My statement you quoted earlier had nothing whatsoever to do with that. I was simply observing that completed DIY projects often cost more than their mass-market commercial alternatives, and that high-end market commercial gear typically costs more than DIY. That's all that I was saying.
I'd like to elaborate on what I meant by, subjective neutrality.
Speaking only for myself, the subjective playback characteristic which I rank most important is the illusion that the sonic event is real. I'm after verisimilitude first and foremost. The instinctive recognition that I'm hearing real live instruments and vocalists. The more realistic the playback sounds, the more sonic neutral, or natural it is for me. I suspect that anyone who has heard playback which sounded convincingly real will readily understand what I mean. I prize this charateristics above strict sonic neutrality, which I mostly define as the lack of tonal colorations.
Whether or not the tonality of the reproduction neutrally matches that of the original event, however, I have no way of subjectively judging anyhow without having personally attended the recording session. This, of course, is an argument in support of objective measurement. However, what I can satisfactorily subjectively judge is the effectiveness of the auditory illusion of the playback sounding live and real. This I can do without reference to having been present at the original recording session. The playback doesn't have to sound accurately real and live to the specific original recorded event (yes, ideally, it would) for it to successfully sound simply real and live. To accurately judge such realism would, likewise, require in person attendance of the recording session.
This is not to suggest that objective tonal neutrality is unimportant to me. On the contrary, I find it very important. It's just that I want it partnered with sonic realism, and if I must sacrifice one to obtain the other, I would unhappily sacrifice tonal neutrality. This, however, seems rarely a choice. It's infrequent that I find an playback system which sounds convincingly live and real. To be extra clear about this, I do not reject with the value of objective neutrality, it's simply subjugated in importance to subjective sonic realism for me. The best of all worlds would be convincing subjective realism combined with objective measured neutrality, while the worst combination would be would be an overtly colored sound which clearly sounds reproduced. It does seem likely there's some correlation between the two.
A parting related thought. It seems to me that an purely objective view implicitly presumes that measured perfection necessarily confers sonic perfection. Logically, it would seem that should be so. Subjectively, however, it seems me, that for one to judge the playback sonics as perfect also requires them to sound live . Assuming, of course, that the original recorded event was of live musicians and vocalists.
Speaking only for myself, the subjective playback characteristic which I rank most important is the illusion that the sonic event is real. I'm after verisimilitude first and foremost. The instinctive recognition that I'm hearing real live instruments and vocalists. The more realistic the playback sounds, the more sonic neutral, or natural it is for me. I suspect that anyone who has heard playback which sounded convincingly real will readily understand what I mean. I prize this charateristics above strict sonic neutrality, which I mostly define as the lack of tonal colorations.
Whether or not the tonality of the reproduction neutrally matches that of the original event, however, I have no way of subjectively judging anyhow without having personally attended the recording session. This, of course, is an argument in support of objective measurement. However, what I can satisfactorily subjectively judge is the effectiveness of the auditory illusion of the playback sounding live and real. This I can do without reference to having been present at the original recording session. The playback doesn't have to sound accurately real and live to the specific original recorded event (yes, ideally, it would) for it to successfully sound simply real and live. To accurately judge such realism would, likewise, require in person attendance of the recording session.
This is not to suggest that objective tonal neutrality is unimportant to me. On the contrary, I find it very important. It's just that I want it partnered with sonic realism, and if I must sacrifice one to obtain the other, I would unhappily sacrifice tonal neutrality. This, however, seems rarely a choice. It's infrequent that I find an playback system which sounds convincingly live and real. To be extra clear about this, I do not reject with the value of objective neutrality, it's simply subjugated in importance to subjective sonic realism for me. The best of all worlds would be convincing subjective realism combined with objective measured neutrality, while the worst combination would be would be an overtly colored sound which clearly sounds reproduced. It does seem likely there's some correlation between the two.
A parting related thought. It seems to me that an purely objective view implicitly presumes that measured perfection necessarily confers sonic perfection. Logically, it would seem that should be so. Subjectively, however, it seems me, that for one to judge the playback sonics as perfect also requires them to sound live . Assuming, of course, that the original recorded event was of live musicians and vocalists.
How's does it border on being insulting? There are plenty of tube emulators that work that way - and some of them sound pretty good. Nelson Pass's H2 Harmonic Generator does the same thing except using a JFET. Nelson handed out a bunch of H2 kits at Burning Amp last year and nobody seemed to take offence to it.
Tom
I think that the source(transducer or digital recording) and the Speakers are the ones responsible for 99.9999999% of that natural sound and the electronics only have to be resolute and powerful enough to pass that information from source to speakers.The fact that some electronic equipment can create the impression that the speakers or the source are accurately reproducing the real thing is just an endless discussion favored by audiophiles on one side and master recordings engineers on the other.
I sent you a private message regarding your previous comment as i don't want to pollute this subject more than i already did.
That's subjectivism for you.
One man's ceiling is another man's floor.
Jan
It was intended as acknowledgement that others may have validly differing playback/listening priorities.
Jan, I'm uncertain. Of course, objective determination is inherently instrumented determination, which is what we have now. If you are asking me about an objective measurement of subjective perception, then I don't know that we will find a solid method. Part of the problem would be determining a way to quantify subjective perception. To accurately communicate it, particularly to consumers. Right now, we only have the flowery prose of most hifi magazine reviews to guide us on this matter. Perhaps, one day, there will be an instrument which accurately predicts the subjective listening experience. I know, you'd say that your test bench is full of such instruments.
I'm no expert on double blind-testing, nor a statistician, however, double-blind testing is probably the closest thing we have to an scientific method for testing subjective human perception. I have always wondered about one issue when seeing test results from a group of subjects. Depending on the nature of the questions, there can be a normal statistical distribution of results. A normal distribution has variance. A variance in subject test response could also signal an actual variance in subject perception.
Indeed, it may indicate that not EVERY test subject perceived exactly the same thing, or did not perceive it to exactly the same degree. The temptation is to dismiss such outlier data points as false, but they are not necessarily false. Such variances are eaily computed into a standard deviation which, assuming the distribution is normal, can then be used to establish a confidence figure for any given data point. However, a figure of low confidence isn't a figure of zero confidence.
It seems possible that some of the test subjects actually had a true difference in perception? I suppose, one way to remove that question would be to test the entire group multiple times, and look for positive correlations in the data points of outlier test subject responses across the many tests. I don't know.
Last edited:
Can't say I disagree with the gist of your post. But maybe we must accept that personal preferences are just that, and not necessarily any indication of reproduction accuracy. As long as some people swear by single-ended tubes, or massive class A, or optimally-biased class B, or hybrid amps, all with different degrees of accuracy, maybe we have to conclude that audio is a fashion industry and there's a market for anything.
That would mean that in the design process you need to accurately identify what aspects of sound reproduction are desired by your target market, and design accordingly, not necessarily for maximum transparency. Something Nelson Pass, to name just one, is extremely good at.
You also see it in this diy world, anything goes, and it is always touted as the worlds' best sound. Which is a logical fallacy because my amps are the worlds' best
Jan
That would mean that in the design process you need to accurately identify what aspects of sound reproduction are desired by your target market, and design accordingly, not necessarily for maximum transparency. Something Nelson Pass, to name just one, is extremely good at.
You also see it in this diy world, anything goes, and it is always touted as the worlds' best sound. Which is a logical fallacy because my amps are the worlds' best
Jan
Last edited:
Ken, I think it's hard to draw a line and say that something is inaudible overkill unless you want to take a hard-line approach and say that everything "decent" sounds the same. That's fine (and that's probably where I am at here), but we have people in some threads saying that phone DACs sound fine, and some in others complaining about -100 dBc phase noise at 10 Hz offset and other phenomena barely measurable.
For every person that thinks they know what makes good sounding electronics, there are a bunch people who prefer the exact opposite. This is what bothers me from a logical standpoint about "subjectivists" that start to claim measurement x or y are irrelevant. You can pick a measurement (THD, IMD, jitter, etc.) and I can go find you people who disagree from this forum in about 5 minutes.
For that reason, I think it's best to appreciate technical posts for what they are and not necessarily from an audibility point of view. It's a slippery slope and a lot of people seem to have opinions that do not align with others.
For every person that thinks they know what makes good sounding electronics, there are a bunch people who prefer the exact opposite. This is what bothers me from a logical standpoint about "subjectivists" that start to claim measurement x or y are irrelevant. You can pick a measurement (THD, IMD, jitter, etc.) and I can go find you people who disagree from this forum in about 5 minutes.
For that reason, I think it's best to appreciate technical posts for what they are and not necessarily from an audibility point of view. It's a slippery slope and a lot of people seem to have opinions that do not align with others.
This is where commercial high-end gear is most intriguing for me from a technical perspective. Designing to the subjective desires of a target market still requires the designer to translate those subjective desires into a set of objective engineering parameters. Knowledge of exactly which parameters then are key, and how to engineer them to obtain the desired subjective result, is the trick for which many dollars are paid by high-end customers.
That presumes they're audibly different in a way that appears obscure to rigorous instrumented tests. Way too many transistors have flipped states over the sound vs measurements debate...
*Excepting those designs which aim for distortions, whether harmonic or in amplitude/frequency response.
*Excepting those designs which aim for distortions, whether harmonic or in amplitude/frequency response.
Last edited:
Chris, I've always suspected that many reports of differing perceptions aren't truly that, but are actually differences in auditory observational skills. Such skills are developed by long experience simply listening to hifi music reproduction systems. Perhaps, notions that iPhones are hifi, and that close-in phase noise is audible, is a dichotomy of perception explainable by differences in listener observational skills/experience.
I essentially agree with your observation. I suspect that the number of people in command of the technical formula to repeatedly obtain audio gear which produces some particular, subjectively desirable sonic template are relatively few. There are many supposed experts around. Since objective perfection (as far as human perceptual limits are concerned) can be found even among mass-market gear, so, it would seem logical that 'desirable' subjective high-end sonics are to be found then in some particular combination of objective imperfections.
Perhaps, that is not only logical, but also true. However, I'm having difficulty in accepting that we can essentially have widespread objective technical perfection, yet rarely do I hear playback which subjectively sounds convincingly like real or live music. Often, far from it. Why is this? Perhaps, there is not enough focus on the dynamic interaction of certain technical parameters. Or, perhaps, we are simply being too limited by our current 2-channel stereo playback paradigm, dominant since the late 50's. I'm sure Ambisonics fans are raising their hands.
Chris, I've always suspected that many reports of differing perceptions aren't truly that, but are actually differences in auditory observational skills. Such skills are developed by long experience simply listening to hifi music reproduction systems. Perhaps, notions that iPhones are hifi, and that close-in phase noise is audible, is a dichotomy of perception explainable by differences in listener observational skills/experience.
I essentially agree with your observation. I suspect that the number of people in command of the technical formula to repeatedly obtain audio gear which produces some particular, subjectively desirable sonic template are relatively few. There are many supposed experts around. Since objective perfection (as far as human perceptual limits are concerned) can be found even among mass-market gear, so, it would seem logical that 'desirable' subjective high-end sonics are to be found then in some particular combination of objective imperfections.
Perhaps, that is not only logical, but also true. However, I'm having difficulty in accepting that we can essentially have widespread objective technical perfection, yet rarely do I hear playback which subjectively sounds convincingly like real or live music. Often, far from it. Why is this? Perhaps, there is not enough focus on the dynamic interaction of certain technical parameters. Or, perhaps, we are simply being too limited by our current 2-channel stereo playback paradigm, dominant since the late 50's. I'm sure Ambisonics fans are raising their hands.
At risk of derailing the thread further...
I do tend to think the stereo paradigm and speakers + room interactions are much more limiting than the electronics. I haven't heard Ambisonics or the SVS Realiser, but I find binaural recordings on good headphones to be a much improved experience.
In terms of this composite op-amp, it's really not too expensive to build. For a good number of diyers I would actually think their own time costs more than the BOM on many projects.
Agreed. I'm going forward with incorporating it in to the analog section of an experimental DAC I'm designing.