Print

Print


Thanks Tom - awesome post... 

>> Hi Tom.  My guess is that those fast "spikes" probably wouldn't make it intact through the output transformers...

> 

> Nor do sharp drum hits. Between output transformers and power supply sag, plus speaker damping, many tube systems sound "soft" and "cushy" to my ears. Not all of them (most McIntosh, for instance). Some people prefer the "soft" and "warm" sound ("warm" to me translates to audible harmonic distortion in harmonics that some find euphonic). I much prefer accurate, "fast" and "crisp," which usually means solid-state throughout. I'm not saying that's impossible with tubes, but I am saying that it's not typical, especially in vintage amps.



Here's another heretical possibility ---  digital was so much more fast than the capacity of 70s transistors and speakers to handle that we hated it. Until systems caught up with it. 

I've been thinking particularly about cloth dome tweeters - which ring for a long time when hit with a spike, and yes Atkinson should be doing pulse testing in his testing of analog gear and and speakers -- but I think the advertisers would scream if anybody knew how much  impulse response garbage is caused by these other parts of the chain - easier to focus on digital. The impulse response testing that was an exotic Syn Aud Con thing only Dick Heyser understood when digital came in is now fairly routine in system design, particularly driver design. 

There's so much mythology and Kudos to Tom for frankly speaking --- 

Look at the square wave of a Shure V15-II from an old Julian Hirsch review --- it looks horrendous compared to Nyquist ringing.... but in our kooky world the digital ring well outside the audio band is a Very Big Deal while the ring of a moving coil cartridge or a condenser mic or a  dome tweeter -which when hit with a transient ring from now until next week, and at a freq well within the audio band... no problem, because it's "analog".  I' don't mean to be polemical, but there is a really fundamental misunderstanding of how all this works and I'm of the belief that some systems sound "better" because they are capable of handling fast attack times, and other systems sound good because they have stages within them that round out those sharp attacks and have lazy easy time smugly sitting back sounding euphonious while actually acting as de-essers. 

And I don't know if this is original or if I just made it up, but the waterfall plot doesn't (yet) tell the story of the FM doppler distortion and IM caused during the ring period...
The focus has been on energy distribution as it relates to audible versus actual frequency response --- think of a C12 --- it sounds lovely -- brighter than it       measures --- because we're hearing 10 or 15 cycles of 14kHz capsule resonance overhang and that's not showing up as dramatically in the chart as we actually perceive it. Ribbons don't do that --- which is why they are employed as the go-to mic for horn sections.

But I think it would be smart to look at the non-linear non-musical sum and difference products caused by the incoming audio mixing with the audio that preceded it... in the decay interval the ringing interferes with the incoming new material and the product of the overhang and the incoming creates nasty IM. This is kind of a new angle I haven't seen expressed anywhere - pitched it to Sean Olive at last AES and saw the wheels turn...

Digital would provoke more exaggerated TIM in an inadequate component than a slow smurfy analog mic-into-transformer-shaped-tape into transformers- into cutter - into vinyl - into tubes - into transformers - into a slow reacting ringing speaker... and it's not the digital's fault. 


Dave Smith ( moment of silence --- we lost him yesterday .........................................................................................................................................................);
and I really dug into this in the early 80s. Neither of us liked what we heard, but we believed in the potential of digital so we played around a lot with differect circuits... I was good at listening --- and assembled two great sounding systems that worked well with early digital. 

Both systems were based on the same D/A converter we custom-designed and hand-wired ... a couple LM318s - the chip in the ATR102 - we knocked off the circuit and attached it to a Phillips 176.4 DAC as the I to V converter and a buffer stage --- fastest chip of its day, and I laugh like hell when all these exotic pieces of  discrete vs tube vs kryptonite hardware float around playing back tape played on the godly ATR which was entirely IC-based... 

System #1 was a little 15 watt class A Musical Fidelity transistor amp that was so sweet sounding because it barely had the power to even say hi to a transient let alone pass it through- just softened everything - to a Fostex concentric horn driver with neodymium magnets and a tiny PC plastic driver --- very fast, little ring. This system sounded fabulous at very low levels on digital stuff. 

System #2 , which really rocked and surprised the hell out of us was based on Dave's insistence on speed, power, and stability --- and my obsession with low ringing --- Spectral amps flat from DC to light and stable --- that amp could pass a 1MHz without batting an eyelash --- driving a Quad Esl 63--- which have fantastic impulse response. 

And on either system the digititis was much reduced. As always, that which he speculated was the case, turned out to be the case. 

Over the past 20 years the industry has improved the resonance performance via beryllium drivers and better magnets. And slew rate in componentry  is much higher than the 70s. 

But yeah how about them transformers and the slow rise time of vinyl - from the cutter to the listener, ain't that awesome... and it's not established fact that that's a lot frickin' easier signal to pass than digital, and when you hit these inadequate systems that round everything off with a sharp rise time they go nuts. 

So the Beatles come out in mono vinyl and everybody goes crazy. 

We're still hoping for the opportunity to play back these tapes (and other vintage material) through an ultra-fast high-bandwidth system with no time-base-jitter, insane headroom and low TIM. 

- Jamie Howarth




On Jun 18, 2014, at 9:35 AM, Tom Fine wrote:
> Nor do sharp drum hits. Between output transformers and power supply sag, plus speaker damping, many tube systems sound "soft" and "cushy" to my ears. Not all of them (most McIntosh, for instance). Some people prefer the "soft" and "warm" sound ("warm" to me translates to audible harmonic distortion in harmonics that some find euphonic). I much prefer accurate, "fast" and "crisp," which usually means solid-state throughout. I'm not saying that's impossible with tubes, but I am saying that it's not typical, especially in vintage amps.
> 
> -- Tom Fine
> 
> ----- Original Message ----- From: "Mark Durenberger Mobile" <[log in to unmask]>
> To: <[log in to unmask]>
> Sent: Wednesday, June 18, 2014 9:20 AM
> Subject: Re: [ARSCLIST] headroom
> 
> 
>> Hi Tom.  My guess is that those fast "spikes" probably wouldn't make it intact through the output transformers...
>> 
>> 
>> 
>> Mark Durenberger
>> On the Road
>> 
>> -----Original Message----- From: Tom Fine
>> Sent: Wednesday, June 18, 2014 7:35 AM
>> To: [log in to unmask]
>> Subject: Re: [ARSCLIST] headroom
>> 
>> Hi Mark:
>> 
>> I don't listen loud enough for this ever to have happened to me, but I have no doubt that it's
>> something that does happen in more raucous studio monitoring environments, and perhaps some
>> home-listening situations. Do you think a typical Williamson or Ultralinear tube power amp would be
>> less likely to blow the tweeter? I'm wondering if their rise and attack times are too slow to bring
>> the full impact of the pop or tick to fruition in the speaker-motor?
>> 
>> One thing I wish John Atkinson would test and publish for all equipment he puts on the Stereophile
>> bench is pulse response. He does this for digital stuff, to look at jitter and distortion. But he
>> should be testing pulse power to clipping with power amps and pulse power to cone breakup with
>> speakers. Music, especially modern music, is full of percussives, if it's not processed and
>> toothpasted to the point of having no dynamics. A thing I've noticed with the new high-resolution
>> remasters of the first 3 Led Zep albums is how much headroom there is between peak drum hits and
>> other transients vs. average level. And they used plenty of dynamics compression, more than typical
>> for that era. Still, there is nearly 12dB difference between peak level and average level in some
>> songs, which is huge for rock music. Turn up the volume some and those Bonzo drum hits really move
>> the air.
>> 
>> -- Tom Fine
>> 
>> ----- Original Message ----- From: "Mark Durenberger Mobile" <[log in to unmask]>
>> To: <[log in to unmask]>
>> Sent: Wednesday, June 18, 2014 7:49 AM
>> Subject: Re: [ARSCLIST] headroom
>> 
>> 
>>> As usual Tom...you're spot on.  BUT don't forget that loudspeaker high-frequency drivers can be taken out by such steep-rise-time transients faithfully passed through a good amplifying system :-))
>>> 
>>> 
>>> Mark Durenberger, CPBE
>>> On the Road
>>> 
>>> -----Original Message----- From: Tom Fine
>>> Sent: Wednesday, June 18, 2014 6:20 AM
>>> To: [log in to unmask]
>>> Subject: [ARSCLIST] headroom
>>> 
>>> Hi Paul:
>>> 
>>> This opens up a whole new area of discussion!
>>> 
>>> I have a phono playback system with a lot of headroom and am very surprised at just how loud tick
>>> and pop transients can be (6-12dB above musical peaks in some bad cases). I definitely agree that
>>> clipping distortion plus tick equals worse sound. I think many of the tube preamps don't clip
>>> because tubes aren't as fast into clipping and then the tick goes by and all is swell again,
>>> especially if there's enough power supply current so nothing sags. My father did tests on mic
>>> preamps, and talked about it at one of his AES NYC Section sessions. He found out that almost all
>>> solid-state designs could be driven into clipping with European condenser mics relatively close to
>>> muted trumpets, drum strikes or hard-hit piano notes. In the midrange frequencies, this was very
>>> audible and distracting. The tube preamps, with the same sensitivity specs, and usually
>>> older-vintage input transformers, did not overload as easily and did not present annoying clip-pops.
>>> Keep in mind, this was circa early 70s testing, but also keep in mind that those early 70s
>>> solid-state consoles are coveted by some to this day. I think that later transformerless designs,
>>> with good input sensitivity control (real low-noise attenuator networks, not just a feedback
>>> adjustment on an input differential amp), mitigated a lot of these problems.
>>> 
>>> I'm not saying you need tubes for good fidelity, in fact I use almost no tube gear in my studio and
>>> none in my living room listening system. I am saying that, if you go with a solid-state design, for
>>> anything involving wild and woolly peaks -- like a phono preamp or a mic preamp -- you better design
>>> in a lot of headroom, more than you ever expect to need. You also better design a rock-solid
>>> over-spec'd power supply, in either tube or solid-state equipment. In tape electronics design, there
>>> is a known limit to headroom (the saturation point of whatever tape you're using). This, for
>>> instance, is why Beyer Peanut transformers work fine for Ampex AG-440 playback electronics (when
>>> operated within spec -- +9 operating level is not appropriate for any older machine) but not so good
>>> for mic preamps (except for low-output ribbon mics in front of moderately loud sources). And with
>>> digital you have a brick wall limit to dynamics -- the distance from wherever you start to digital
>>> zero. But even there, new debates rage about headroom. See the design notes for the latest Benchmark
>>> DAC2 units. They were purposely designed with 2dB above digital zero headroom, explained better than
>>> I can attempt in their white papers and technical notes.
>>> 
>>> -- Tom Fine
>>> 
>>> ----- Original Message ----- From: "Paul Stamler" <[log in to unmask]>
>>> To: <[log in to unmask]>
>>> Sent: Wednesday, June 18, 2014 12:21 AM
>>> Subject: Re: [ARSCLIST] Upgrading Audio Systems
>>> 
>>> 
>>>> On 6/17/2014 12:53 PM, Miller, Larry S wrote:
>>>>> Regarding your reluctance to upgrade your system to the point where
>>>>> you can't stand to listen to your favorite recordings, I completely
>>>>> sympathize.
>>>> 
>>>> Funny, I'm just getting ready to address that question in an article I'm working on (about a phono preamp design that's been percolating for several years). My suspicion is that the electronic gear that "makes pristine records sound better but makes imperfect records sound worse" is not, as is often asserted, "revealing more accurately just how bad these records really are". I suspect, instead, that imperfect records sometimes, through their imperfections (scratches, wear, etc.) stimulate misbehavior in the electronics which then causes us to hear the records as sounding worse.
>>>> 
>>>> Back in the 1970s, when tube equipment was beginning to reappear in the home-audio world, a lot of people reported that certain tube preamps seemed to emphasize scratches less than their solid-state brethren -- I heard this a few times myself. A well-done article in, I think, JAES, pointed out that the tubed preamps in question all had significantly more headroom than solid-state preamps of the time, and suggested that this might explain the lesser audibility of scratches. That sounded reasonable to me, and my own experiments (written up in audioXpress) suggest that the worst scratches on LPs and 78s may be as much as 26dB higher than the 5cm/sec considered "nominal level" in disc cutting. I designed the preamp with that kind of headroom in mind, and its various incarnations so far have sounded very good.
>>>> 
>>>> I hope to submit the article for publication within the next six months. The design actually comes in two flavors -- one intended exclusively for RIAA discs, and one with adjustable compensation (mostly for 78s).
>>>> 
>>>> Peace,
>>>> Paul
>>>> 
>>>>  One reason I'm so fond of the Stanton 881-S cartridge is
>>>>> that it seems to make recordings sound good without getting to the
>>>>> point, as do some moving coils, that mediocre recordings sound
>>>>> unpleasant.  I often wonder why this cartridge isn't used in
>>>>> transcribing 78s since it was available with an off the shelf 2.7 mil
>>>>> stylus and, to my ear, sounds much better than the commonly used
>>>>> Stanton 500. But getting back to upgrading, I'll offer an alternate
>>>>> view.  I'm not one who frequently upgrades my system, but when I do,
>>>>> that is, when I hear some piece of equipment that makes some of my
>>>>> favorite recordings sound better, then acquire it, I'm essentially
>>>>> rewarded with a new record collection, sometimes hearing things I
>>>>> never heard before.  That's why I upgrade.
>>>> 
>>>> 
>>> 
>>> 
>>