I don't read audiophile revues, so, I don't know what is the fashion about DACs, but I noticed differences between them since a long time. Well, it seems they are, at least, in their dynamic behavior. It looks like it is mostly, but not only, in the analog part of them:
"Its a fact of life that sadly the array resistors on the ESS DAC's suffer from "Poor" Voltage coefficient (that is there resistance value changes with voltage across them) - VC modulation is seen as ODD Order distortion products (Change of "Gain" over cycle) - as the distortion is Odd order in nature it is not cancelled by differential summing." ©JohnW
That could explain my recent found that, sometimes, some recent DACs sounded "too much" in their dynamic behavior, making some of them impressive but fatiguing ?
Anyway, nothing I know sound exactly the same, nothing in real world is never perfect, preamps, integrated circuits, caps and even resistances. It is all the fun about electronic design and has nothing to do with fashion. Sometimes, we can measure why, sometimes, we don't know how. It can depend of various factors and what we are looking at.
As long as we will continue to measure standing waves, we will miss part of the landscape. A BIG part.
As long as we will not have a real knowledge of the human audibility thresholds, the measurement numbers will not tell so much.
All this being, of course, a personal opinion, that I believe is shared by lot of people and obvious in my book.
Hats off ! ;-)
Last edited:
"Fairly accepted", yes probably, but it doesn´t ensure that all parties agree on the meaning per se. So synchronizing any describing vocabulary is an essential part when doing qualitative research or communication about reproduction of music.
I don´t think the "scientific crowd" is generally freaking out, certainly not the people who are doing research on this topic.
@ PMA,
I'm still interested in your comment:
"Oh, that one.
Thanks for your answer, but did i misread your post about it?
I'm asking because Ashihara did not show that the reason for the audibility of any high frequency content (like in the Oohashi et al. paper I've mentioned) was an intermodulation effect."
@ Evenharmonics,
Speaking of strange; weren´t you the Evenharmonics that just recently wrote about things that allegedly happened in other forums, but wasn´t able to bring the evidence?
I'm still interested in your comment:
"Oh, that one.
Thanks for your answer, but did i misread your post about it?
I'm asking because Ashihara did not show that the reason for the audibility of any high frequency content (like in the Oohashi et al. paper I've mentioned) was an intermodulation effect."
@ Evenharmonics,
Speaking of strange; weren´t you the Evenharmonics that just recently wrote about things that allegedly happened in other forums, but wasn´t able to bring the evidence?
He did. The content with both <22kHz and >22kHz frequencies was distinguishable from that with only =<22kHz file only in case that it was tested through speaker with one driver. Not distinguishable when tested through a 2-way speaker with 2 drivers.
This is well known to us who make real tests in a real life. Make a 19kHz+20kHz test, of course you need low distortion amplifier, through a good 2-way speaker with crossover that would not let the signal to get in the woofer and a person of our age will hear nothing, provided that amplifier output has reasonable level and is not overloading the tweeter. Do the same with distorting amplifier or one wideband driver without crossover and you hear 1kHz intermodulation clearly.
If you have more questions, I will be back on Sunday evening, as I am leaving just now.
Thanks, but I can read and been there done that.
But, Oohashi et al. already considered the intermodulation problem and therefore designed and used a speaker in their experiment which reproduced the high-frequency content (above ~21 kHz) with a separated ribbon tweeter (ribbon tweeter, afair).
Further, it is well known to us, who use real scientific reasoning (SCR), that one has to really carefully analyze two experimental designs (and executions) before concluding that one invalidates the results of the other if it was not a replication.
And a replication Ashihara's experiment was not, in fact, it was a very different experiment, and so one has to compare the high-frequency content, has to compare the level of the high-frequency content and has to consider the results of the auditory evaluation, especially when evaluating the reproduction of the high-frequency content alone.
What Mallinson is saying (& JohnW also) is explained more fully in this (written by Mallinson & Dustin Forman)
As you can see he is saying the design process is not appropriately solving all use cases & the anomalous behaviour is perceived by some listeners
It remains a mystery? Perhaps it's not important.The topic of "emotion" in reproduction? What's the synchronized vocabulary on that?
More from that ESS paper above to provide some context
We hear all this talk about the perfection of well designed DACs & the accuracy of the design & modelling process (wishful thinking, taught in schools) but when real information is provided from DAC designers we see the picture is not the perfect vision it's made out to be.
So who's going to be first to ignore all of this technical info & instead talk about how could anyone trust information coming from DAC designers involved in commerce?
We hear all this talk about the perfection of well designed DACs & the accuracy of the design & modelling process (wishful thinking, taught in schools) but when real information is provided from DAC designers we see the picture is not the perfect vision it's made out to be.
So who's going to be first to ignore all of this technical info & instead talk about how could anyone trust information coming from DAC designers involved in commerce?
Howie, thanks for sending me down on memory lane. I used these for my main stereo while being a student. They played VERY loud. But the tweeters burned out easily.
Yet art exists, symbols move the world, fashion designers work. And a jeweler sells a precious ring without trying to convince us of the exceptional performance that that finger will get (but nevertheless jewels work).
I believe this signifies the great problem in digital audio - the lack of understanding about how the 'digital concept' is implemented in the real world. (Remember digital is a concept, not a physical thing - it has to be implemented in physical things which don't behave digitally) The divorce from physical reality by relying on mathematical models leads to the sort of confusion seen here
As Mallinson said they plotted one state variable - that doesn't mean they can model the whole system, mathematically[/quote]
Maybe this paper from Philips research explains it?
In which it states that
Note PRAGMATIC doesn't mean accurate - just good enough.
And when evidence from listening impressions that 'just good enough' isn't applicable in all cases what is the reply? Do we go back & look at the underlying assumptions inherent in the design process or do we ignore all such listening impressions as "delusions" (Mallinson, an engineer, says he can't hear it but he trusted those who could & discovered the problem in measurements)
IMO you guys are talking about two different things a 1 bit SDM (or modulators in general are not purely digital systems, an FIR or IIR filter is.
The modulator front end is not a "digital concept". I took a lot of flak years ago (almost 30 by now) for daring to suggest a two chip converter with SOTA bi-polar integrators and comparators glued to CMOS logic. Talk about not politically correct.
More from that ESS paper above to provide some context
We hear all this talk about the perfection of well designed DACs & the accuracy of the design & modelling process (wishful thinking, taught in schools) but when real information is provided from DAC designers we see the picture is not the perfect vision it's made out to be.
So who's going to be first to ignore all of this technical info & instead talk about how could anyone trust information coming from DAC designers involved in commerce?
The problem is, on this thread there are a few real experts. And then there are people that think they are real experts and know everything about everything when in fact, unlike the real experts, they know only a little bit about a few things.
As a famous politician once said 'because as we know, there are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns—the ones we don't know we don't know.'
As Einstein is purported to have remarked ' . . . some things can be counted that don't count. And some things that cannot be counted do count . . .'
Last edited:
The heart of all SD DACs (whether one bit or not) is the nose shaping modulator. Maybe Chris is taking about something else not mentioned in Mallinson's paper & presentation - I'm not - I'm talking about what JohNW & Mallinson wrote - that the noise shaping modulator in SDM may not be unconditionally stable.
So I don't get your point SW?
Last edited:
Indeed, being aware of one's own limitations & the limitations of tests/measurements is all most of us expect for an honest & open discussion. So often that is not on display here - rather it's a display of ego/arrogance usually done in the guise of science/engineering - very far from scientific thinking where truth is the goal
Last edited:
- Status
- Not open for further replies.