing.
Which is why scientist make such a fuss about independently reproducing results. But having a neutral third party run the test, and have witnesses that verify the testing, is another way.
But in any case, that is not really an argument against ABX - or are you saying people are more likely to believe your subjective claims than your claimed ABX results?
Again, not really an argument against ABX - the same argument applies against subjective "testing". Basically you are saying "I can not hear differences if I am not at home". The method doesn't really affect that.
Then don't.
If your test setup allows instantaneous switching, you only need to listen once through the sample (or as many times as you like) - just keep switching back and forth at will.
So, to rephrase:
1) can be a pain in the *** to set up corretly.
2) Some people might not believe me, no matter how I do the test, unless independently verified
3) some people have to listen at home - doesn't matter if it is sighted or blind
4) For some people, playing same piece over and over again does not produce very good results, so they shouldn't. No matter what the listening method.
I agree for the most part, but...
1) But ABX testing amps is in the pain in the *** category to set up. Testing, say mp3 vs lossless is not. But that is basically the only thing that is not and in some cases testing balanced vs single ended inputs. Testing cables is arguable (please don't). Internet + independent party both tester and "skeptic" is happy with = won't happen
2) Some -> most. I have yet to see a skeptic that would believe an ABX result not foreseen. Same goes for subjectivists too.
3 + 4) ABX is just not very effective way to require for proof of things when it comes to hifi. You need some dude both parties trust, but who is unbiased. Then you need to do the testing in someone's home. Then you need some people, who do the testing right and smooth and relatively gapless. Even then the end result could be just a dud.
This can be done, but it requires arranging, time, effort and sometimes money too, to do right. Why bother? You could just have a beer and listen to music without all these shenanigans.
1) But ABX testing amps is in the pain in the *** category to set up. Testing, say mp3 vs lossless is not. But that is basically the only thing that is not and in some cases testing balanced vs single ended inputs. Testing cables is arguable (please don't
). Internet + independent party both tester and "skeptic" is happy with = won't happen2) Some -> most. I have yet to see a skeptic that would believe an ABX result not foreseen. Same goes for subjectivists too.
3 + 4) ABX is just not very effective way to require for proof of things when it comes to hifi. You need some dude both parties trust, but who is unbiased. Then you need to do the testing in someone's home. Then you need some people, who do the testing right and smooth and relatively gapless. Even then the end result could be just a dud.
This can be done, but it requires arranging, time, effort and sometimes money too, to do right. Why bother? You could just have a beer and listen to music without all these shenanigans.
Last edited:
Well I have experienced the result of not enough current when I connected Sonny's Mirand Discrete output stage directly to the unbuffered inputs of the NC500's. This results in a very thin sound, and when pushed starts to harden. I wouldn't use any IC or discrete opamp with less current capabilities than the LM4562. Look at that opamp as an example of bare minimum.
Now, now, never underestimate the power of the Sheddi.
Nice one julf
Nice work boggit, very handsome. I'm looking forward to seeing your final-final.Did you do any measurements to find out exactly what was causing the hardening? Was it clipping, or something else?
Power amps and speakers are the hardest ones. The rest can pretty easily be done with an ABX switching box.
You are making a lot of assumptions about the "skeptic" - in my experience, it is usually it is the subjectivists who won't agree with the results, and usually because of the same "the test circumstances didn't allow the differences to be heard" reasons paranormal phenomena fail to materialize when a sceptic is in the room.
An independent supervisor is one way of dealing with that issue, if you worry about being believed. But are we here to convince people to believe us, or are we here to learn? I find ABX extremely useful just to test my own assumptions, so no need to share the results - but sometimes sharing the results is very useful.
Solid, verifiable and repeatable evidence is only required when someone is making claims that go totally against well-established theory and engineering experience. "Extraordinary claims require extraordinary evidence".
Unfortunately it is pretty much the only way we can remove the effect of (unconscious) biases - and those biases are by far the most significant component in "high-end" audio.
Sure - and that is what most of us do (the beer and listening part). The problem is when somebody, after too many of the beers, claims that pasting a magic lacquer on your transistors makes them sound like tubes.
No I didn't measure it. It was most likely clipping at high output. It just didn't sound as good when it was flat out. Much better with the additional buffer stage in the path.
Yes, most probably. Which is why "opamp rolling" without properly understanding the characteristics of the opamps is somewhat futile.
It's not a mystery that this output stage didn't have enough current. It's was known all along. But luckily most discrete opamps also have datasheets to confirm if they have enough current capabilities or not.
It was known by the ones who actually read the data sheets. How many "opamp rollers" actually do that, instead of relying on random recommendations off the Internet?
Sure, but current capability is only one of the many relevant parameters (such as open-loop gain, bandwidth etc.). I have seen amplifiers being turned into quite efficient high frequency oscillators by people who didn't understand phase issues...
Naw- short term preference is a poor judge of performance or even long term preference- hence so many "smiley face" (boom and tizz) frequency response speakers that sell well but people tire of rapidly once they have to live with them.
Early in my time in the hobby I did everything by ear. It worked out pretty well but I definitely wasted time and money by not going the engineering oriented path. Having an oscillator can blow tweeters, etc, and not necessarily right away. It's important to do the homework, especially given that the philosophy of NCore is extremely engineering-oriented.
Sounds familiar - did most things by ear, but also listened to a lot of the "conventional wisdom and experience" that turned out to be mostly hearsay, superstition and old wives tales. Fortunately I started out when there still were proper, rational general publications (such as "Wireless World") available, where the Brunos of their day (Williamson, Baxandall, Linsley Hood, etc) actually explained the rationale of their designs in engineering/scientific terms.
Going to university and studying electronics was an eye-opener, but the real eye-opener was looking at both sensory psychology and pseudoscience and belief systems, and realizing how fallible we human beings are.
Been there, done that, but didn't keep the burned-out tweeters
Hypex really addresses the pro/OEM market (and, fortunately for us, the somewhat technically oriented DIY community), but it is interesting to see that even Mola-Mola, the more "high-end" oriented venture of Bruno and Jan-Peter, states on their front page "I analyse every subcircuit mathematically and look for ways to eliminate every error term. When the practical circuit measures as predicted, I listen to search for unexpected sources of colouration. These are then included into the maths and the whole process repeats". Sounds like the scientific method to me...
hypex ncore
No. But you are referring to General opamp rollers. If they enjoy what they hear, then what's the harm?
One could spend their money on cocaine and prostitutes instead if they wanted for entertainment. Which is more harmful?
I'm sure Boggit will ensure the circuit is appropriate for the opamps he recommends before offering his buffers for sale on the commercial market.
No. But you are referring to General opamp rollers. If they enjoy what they hear, then what's the harm?
One could spend their money on cocaine and prostitutes instead if they wanted for entertainment. Which is more harmful?
I'm sure Boggit will ensure the circuit is appropriate for the opamps he recommends before offering his buffers for sale on the commercial market.
If it is only for themselves, then no harm. If they go on the net, and recommend their choice to others, who fry their amps and tweeters, it is a different story.
hypex ncore
I guess it will be a good lesson. Life will go on.
The discrete opamps I tried with the NC500's were outstanding.
I guess it will be a good lesson. Life will go on.
The discrete opamps I tried with the NC500's were outstanding.
Last edited:
So the lesson is "don't believe everything you read on the Internet". But isn't the implication also that we shouldn't believe anyone that bases their argument on "trust me, I know what I am doing" without being able to support their view with rational justification and evidence?
Well this is diyaudio. So you have to pick and choose who to take advice from. At times you may even need to take risks. But this is all part of the learning process and DIY experience. If you want to play it 100% safe, buy a finished commercial product and call it a day.
- Status
- Not open for further replies.