Rowland Research Response to Class D Amp Article...
Audiophile Realist -- Tue, 11/28/2006 - 22:18
Will TAS be responding to Jeff Rowland's response on the sound of his 201? He claims that his test measurements produce nothing that would support assertions about high frequencies being "cut off". Will TAS run any measurements to support their claim? Just wondering.
AR-
I have a dealer friend who sells the Rowland products and has noted how sensitive the switching amps are to their environment and associated equipment. One of the facts of life in the audiophile world that is routinely glossed over - although I do think TAS does the best at bucking the trend - is how much set-up factors in to system performance. Grounding issues take on even greater significance, and that's long been a potential (sorry) problem area way before switching supplies. Power cords and their filtering design are crucial as well... speaker cables... the list goes on. Anecdotal, but here at home in an fairly complex system with a pre-amp using a switching supply, one grounding plug makes the difference between hashy, fatiguing sound and lovely, rich timbres.
All to say, it would have been fascinating to have Jeff Rowland's take on how the 201 sounded in Jonathan's kit.
best,
Scott
I think the comments are so opposed to one another, that either of the two simply has to be wrong. It's interesting that Rowland did not mentioned the specific roll-off figures that he has - that would have been helpful. At the same time it sheds light on the problem of TAS' reviews being entirely without any scientific measurement to back up the anecdotal and subjective review format. You have so many different variables that it is impossible to fairly judge a product without nailing something to the floor.
Audiophile Realist wrote: At the same time it sheds light on the problem of TAS' reviews being entirely without any scientific measurement to back up the anecdotal and subjective review format. quote]
AR-
I'll have to agree to disagree there as it would have to be demonstrated there is a measurement ( I do believe if we can hear something, it can be measured, but point out there seem to be things that science hasn't caught up with yet) to nail something like that down.
One thing a wise audiophile friend taught me early on was that until you sit right down next to someone and listen at the same time, it's never completely possible to understand their perspective.
kind regards,
Scott
At the very least TAS should include challenge Rowland's comment with a frequency response test of their own. And if JV and others still conclude that the highs are cut off, even if the tests report to the contrary, then this is even more reason to study what makes Class D amps tick. An amp that sounds rolled off even though its frequency response chart demonstrates the opposite is something to be looked at. The answer is not to abandon all scientific testing because there is no "absolute" test to prove what we are hearing.
As it stands, a reviewer who swears by "the absolute sound" might actually have a fondness for a certain frequency range, or perhaps their room accentuates a certain frequency range. Their praise or complaint for a product's sound may have more to do with them, their room or the associated equipment than the actual product under review. We would be in a better place to judge Rowland's complaint with actual hard data. It might not allow us to take a definitive position, but it would certainly shed more light than not doing it does.
If the rest of the world took the anti-testing stance advocated in TAS, the human race would have probably been wiped out by the common cold by now.
See the single item thread by DOT*SYSYEM in this Forum, regarding ICEpower interfaces. The Rowland site lists the input impedance of the amp as 40 K-ohms. This means there is an analog gain stage ahead of the ICE module. If RF noise from a switching power supply is not the root cause of the treble's hypersensitivity to cords and setup (which is my expectation), then the remaining candidates appear to be either this gain stage, the output filter, or something else like an unexpected internal mechanical resonance.
I reread the comments in the original article, the reply by Jeff Rowland, and the Neil Gader review of the Chapter Precis in TAS 167. It seems the complaint is not one of rolled-off treble, but a qualitative shift in treble sound suggesting complete loss of information below some level. This is consistent with my experience in finding and treating RF noise sources within my house. Reducing external noise reveals more information at the lowest audio levels.
A measurement to reveal this property of playback systems would have to plot the resolving power as a function of frequency. Bench measurements of single pieces of equipment would not be conclusive, as the resolving power is an emergent property of the system including its domestic noise environment.
Testing if low level treble data is being smothered by noise sounds like a pretty simple test to me. Establish a benchmark using an amp that does resolve the data, measure, then flip in the Rowland or any other digital amp that supposedly cuts off the treble.
Fellas,
I actually liked Jeff Rowland's amp with the exception of the treble. The 201 was, as I've noted in print, extraordinarily detailed and bloomy and realtively dynamic in the midband and quite acceptably good in the bass. But the highs...well, to my ear they just weren't there or there with sufficient air and abundance.
About this subject of treble response...all I can say is that four out of the five reviewers who took part in our survey reached the exact same conclusion about the treble of Class D amps, including the Rowland 201, and our conclusions were reached entirely independently. One of us might have had a system-related problem that affected treble balance, but all four of us? I think the chances of that are slim.
I know what I heard, and I do not use a system that rolls off the treble. Indeed, the MAGICO Minis are anything but rolled-off on top. (It is my understanding that Jeff Rowland himself uses MAGICO Minis as one of his primary references; I know he has showed with them.) I did use LPs as my source. I'm not sure how that would play into this but, for the record (so to speak), there it is.
As for Jeff Rowland's measurments of the bandwidth of his own amp...I don't want to seem insenstive or insulting to a desgner whose products I have always admired and continue to admire, but, honestly, I think such measurements are beside the point. Whether Jeff says he measures a roll-off or not , the 201's treble was missing-in-action in my system (and in Neil's and in those of several other audiophiles I've spoken with).
As it is, and in spite of its treble quirks, I gave the Rowland 201 a tentative, system-dependent recommendation. To make that recommendation an unqualified rave, all Jeff has to do is make the 201's treble sound like its magical midband.
Jon
There is a huge question here about whether the highs are rolled off at the top end, or if the noise of these amps is blurring the finer detail of treble at low volume. Why leave the question out there when it would be so easy for the audiophile community to look further into why the treble sounds cut off?
:roll:
It's like testing is some kind of heresy or something. Is it that big of a deal to give your readers anything more than the anecdotal "my friend's cousin heard the same thing"? Just measure the thing against an amp that doesn't have this problem and see what you can come up with. If the results don't explain a thing, then you conclude that the tests don't explain the phenomenon. Sheesh!
JV probably isn't intending to engage in a farce or to look foolish. However, I would not take his comments very seriously.
In the 2006 TAS Editors' Choice Awards, JV includes the ARC 150M amplifier. His "review" (p. 80) states "This... cool-running, Class T amp produces a surprisingly ARC-like sound. A little bright, forward, and weighted toward the treble, the 150M has MUCH of the NATURAL AIRINESS and LIGHT of ARC TUBES, with TERRIFIC SOUNDSTAGING, DYNAMICS, AND CLARITY for multichannel." [EMPHASIS ADDED] (Note the full review is not available at the AV Guide website.)
Of course, these comments are printed just one month prior to the following comments "For me, Class D on the whole -- with the exception of the Kharma MP150... -- has been a disappointment. There's something wrong with the treble of Class D amplifiers.... [T]here's a kind of compression of dynamics, harmonics, and spaciality in the top octaves of the Class D amplifiers I've heard. ...And air, which is quite audible on a linear amplifier, ... just ... vanishes."
It seems that JV needs to justify his current opinion (of the month?) by stating that others agree with him. However, his credibility might be best served by first reconciling contridictions in his written opinions.
Designers such Bruno Putzeys might consider that chasing JV's tastes will be a never ending battle.
JA
"Bright" and "forward" are hallmarks of RF noise. These qualities may be present with treble problems as discussed.
The measurement issue seems simple until it comes time to actually do the measurements, make sense of the data, and correlate them with listening observations. There is nothing wrong with TAS' approach as long as it is not claimed to be the only valid one.
My concern here is what appears to be a generalization upon insufficient data about switching amp treble. Insufficient data in this case means the sample population. Many of the sampled amps share core components and design methods, so they are not independent. The sample size is smaller than it appears.
This difference in the treble with class D - could it be the presence or absence of noise? Could it be either in a given situation? Or perhaps the absence of some class A or AB circuit signature previously assumed to be correct? Air in the recording or air added to everything by a component?
FWIW, the 201 review reminds me of descriptions of the sound characteristics of some transformer-based moving coil step-up devices from ancient times...
Bob
DOT makes the same point made in Jacob Heilbrunn's freshingly-unpretentious review of the Rotel 1091. I really must tip my hat to this reviewer, because he asked questions about the technology and gave a down to earth commentary on what he heard. He actually even proposes that the existing paradigm might be wrong, and that this might be the reason why reviewers are unsettled by the sound of D amps. I'd love for more writers to be as succinct and humble as he was in this review. It lacked the puffery and BS that would have made the review two pages longer. He didn't pretend like he had the answer when he didn't have the facts to back up his opinions. This I can live with.
Jba,
The ARC recommendation that you refer to was excerpted from a much earlier review in TPV of an ARC five-channel Class D amp (the 150.5), being used in a home-theater system with low-res (Dolby Digital 5.1) sources (note the words "for multichannel" at the end of the blurb). If this wasn't clear enough, I will edit the blurb in future issues to make it so. The Class D ARC amp reviewed in our roundup (the 300.2) was a two-channel amp used in a two-channel system with high-res sources. It didn't fare as well.
As for measurements...given the number of amps that measure well and don't sound good or measure indifferently and sound great, I'm not convinced that measuring per se would prove anything. I am also not sure that claiming that Class D is somehow lower in noise in the treble and that's what we were hearing--the absence of Class A/Class AB artifacts--holds water. When you can't hear certain high-pitched musical notes or overtones or ambient cues on one amp and you can hear them on another, you tell me if that's an absence of noise or an absence of information.
JV
The point is to measure a phenomenon - this one being the absence of subtle treble information. No one has advocated a position that there is a test that can determine whether or not a product sounds good in entirety. What we do know is that testing something gives us some objective information while testing nothing gives us nothing but subjective anecdotes.
TAR,
OK, while I don't see how testing will prove anything either way, I don't see any harm in it, either. I will propose that we conduct bandwidth/distortion measurements of Class D versus Class AB amps to Robert. (In a former life, Robert used to do all the measurements for Stereophile.) Let's see what he thinks of the idea.
Jon
BTW, you might want to take a look at my Parasound review in the new TAS, where I draw a comparison between a modestly priced Class AB amp and its Class D competition, including the Rowland 201.
jvalin wrote:TAR,
I will propose that we conduct measurements of Class D versus Class AB amps to Robert.
Jon
Brilliant. You guys can't be pro measurements out of one side of your mouth (when reviewing the Magico Mini), then anti-measurement out of the other (Class D).
Speakers, IMO, are somewhat of an exception when it comes to how their perceived sound correlates with frequency and distortion measurements. (The MAGICO Mini happens to be a case in point, as I noted in my review.)
That said, I know plenty of superb speakers that aren't as "flat"-measuring as MAGICO Minis but still sound remarkably lifelike--the Avantgarde Trios, the Wilson MAXX IIs, the Magneplanar 20.1s, and the MBL 101 Es, for instance.
As "case-dependent" as the correlation between measured response and perceived response is with loudspeakers, it is even more so with amplifiers, CD players, preamplifiers, and other electronics.
Not that any of this really matters to you, TAR. Obviously, anything I say is going to be attacked by you, whether I agree with you or disagree. it's almost amusing.
Dot*System asks a good question:
"This difference in the treble with class D - could it be the presence or absence of noise?"
IME, RF noise in an audio system manifests itself both as added spurious overtones (enhanced sibilants, mushy cymbals, crispy consonants) and voiding of low-level information (subtle vibrato in singers like Ella Fitzgerald and Shirley Horn, midrange warmth, piano body tone). Any class of amplifier may suffer from external RF noise pollution, but my interpretation of this controversy is that most of the amps (or the systems into which they are inserted) sampled by TAS suffer more so due to use of switching power supplies.
Solid-state Class A-B amps, on the other hand, may suffer from self-modulation of the bandwidth. Transistors contain internal capacitances that vary with the voltage on them. These capacitances help determine the stage bandwidth. Tube amps do not have this property, but there is no tube that does not have some degree of microphonic behavior.
If the sonic signature observed by JV and the other TAS reviewers is truly a characteristic of switching amps, the first place to look for a physical cause is the low-pass filter. If this filter is audible, then the self-modulation of output transistors in Class A-B amps should certainly be audible as well. There will be familiar and unfamiliar sonic signatures depending on class.
If the switching amp low-pass filter is not audible, then the cause is something that can easily be designed out of switching amps once it is understood. They will then eventually replace conventional amps because they will work better and cost less.
Quote:
Speakers, IMO, are somewhat of an exception when it comes to how their perceived sound correlates with frequency and distortion measurements.
That's hardly surprising, considering the magnitude of the role that speakers play in the total sound output. Contrary to the hype written about amps, front ends, CD players, etc., from my experience the differences are far greater amongst different pairs of speakers than anywhere else. I'd upgrade my speakers long before I spend a significant amount of money on a CD player, DAC, preamp, or amp (probably in that order). Nevertheless, frequency and distortion measurements anywhere in the chain should be pretty revelatory as long as the products further down the chain don't significantly pollute what is being measured. Assuming that the Magico Mini is flat and of very high resolution, then I'd imagine that it wouldn't be hard to do a frequency sweep vs amplitude test for an amplifier (or something like that).
It would also be nice to know if this problem being placed on Class D amps is really just a problem for amps in this price range. Do other classes of amps below "reference level" also not reveal the finest treble details?
In my system, the Roland 501s sounded harsh and strident. There was a bump in the lower treble. This could well have been EMI pollution, from the Roland or from the power grid. It could also have been a mismatch between cables, preamp, or speakers. There are so many variables.
The lower treble problems inhibited my ability to discern the quality of the upper treble.
So a totally different experience from JV. My opinion, is that the switching power supply is the culprit. All Ice power I have heard has had this quality too, with the exception of the Red Dragons (which I have heard only at shows).
I have heard the Ice Power amps on the Magicos in the Berkeley home of the designer too, and had the same opinion.
I used an Audio Prism Noise Sniffer once to examine noise in my system. At that time, a Pioneer player with a switching power supply was dumping noise backwards into the rest of the system. QuietLines and other filters could not remove the noise.
Bob
Jonathon,
your comments re treble are exactly right, I have to say I love my 201's and am massively impressed with Rowlands approach, there is no doubt though you can hear subjectively the roll of or db cut you speak of
Heres the problem,they dont sound good,sure walloping bass, which if u ask me is quite unrealistic anyway.I dont know what the proble is but there is a problem.
I owned a 501 for a while and my initial impressions weren't that favorable, however i was stunned to hear the amp improve over time, and I mean a lot of time. Break in took..get this...a month!! Apparently Rowland agrees but I don't know the reason why tbreak-in takes so long. Something about magnetic flux? No doubt the Rowland 501 is indeed a very good amp (and these were used on Pipedreams with all other ref quality stuff i,e ARC Ref 3, all Nordost cables, etc) . So here's my question. Jeff Rowland doesn't make anything less than good. SOA? I don't know, but has anyone heard his model 500? That ought to be a pretty damn good amp. And what is the new power supply device he now says makes his amps sound better? There isn't much on these (either the amps or the power supply adapter) in the world of audio reviewers. Wuz up? We need a good review!!
Correction. I meant his flagship, the model 300.
Yes, I did hear the Jeff Rowland Continuum 500, and I think it is totally another thing than the 201 or 501. By using a new power supply mit PFC (Power Factor Correction) Jeff tages the ICE-Power-Modules to a fully new level. I am quite sure, that the quality of the power supply makes the very difference, either a Class D amp sounds good or not..
I did not ever hear the new Model 312, but I cannot imagine, that there would be any lack of high-frequency response.
job
There appear to be several reasons products don't get reviewed by the most influential audio press. If it is extremely small production and artisinal in its manufacture or design, one might think there may not be enough consumer interest to know about it and thus no review is performed. After all, audiophiles are generally not as willing to by products from a "garage" operation as opposed to the efforts of a more reputable firm (unless there's something extraordinary there). I sense another reason not to review product is best summarized by the adage my mother taught me "if you don't have something nice to say about somebody, don't say anything at all". I wonder if this applies to Jeff's stuff? Jeff is, forgive the term, audio royalty. He is a long-time earnest and sincere contributor to high end audio with a good track record of excellent and occasionally SOA products. Could it be that his products are getting neglected because nobody wants to say they are "this or that" if they cannot extend some lavish praise upon them or conclude they are genuinely superb? I don't know. But here is an American manufacturer that everyone likes, respects, and wants to succeed. And he never makes bad stuff! (And whatever the hell he makes, it is certainly always beautifully made). Furthermore, he is hardly a garage operation after several decades in the business. So I ask again, why isn't anyone reviewing his SOA effort, the model 312? One would think that if anyone could do Class D right, it would be Jeff now that he's had some experience doing this for a while. I think many audiophiles would feel that some enlightenment about his most current efforts would be most welcome.
Post new comment