I'm going to be a bit rude, and throw in a piece again, by Allen Wright, that he wrote elsewhere:
That vinyl often gets this right without much effort put into getting it to happen, is just a happy circumstance of how that reproduction chain works ...
I emphasised the "most digital systems", because this failing is so often heard - but is totally unnecessary: the trouble is that most audio people don't know or understand how to fix this - and fixable it most certainly is - especially in RBCD.
That vinyl often gets this right without much effort put into getting it to happen, is just a happy circumstance of how that reproduction chain works ...
That is some pretty self serving BS as far as I'm concerned. I have never encountered evidence of a lack of low level details because a system is digital. I have seen many examples of analog recording systems showing the degradation of low level accuracy in the presence of a high level signal. Badly executed digital can have problems but following the datasheet on even the least expensive DAC or CODEC from a reputable company will yield dynamic performance that will easily exceed any analog recording medium, not to mention any known transducer.
Further, the claim that hearing doesn't show this effect is just plain wrong. The muscles in the middle ear are dynamic volume controls and absolutely control the level to match the current maximum dynamic. In cases where they are damaged the dynamic range is severely limited. And in any case the background levels at the sensing point will change a lot with music dynamics. You will perceive something different but its not necessarily what your ears are hearing.
Further, the claim that hearing doesn't show this effect is just plain wrong. The muscles in the middle ear are dynamic volume controls and absolutely control the level to match the current maximum dynamic. In cases where they are damaged the dynamic range is severely limited. And in any case the background levels at the sensing point will change a lot with music dynamics. You will perceive something different but its not necessarily what your ears are hearing.
Joachim, for the second time, do read the original O/L paper, and you will easily see that one of their key points was using only 20 dB of global NFB.
Their other point was having an open loop power bandwidth out to 100 kHz. So, with 20 dB of GNFB, if memory serves, their THD at 100 kHz was something like 0.5% or so. There's a graph showing it.
The original Electrocompaniet was NOT up to the job. They did some minor changes, but that was quite enough for them to fall behind the original. For one thing I reember, they changed the place and value of the VAS compensation capacitor.
This is not to say that the EC was a bad amp, rather it's to say that it was simply not as good as the original.
I on the other hand have heard this over and over again, it's been standard behaviour of systems over the years. I take recordings where I know what the low level detail reveals, having heard them repeatedly on decent playback - and play them on unknown, frequently pricey systems; and whole levels of information goes completely missing, it's as if a scapel has perfectly sliced off the underbelly of the sound - once the complexity of the sound builds up, it all turns into a bit of aural mess, becomes something not worth listening to ...
Edit: I'm fooling around with a decent digital keyboard right now, and it's easy to hear this happening - when not up to par, all the separation of the constituent sounds fails to occur, it just comes across as a bit of a muddled mess; as the instrument improves each separate strand stands out with its own identity, easily perceived in its own 'space'. Absolutely nothing changes during this, apart from the state of tune of the electronics ...
Last edited:
I have also heard this time and time again, as you say, on indecently priced systems as well as on low priced ones.
But I have been bushwacked by a few El Cheapo products you would never expect to work so well. Toshiba/Aurex SB-45 is a great example. It was the second or third product from the bottom in its day, this being circa 1982, rated at just 40W/8 Ohms, flashy LED špower meters, by looking at it you'd think it was cheap junk flashing its way. But turn it on, throw whatever you like at it, and it surprises you.
It's low price is reflected in its PSU, it will keep it all together up to about 25W/8 Ohms, but then it will start to lose it. Just for the hell of it, I replaced its original caps with a pair of larger caps, and sure enough, it benefited from it. Unfortunately, there's not enough space inside its enclosure to fit in a larger toroid, I suspect that would give it a still better kick.
Evidence of digital error is jitter and pre-echo I suppose.
Almost all digital components like DAC chips tend to have a slight sound difference, however their specs are naturally vastly above analog, so I'd say you're right there.
"the ratio of max signal to zero signal - but in practice the real audible dynamic range of most digital systems is far less than the quoted figures if you are listening and trying to access vital low level data"
The valid point here, at least pertaining to transducers, is that DNR is not only the total maximum to minimum, it's +-1 dB at x/y/z Hz as well, which is never ever measured.
Are there speaker amplifiers using AD797 and LME49990 though?
Sorry if it's a naive question.
Yes. I asked for a test disk, that should be high quality pressing.
Jcx quoted number (1000) is close to the disk run number for which you got the basic costs. A second stamper -if required-will raise the cost per test disk.
Frank
IMO, most digital systems fail (compared to vinyl and home tape) in one aspect: They don’t mask low level information.
Digital systems excel in measured dynamic -under any definition- range.
It is the way our mind works that gives us the impression that analog has more low level detail.
That is not bad, it is very beneficial to the end user.
This does not make the content of A.W. quote correct. I would say it is misleading but not necessarily intentionally.
George
Last edited:
George, I would agree with you if I just looked at the figures, and had no other basis for assessing - however, IME this is where measurements lead one very much astray; because I've gone through the exercise innumerable times: a system starts in having a classic "digital" sound, for want of a better word, and I start to improve it - the key approaches are extended conditioning, power supply improvements and interference abatement measures - and the more I do this the more that digital system starts to produce "analogue" sound - low level details, rich in information about the sound structures within the recording, with no fatigue in the listening. I can't accept that this is a 'trick', or fake detailing, because it does not compute, makes no logical sense - if it is, then live sound is a con, a sham - and is not worth listening to ...
That was computed like Barrie Gilbert's article did. He showed several degree's of phase on a 741 at 10kHz just going from 2V to 5V peak. I found <.0005 degree going from 1V to 10v peak output at 20kHz. I guess nothing will stop the nonsense that keeps getting posted here.
An interesting observation: some of the lounge threads here give good, factual information, while many of the so-called tech threads have all the BS.
Maybe we should swap the area names ?
Jan
So some folks can measure it but claim it's not audible, while others can hear it but claim it's not measurable?
Godfrey, each side of the coin has its dedicated believers, the hard core, if you like.
One side claims that if we cannot measure it, it doesn't exist, it's just a figment of imagination.
The other side says damn your measurements, you can't measure sound quality, at best you can hear it.
In my view, both sides are wrong, as extremes usually are. The truth has a way of finding the middle ground.
If we can hear it, but cannot measure it, this simply means that we don't know enough about a phenomena yet, and in time, we probably will.
If we can measure it, we also need to be able to interpret the measured data properly, which is something far too many "golden eared" people cannot do, or worse, do not want to do.
Even I, as a layman, have learnt to interpret some data by correlating what ma 'scope is showing me with past experience with similar views. On basis of how THD peaks behave, specifically what their decay order and timing looks like on the 'scope, I can have a relatively fair idea of what I can expect from that amp. A general idea, mind you, certainly not precise, but very informative.
Frankly, I care about that view way more than what the absolute values are like, because experience has taught me that amps measuring wonderfully can still sound just so-so even if they cost an arm and a leg, while other amps, which do not measure nearly as well nevertheless manage to sometimes sound wonderfully.
All of which tells me that we have yet to learn a lot more before we're home and dry.
If a speaker is 2 feet tall or 5 feet tall, I'm pretty sure everyone can hear that.
Now, show me the measurement.
You want to play hopscotch in a mine field?
We cannot yet define good sound by measurement for the electronics only, and you want to bring in the speakers, which are electro-mechanical devices, as transducers are?
Yes, for example I've always suspected that the electrolytics in the amplifier's power supply, which are in series with the output devices, have a substantial effect on the sound. Maybe a dynamic shunt regulator for the output stage (that didn't waste too much power at low levels) would improve the sound quality enough to be worthwhile.
- Status
- Not open for further replies.