I can evade any "subjective" test with a quote that runs something like, "well, I can hear a difference, so maybe it's your measurement that is wrong". It's the ultimate cop-out--no matter how controlled the test, you can make this excuse. Of course, there's entire fields of study that tackle these problems and metrics of confidence change based on the field (APS using 5-sigma to define a discovery, for example).
All the "holes" you're ostensibly poking in proposed testing protocols run exactly this gambit. You've raised them elsewhere, and the whole, "everything is unknowable, so I just trust my own perceptions," language gets old fast.
I'll propose 2 much more useful hypotheses:
1.) The circuit effects are audible.
2.) The circuit effects are uniquely audible due to their location in the signal chain.
No, there's no way any blinded listening test (single or double) will achieve the sort of confidence we would want to hang our hat on. It'd certainly be interesting if SY has the ability to do a single blind measurement protocol, even though I fully expect that he would test each box equally. It'd certainly be easy enough to ship SY two boxes labeled "A" and "B" and it remain blinded as long as SY doesn't take off the lid.
Or, probably more valuable, would be a single-blinded evaluation of the measurements themselves. I.e. SY (or whomever) emails his captured data, along with protocols/descriptions to another party to do the analysis. All of our clinical samples are done in this fashion--sometimes we'll look at the raw, blinded data and try to figure out which population is which, but only after sending the results back upstream for the stats that will actually come out of the study.
Hi Coris.
Your subjective appraisal points to change in subtly audible errors.
Reducing jitter could describe your findings.
Adding low levels of lowest order harmonics can also increase perceived focus, positioning and impact of individual sound sources without significantly affecting system low level white noise.
IIRC some measurement graphs posted recently showed increase in lowest order harmonics ?.
You obviously have suitable modifiable DAC board running.
Would you like to rerun some experiments for us all ?.
1 - Revert schematic to original/normal configuration, take a listen.
2 - Add 1uF shunt caps, take a listen.
3 - Remove 1uF shuntcaps, take a listen.
4 - Add 0.33F supercaps, take a listen.
5 - Remove 0.33F supercaps, take a listen.
6 - Add 22uF SMD, take a listen.
7 - Add 0.33F supercaps, take a listen.
8 - Add 1.0uF shunt caps, take a listen.
Write notes at the end of each trial.
You get the idea, listen to each component change individually, then listen to each of the permutations available with the three components in question, but revert the schematic to original condition between trials in order to reestablish/reset hearing memory.
This will take some time on your part but would be much appreciated and ought to help the discussion.
If you are able to provide measurements of each experiment would be very useful also.
Dan.
Chris, you are defending a certain orthodoxy, I get that. But it's also stiff and rigid, way too much for my liking. I honestly think you are wrong. It's also unnecessarily personal.
I walk the talk every day and I give value big time - just ask those who know me and how generous I have proven myself over many years. I get a lot of pleasure and satisfaction. Once again I am being judged through the prism of money? That's really sad. No, once I disclosed what I knew up to that time and invited further discussion, that was accepted by fair people as the fair thing to do. That's my decision and it should be respected. I don't judge others and - we are all free to do what we think best. I have no regrets, but I am surprised by the reaction.
I come from a background where we were allowed to say the craziest things, mostly all wrong, but you know what? Crack open a lot of shells is messy business, but then you might find a pearl.
Having said that, I do believe that the newer posts seems to be a little more civilised...?
Last edited:
Joe, I'd like to know how one listens to the effects of jitter at one 1/2 a cycle/day. I don't have any 12 hr. songs.
Where do i say that? Unfortunately we currently have no objective or subjective tests that do not have some problems.
We can get to a certain point with measurements, but they are not proof. We can only get so far with subjective tests, but they can only prove 2 DUTs are different. And they have lots of places where the test becomes invalid.
I had similar thots on this current situation.
And i have executed similar blind tests wrt EnABL. Test situation set up, with individuals running the test by themselves blind. These test results were disregarded by most because they were not formal. Not by the people that heard the difference (most did, some did not)
dave
The results of any music listening evaluation is necessarily partly a function of music selection, system components, listening room acoustics and familiarity with the sound character of the three. I had also observed soundstage and imaging benefits, but simply choose not to focus on those in my brief subjective report, which is a paragraph in length.
I'm not surprised that Coris' subjective descriptions might differ from mine, partly because the factors listed above are certainly different for him. I think that the key point is that we both would describe the effect as being subjectively beneficial to the music. That's the essence of what's been reported. Note that I made no statement about the sound being more accurate to the source. I don't know whether it is or isn't.
Last edited:
Scott, I might agree that you have a point if this were a matter of many having conducted the experiment and reporting a null subjective result. However, that's not what we have here. Here, we have many doubting the subjective reports of those who've conducted the experiment, yet steadfastly refuse to find out for themselves, which is quite a different state of affairs.
Yeah, I am shopping for that and using ARTA. Seems it will 24/96 as they are reasonably priced down here.
I do have a ClioFW 24/192 Spectrum Analyser with its own prop. software and will use that too - many think of it as just for speaker designers (many consider it the best there is and not cheap), but it is much more than that. THD <0.002% and THD+N <0.007% simulating the ARTA chapter 2 test.
But I need to use ARTA and let others here do their own confirmation/parallel testing. Again, I think that would be great, but it comes down to whether we can catch this thing. It must come down to generating the right stimulus, and I honestly think we should pull together on this and less arguing getting nowhere.
I think if we could capture the 'event' that has been perceived, then it would be a slam dunk. As Steve has said, you can't argue hard data. Even Scott has rightly honed in on the target -1dB @ 20KHz, there is some kind of 'corner' or 'knee' happening. We trust our ears because even skeptics have heard it - and scratched their heads.
If we could achieve that, then I have greater confidence that any listening test that is derived from a known positive, is more likely to succeed, because we won't be testing for fairy floss. This is despite my agnosticism towards double-blind listening tests.
Let's explore this a bit further. Suppose that Joe were to present us all with the details of a valid scientific listening test which conclusively showed the effect to be plainly audible. Would that be sufficient evidence for you to accept that reality, or would you then want to conduct the experiment to prove it for yourself first?
THD of -94dB. Doesn't sound like 24bit ADC performance unless I'm missing something (and I often am)
If I'm lucky, I'll be able to do my sims before it's shut down. Might learn something from that!
It's another entrenchment of different philosophies buoyed by the fact that we're discussing luxury goods.
You would what, fully accept the reality, or then want to experimentally prove it to yourself first?
- Status
- Not open for further replies.