Saturday, 7 August 2021

SUMMER MUSINGS: No... Not all amplifiers sound the same (but many do! ;-).

With both my Class D Hypex nCore NC252MP amp and linnrd's Class A Pass/First Watt SIT-2 in my sound room a couple months back, this allowed me to switch reasonably quickly between them to listen. You might be wondering about some comparison comments on how the amps sound. Certainly much has been said in the press about sonic differences between amps over the decades, but relatively little has been shown with measurements in an actual room using what comes out from the speaker.

First thing I'll say is, "No, not all amplifiers sound the same!". This should be obvious actually when looking at the objective measurements of the amplifiers and what they mean if we extrapolate the results to a known system.

Like I said, it seems uncommon to see measurements "in situ" with amps connected to the same DAC, preamp, and speakers to show differences. To demonstrate the differences in my room with my speakers, I set up the miniDSP UMIK-1 microphone in the room like so to run some sweeps and tonal FFT's for comparing:


So what I've done here is set the microphone a bit further than 1m away from the right speaker along the front tweeter axis as the consistent point of reference. I have the laptop shown on the left connected to the RME ADI-2 Pro FS Black Edition DAC for signal generation. In the picture above, the small Hypex nCore NC252MP amplifier is currently turned on, playing the test signal, sitting on top of the Pass amp.

Here are the frequency sweeps using the Hypex nCore vs. the Pass/First Watt SIT-2, targeting around 85dB SPL (1kHz cursor at 86dB SPL using both amps), playing to the right Paradigm speaker, room correction DSP obviously turned off:


As you can see, even though we're using the same DAC, preamp, cables, speaker, measurement microphone, and room, the in-room frequency responses with the 2 amplifiers are clearly not the same. We can appreciate that the frequency response from the Hypex nCore Class D amplifier has maintained stronger upper and lower frequencies while the Pass SIT-2 allows for a relative reduction in both the highs and lows, thus relatively accentuating the midrange in comparison. This is despite both the SIT-2 and Hypex NC252MP having relatively flat measured frequency responses with typical 4/8Ω resistive loads; a nice reminder that resistor loads are not typical speakers.

To make it easier to see the difference, here's a plot of the delta between the Hypex over the Pass SPL across the audible frequencies:


As you can see, there is up to 5dB difference between the amps such as at 100Hz where the cursor is. Clearly this kind of tonal difference would be audible!

Notice that the 3-way Paradigm speaker has impedance dips in the graph above reaching their most extreme down to 3.5Ω both in the high and (more chaotically) lower frequencies. This would at least to a significant part explain the frequency response variations between the Hypex and the high output impedance (low damping) of the Pass SIT-2. There is more here in that I am not showing the electrical phase variation which would interact to create an even more complex "effective" impedance (this is where calculated EPDR "Effective Peak Dissipation Resistance" could make a difference).

For completeness, here is the very significant spread in damping factor offered by the two amplifiers. On average, the Hypex achieved 340x damping factor compared to the SIT-2's 1.6x into 4Ω! Use of feedback in the amplifier design will play a huge part in this result:

Notice the log scale used!

Furthermore, yes, while harmonic and intermodulation distortions might be difficult to hear, they can be shown to be different during speaker playback at normal volumes. For example, let's have a look at the Triple-Tone Distortion signal I use in my amplifier tests, played back at 80dB SPL level using the Hypex vs. Pass. With an in-room sensitivity of 92dB/W/m for the Paradigm speaker, we're only asking the amplifier to produce around 0.1W:


Notice that even using a little USB microphone like the UMIK-1, we can see the difference in distortion levels between the amplifiers. While the TD+N result isn't all that different at this output level (Hypex NC252MP -32.6dB, Pass SIT-2 -31.4dB), we can certainly see the higher harmonic and intermodulation frequencies created by the SIT-2, especially notable are the pair of +/-48Hz intermodulation sidebands around the 960Hz tone. These distortions will only get worse when we increase the volume. For example, let's increase the volume by +5dB:

BTW: Notice the 60Hz hum with the Hypex nCore. Normally this is not present when I'm feeding it with balanced input. In this case, I used an XLR-RCA adaptor for the preamp RCA input to the amp which picked up some hum.

We can see that the average volume level has now gone up by about 5dB from -53dBFS to -48dBFS for both amps (value in the upper right side box). With the Hypex NC252MP, this has resulted in a relatively clean 5dB improvement in TD+N from -32.6dB to -37.4dB. But look at what happened with the Pass SIT-2. TD+N got better, but only marginally by -1.4dB from -31.4 to -32.8dB. With higher amplitude, more power usage, the SIT-2 also added more distortion as predicted by our THD+N vs. power curve in the amp measurements. We can see this visually with those +/-48Hz sidebands around 960Hz and also very noticeably the elevated 1920Hz 2nd harmonic. Furthermore, we can also see that the high frequency 5472Hz tone "drooped" a bit further with the SIT-2 at higher output levels, likely indicative of the poor amplifier-speaker impedance match; again, in effect accentuating the midrange frequencies compared to higher frequency "presence" and "brilliance" regions.

By the way, the intermodulation sidebands (like those +/-48Hz ones around 960Hz) are obviously much higher than any jitter sidebands from modern DACs! In general, this is an example of why it's not worth fretting over jitter, especially with amps like the SIT-2 or when using tube gear. The idea that audiophiles would use stuff with high non-linear distortion like tube amps or tube DACs (such as this horrific mod job) and then claim that they can hear the difference "hi-res" audio makes or need low jitter digital equipment is simply preposterous! Be careful of anyone who claims he can supposedly "hear" jitter using non-hi-res, non-solid state gear. Be even more suspicious if a vinyl-lover criticizes the sound of digital gear claiming it's because of jitter! ;-)

As I mentioned above, I don't remember seeing measurements like this in the audiophile magazines, but I think there are some basic "truths" here which audiophiles need to know about when selecting amps. Understanding objectively how the technology affects the sound is important. With that insight, hopefully then we can also wisely choose what would likely best fit our needs. This will also reduce the amount of false claims made in public or in print which is important for the integrity of discussions.

So, to summarize the main thinking points:

1. Aim for higher damping factor (lower output impedance) amplifiers if you want a system which maintains a flatter frequency response with varying speaker impedances. But no need to go overboard! Most of today's speakers are rated nominally around 4-8Ω which is why I usually measure amps using the "more difficult" 4Ω load and I like to see something like around 20x or more damping factor across the audible frequency spectrum, not just single average value. High damping will allow the amplifier to act with less "load variance"; achieving a flatter frequency response. I believe this is a huge part of the difference between the sound of the Hypex nCore vs. Pass SIT-2 I'm showing here.

2. Aim for adequate power for your needs. A number of years ago, I posted this article based on the work of others which I hope can help you find out for yourself approximately how many watts (into 8Ω) you might need from your amp using your speakers and preferred listening level. Obviously, an amp that can exceed what's "needed" is good practice since that will cover the dynamic range for loud transients. For this reason, the SIT-2 would not be appropriate for my speaker system especially when I implement ReplayGain level normalization plus usually a bit of attenuation when running room correction DSP. Also the SIT-2 is definitely not going to work when I need the dynamic "oomph" with movies!

An amplifier that is able to provide enough power for your needs (and a bit more) will allow you to run well within the device's performance envelope and reduce strain in the form of distortion at high overall amplitudes and during strong transients.

3. Related to (2) above, "clipping" can be more pronounced with solid-state amplifiers. (Even though the SIT-2 is solid-state, it acts tube-like in this regard.) For example, we can see the very steep rise in distortion of the Onkyo TX-NR1009 receiver in the THD+N vs. Voltage Output graph. So long as you stay within the low-distortion power limits (which is not difficult since this amp can provide ~200W into 8Ω), there's no need to worry.

I see tube amp advocates making a big deal about solid-state amps clipping and sounding harsh sometimes, or play up how the "graceful" rise in distortion can make tube amplifiers sound "more powerful". As far as I am concerned, if you want more power, just get a more powerful amplifier to begin with! It has become a bit of a cliché in tube amplifier reviews when the reviewer says something like "this tube amp sounds more powerful than 10W!". So what if the tube amplifier can play louder but with progressively worse distortion? IMO, there's no need to be all that excited or even glorify an amplifier like the SIT-2 rated as 10W with a relaxed harmonic distortion spec up at 5% (instead of the usual 1%).

4. During discussions among audiophiles, it's still good to respect the "more subjective" minded of us who presumably like the effect of lower damping factor, higher overall non-linear distortions, and the gradual rise in distortion with higher output (or some combination).

As discussed, I would not use the Pass SIT-2 in my main system due to the lack of power and also my philosophical leaning towards achieving "high-fidelity sound" which means "accurate" sound. My reference is a clean sonic signature with an amplifier that doesn't editorialize what's being fed in and hopefully speakers likewise not "coloring" the sound too much (as usual, it's your speakers and room that make the most difference).

Having said this, I can hear why some would like tube amps or the distortions that the Pass SIT-2 adds. It can be "sweet" particularly at certain frequencies and with various genres like female vocals in the case of the SIT-2, likely due to the relative accentuation of the midrange "warmth" and reduction of the higher frequencies which can take away some "harshness" especially in bright recordings. Also, for acoustic and live performances, the amp's higher (but not too high!) background noise level and distortions might even add a subjectively pleasing "ambiance" to the sound which I think could make it "rounder", "fuller", more "airy". Some listeners might describe this as the sound being more "3D" even though I would not say this is accurate to what's on the recording itself. As usual, a "euphonic" sound doesn't necessarily mean better fidelity nor higher resolution. Obviously, vinyl and tube gear in general are not high resolution devices. Perhaps ironically, as one upgrades to higher performing tube gear, the sound likely approaches that of good solid state.

[As an aside, it's always fun to read what was said back in the day, for example this J. Gordon Holt article from Stereophile on the solid-state Dynaco Stereo 120 from 1966 echoes from a time when transistors first made their appearance on the audiophile stage.]

The euphonic sound is probably why some audiophiles prefer the simple, Class A, flea-watt SET tube amplifiers mated with efficient speakers; not because of the absence of crossover distortion from Class A. Crossover distortion in Class AB amplifiers these days isn't really an issue. Realize that psychologically, some audiophiles might experience the sound of SET amps with sentimentality and describe the cleaner solid-state sound as being "more hi-fi", or use the term "clinical", even pejoratively "sterile". I personally prefer the "clinical" sound when the recording calls for it even though I know that this will lay bare poor-sounding recordings (for example, NoNoise post-processing can be rather unnerving on hi-res playback gear).

I would prefer to let the recordings (the product of the artists and production teams) dictate whether it's supposed to sound clean rather than have the playback gear impart its own sonic signature. "Clinical", accurate amps can still get down and dirty with my grungy Nirvana, Soundgarden, and Pearl Jam albums when the situation arises, so that's all good. ;-)

There's nothing wrong with wanting to achieve different goals in life. So long as we understand ourselves and each other! No point in "objectivists" declaring a non-hi-fi-measuring tube amp as "garbage" or "defective" - just say that this is not a "high fidelity" product. Likewise, I think it's poor practice for a "subjectivist" to declare some highly distorting tube amp as "best sound ever" without clearly defining it as "best sound ever for me, my system, in my room" as the case may be, identifying accurately that the claim is not meant to be universal (which by definition is why we use the term subjective).

Appreciate that difference, communicate that difference, especially in mainstream print and online. In doing so I think we'll enjoy each other's company a lot more and build up a hobby with higher collective IQ/EQ. Only then, I think, can we grow the hobby into something that "audiophiles" can be proud of (and in the process embodying rationality of course). 

Where consumers go, the manufacturers I suspect will follow. If as an audiophile community, we can speak insightfully and realistically about why something like an amplifier can sound different, looking for characteristics like damping factor (output impedance) and understanding the relationship between distortion and power in the context of our sound systems, then hopefully manufacturers can also present their products in a realistic light without resorting to excess hype which ends up sounding ridiculous to knowledgeable hobbyists.

5. While I can show and hear a difference between the Pass SIT-2 and my Hypex nCore quite easily, realize that within reasonable objective operating parameters, amps will sound alike! Suppose you put two very disparate (say an amp costing $10,000 and another at $500 of various dimensions, power, features) but "clean" amplifiers side by side with the following characteristics:
- Flat frequency response 20Hz to 20kHz measured with a typical resistive load (like 4Ω).
- Low distortion <1% THD for whatever maximum watts you need (100W into 8Ω likely adequate for most of us).
- Adequately high damping factor say 20x across the frequency spectrum for your speaker's nominal load assuming there are no prolonged portions of the electrical impedance curve dipping far below that.
- Low amplifier noise floor devoid of hum, hiss, noise spurs above your listening room's ambient noise level when the amp volume control is set to desired listening amount.
Likely, you'll find such "clean" amps will be indistinguishable in a volume-controlled blind test with actual music regardless of whether they're Class A or AB or D. Tube or solid-state. Lower or higher feedback.

Manufacturers love to differentiate and imply sonic differences using technical descriptions (eg. Class A, low feedback) and some audiophiles likewise might swear by them without evidence that such claims are true - or take the time to listen in a controlled environment, much less blinded!

As usual, some folks will also gravitate towards luxury products and the pride of ownership that can come with that. Yeah, tube amps can glow nicely and their physical "warmth" might accentuate the perception of sonic warmth. Likewise, big price tags and seeing an imposing 100lb behemoth amplifier in a fancy showroom may have strong psychological effects. When it comes to sound quality, let's try not to conflate "high fidelity" sound with the "non-utilitarian benefits" of luxury products.

I hope you're all enjoying a good summer, dear audiophiles. Stay safe, and enjoy the music!

Addendum: October 2021
Here's an interesting PDF of an old blind test article on amplifiers by David L. Clark and Ian G. Masters I believe from January 1987. There have of course been discussions and rebuttals like this.

12 comments:

  1. Great approach. Rare opportunity to look at two such antagonistic amps compared. What is most intriguing for me is distortion vs amplitude graphics. Starting from the volume you almost still hear in your room.
    Of course such minimalistic amp as SIT deserves more delicate speakers, let’s show some respect here :)

    ReplyDelete
    Replies
    1. Hey N,
      Good point about running the distortion vs. amplitude. Alas, didn't get a chance to when I had the SIT-2 here but maybe in the future I'll try rigging this up when I have similar type amplifier. I'm mindful about the resolution limits of the UMIK-2 microphone of course so maybe someone with better gear can give it a shot.

      Yeah, "delicate" speakers are needed with something like the SIT-2!

      Delete
  2. Great points. So many needless arguments when the whole purpose of this hobby (at least for me) is the enjoyment of music. For myself, I am finding I prefer a system that actually is somewhat tolerant of poor recordings. Yes I may miss out in the overall fidelity to accomplish this, but it adds to my enjoyment of the music I have.

    The point about similar measuring amplifiers sounding the same regardless of cost or topology is still a hard principle for my human brain to accept. The engineer in me absolutely agrees, but psychologically my brain whispers to me, "c'mom a $10,000 amp absolutely has to sound better than a $500 amp!"

    ReplyDelete
    Replies
    1. Hey Joe,
      Love that little voice in the head ;-). Important as audiophiles I think to appreciate and express the 2 things you said there:

      1. Know what you're after. As you expressed, and I cannot disagree, it's a good goal to seek a system that is able to make as many of the recordings out there sound as "good" as possible. "Good" being a personal judgment and recognizing that it's OK if what we seek might not be the most resolving or highest fidelity.

      In this case, you will have to experience and find for yourself the balance you seek between DACs, (pre)amps, speakers, room setups that achieve that goal with the music and genres you enjoy.

      For me, as one who's interested in "fidelity" in my soundroom, I think the job is actually easier to some extent! I try my best to objectively measure and confirm the sound of the gear I'm listening to, make sure I have room-correction filters that achieve the objective goal within reasonable limits. I have personally never found myself not still loving the music even if the gear is showing me the limits of the recording (heck, I can enjoy music on my car stereo and FM radio even if I didn't have all the hi-fi stuff).

      Both these approaches will at some point reach a "good enough" stage I trust so neither of us should go into an OCD overdrive. Regardless, I think getting older probably will increase the ease of satisfaction over the decades. Unless we end up being grumpy old men never satisfied about anything - but that's a personality issue. ;-)

      2. You demonstrated great insight into the psychology of self and in relation to audio.

      It's not easy to accept some facts that may seem counterintuitive. After all, "flat earthers" still have videos on YouTube showing supposedly how the Earth doesn't curve in the distance and the sun revolves in some unusual trajectories. ;-) How is it that the Earth is a globe and we're travelling thru space at 30km/s if I can't personally feel it!? That seems psychological dissonant yet do we accept the science or are we going to go by personal intuition alone and say that the astronomers are all quacks and there's some kind of conspiracy going on?

      The third option is to test oneself! No, we might not be able to go up in space and see the Earth curvature ourselves (so let's just have Richard Branson and Jeff Bezos do it for us ;-). But in the audiophile world, we can certainly measure the equipment, cables, room, etc. ourselves as hobbyists and use that to test with what we perceive.

      Delete
  3. Finally! I've always wanted someone to do the test you just performed here Archimago! I mean, if a component changes the sound of a system to a significant degree, it should be measurable at the output of the speakers.
    (And, in principle, that *should* go for many of the claims made for differences between cables, CD players, DACs, AC cables, power conditioners and other tweaks. Very often the subjective claim for these seem well in to the measurable range "highs extended, bass deeper, tighter" etc. Though I would bet on very few if any of those holding up in the test you did, like you are getting between those amplifiers, and as you might find among certain tube amps as well).

    ReplyDelete
  4. In addition...

    As I've mentioned before I've used my CJ Premier 12 tube amps for 20 years, and through that time have also used solid state amps, that I've owned or borrowed. Most recently I borrowed a friend's Bryston 4B3 SS amp for a month to switch back and forth and compare. Since I am generally a fan of carefully arrived at empirical conclusions, I wished there was some easy way to do a blind (or double blind) switching between the two. However, having seen the attempts of someone else on a forum to make a switcher that would safely switch between his tube and SS amp for blind testing, and also reduce variables, adding resistors etc, it was clearly a much harder task than I had figured. And frankly, I'm too lazy to go through that(and not an electronics wiz anyway, though I have managed various blind tests for easier cases, like DACs, cables, music servers).

    Anyway, I have to roll with my sighted tests, though I'm left with very distinct impressions I'd pass a blind test between the amps.

    So, to relate my own experience/goals to your adherence to hi fidelity:

    I'm a nut about instrumental tone and timbre. I'd always sit up close to the symphony because I preferred maximum vividness of tone and texture, vs the mixing effect of the hall from further seats (I know, most prefer the latter).

    Similarly, the more differentiation a system can give me of instruments timbre and tone, the better (and more realistic) to my ears. Same with recordings themselves: I want the character of recordings rendered with maximal difference (that exist on the recordings).

    And that is what I hear from the solid state amp vs my tube amps. Which makes sense because if the tube amp sounds different it's distorting the sound, and if it's distorting it's distorting the timbre of instruments as well. So, yes, I hear a very slightly more transparent sound with the SS amp that gives a teeny bit more differentation of instrumental timbre.

    But what the tube amp seems to give in return, to my ears, are characteristics that sound a bit more natural and believable to my ear, a bit more richness, roundness, a sort of slight emphasis of presence while also sounding a bit more relaxed in terms of transients not sounding so hard and electronic. So it actually introduces some characteristics that are a little bit more in line with what I hear in real sounds. Just a bit.

    So it's a trade off. And ultimately I find it the trade off I choose, because even with the mild distortion, instrumental timbre is still highly and richly differentiated on the tube amps (just not to the nth degree as the solid state amp), and essentially all the details of the recording seem presented, all the musical, production choices, down to the teeniest barely heard reverbs, everything that makes one track sound different than another.

    So ultimately I find that the tube amp distortion is subtle, but *just* enough to add a flavor I personally appreciate, yet not nearly enough to suggest a "lack of fidelity" to the recording in any practical aspect.
    The characteristics of records completely swamp the level of coloration from the tube amp, and I have no problem hearing essentially what the artist laid down on the track. (In fact I have musician friends who bring over their own music and also different masters of their music, to evaluate on my system, and it does a fine job showing even the most subtle differences in mastering).

    Cheers!

    ReplyDelete
    Replies
    1. Greetings Vaal,
      Nice discussion and description of your experience! Yup, no surprise that the difference between your CJ Premier 12 sounds different from the Bryston 4B3!

      Just look at these measurements ;-)
      C-J Premier 12: https://www.stereophile.com/content/conrad-johnson-premier-twelve-monoblock-amplifier-measurements

      Bryston 4B3: https://www.soundstagenetwork.com/?option=com_content&view=article&id=1699:bhk-labs-measurements-bryston-4b3-stereo-mono-amplifier&catid=97:amplifier-measurements

      Yup, very fair tradeoff between characteristics like measured distortion level and enhancing psychoacoustic illusions without sounding excessively colored!

      "I mean, if a component changes the sound of a system to a significant degree, it should be measurable at the output of the speakers."

      Absolutely Vaal, I agree. Hi-Fi audio ultimately has to do with the sound waves hitting our ears. Yes, there is a lot of complexity we should consider, but nothing's traveling near the speed of light (other than electrical conduction I suppose), nor does there appear to be anything audio-specific operating on the quantum scale to suggest deviation from basic physics.

      I find it fascinating when some extreme subjectivists automatically jump to the conclusion that there's "something" being missed in measurements or that current audio science is inadequate to explain what they claim to "hear" with stuff like cables and magical tweaks. All this without them actually taking the time to systematically verify that their beliefs are even true... Very strange considering that they're typically just talking about obviously measurable things like noise level, relative frequency changes, or dynamic variations!

      Ethan Winer (The Audio Expert, 2013) puts it quite succinctly:
      "In truth, only four parameters are needed to define everything that affects the fidelity of audio equipment: noise, frequency response, distortion, and time-based errors. Note that these are really parameter categories that each contain several subsets."

      I bet some folks would be upset by such a statement ;-). However, unless evidence can be brought forth to show that "something" else exists, I agree with Winer that we should deal with achieving excellence in those 4 parameters before going off on nebulous spiritual exercises.

      Or worse - paying money to people who claim they've achieved certain levels of audiophile enlightenment.

      Delete
  5. Hi Archi

    Great effort. Thanks.
    BTW: The Audix TM-1 would be a great update for you.
    (or the Earthworks M30, if you want to climb even higher)

    All the best, from my side

    ReplyDelete
    Replies
    1. Great, thanks for the tip Juergen!

      Will keep an eye out for the microphone ;-). Hope you're enjoying the trip!

      Delete
  6. Hi Archi,
    very interesting to see that amplifiers can provide different results not only because of their overall engineering (class, type of transistors, capacitors, etc.) but their ability to drive speakers: we easily forget that speakers are not an easy constant resistive load !

    I am personally in the same team as you: I do prefer to have an accurate system over a "warm square rounding" one (that keeps me far away from turntables and tube amplifiers !). It may be annoying when listening to bad recordings, but OMG, the reward is when you listen to high quality recordings, what a pleasure to have it all !

    ReplyDelete
    Replies
    1. Exactly DColby,
      My feeling is that if there is something in the recording to hear, laid down there by the artist/engineer, then I simply want my system to be able to get that out for me.

      Whether this level of detail will sound "good" to me is not the business of the sound system or the designer of the amp/speakers/DAC/etc.!

      All the best...

      Delete
  7. What is best is that which connects you to the music and that is likely to be different to different people.In a perfect world all speakers would have a perfectly flat impedance of say 8 ohms.If anything modern speakers have moved away from this ideal and it is almost normal for them to dip down to 3 ohms less in the bass and also kick up to relatively high impedance at higher frequencies.This is really not good enough and yet manufacturers have been allowed to get away with it.So then you need a high damping factor/high negative feedback/high current/low output impedance amplifier to drive it properly.And that is two wrongs trying to make a right.

    ReplyDelete