Hello Steph,
There's a lot of good sense in your thread.
A few comments: there is a LOT of documentation and knowledge about what you call 'cognitive amplification factor'. The scientific arena of cognitive and neuro-psychlological studies is vast and the area of how we perceive musical signals and sound in general has been beaten to death several times. It's just that the average audiophile doesn't know it and/or is not interested.
Secondly, your last statement is on target, or even better: ALL our perceptions are the result of a 'model output'. One of the very many inputs to the auditory model is the air vibration entering your ear; it is by no means the only input.
jan didden
Even if done double blind or otherwise dismissing any "latter design must be better" bias, would this truly result choosing the amp nearer a "straight wire with gain," or would you end up choosing the amp that gives the distortion that most pleases you?
Perhaps in these comparisons you should NOT pick the one you prefer, but rather the most "sterile" sounding amp!
Regardless of Self's intentions, this could be a serious product and I thought of a separate "distortion" box in front of the power amp while reading the thread before getting to Self's "niceness" comment.
The "preamp" is traditionally both a source selection box (phono/CD/Tuner/Aux) and a sound processing box (bass, treble, stereo/mono). Why not add extra processing controls to add "appropriate" distortion? I can see a rotary switch with positions marked SET, P-P, SE/SS, P-P/SS, (perhaps separate settingss for bipolar and MOSFET output) - these would determine the basic transfer function and thus spectrum of distortion. An "effect" control (perhaps not what it would be marked! but a good enough label for this discussion) would vary the amount (logarithmically, much like a volume control - 20 percent on gives maybe 1 percent effect, helps to more easily adjust the lower levels). Again, I don't know how the "user markings" of the ranges would be, but fully counterclockwise would of course be "off", turned slightly clockwise would be "good" (or "Nice"), and turned mostly or fully clockwise would be "bad" (which might actually be GOOD for some listeners who like listening to hard clipping).
One more thing, the straight-wire-with-gain setting would be labeled "sterile." One problem is writing the marketing and manual section that explains why the "effect" control has no effect on that setting. Perhaps just say in Sterile mode the effect of the "effect" control is very subtle.
Last edited:
It would be an interesting project for sure, but I suspect a lot harder to implement than we might imagine... at least to make it fully adjustable.
Sure, but the box is a bit big to stack on the amp. Maybe in a nice matching preamp case with suitably market-oriented labels as benb suggests 😱
Are there weblinks available ? I would like to read the progresses made in this particular field.
Last edited:
I assumed my readers would realise that. However, that is not a reason to avoid even trying to get reasonably close to it.
Those who want a separate distortion box to add some nice colour should just buy a "tube buffer". These are usually just a cathode follower in a pretty box. A good one will add a little noise and low-order distortion. A bad one will add more noise and higher-order distortion. In either case they are only likely to genuinely improve the sound of a CD player with very high output impedance or an amp with rather low and non-linear input impedance. So most effect with poorly-designed partners, yet some claim this is a sign of "discrimination"!
Is there a link to Doug Self's book (Amazon etc?).
As for all amps being the same - that's not my experience at all..
You also should not take an amp away from the speaker - it is a symbiotic relationship which affects some speakers and amps more than others.
As for the 100s of opamps being in the recording path, I suspect the real number is somewhat less than 10. I'd guess on about 2 on a modern digital deck - if that.
Despite what goes on in the chain - I feel the amplifier/speaker interface is the one that primarily affects the sound you will get. Many topologies do this well, but my own feeling is that a tube amp with it's transformer is naturally the best (IIRC a McIntish SS amp also has a transformer). This is because I feel the secondary to voicecoil forms a complete and sympathetic circuit that you simply do not get with 'direct drive' transistors.
People who design amps with resistor loads are deluding themselves - connect up a real speaker (A/B switch) and watch how a square wave changes shape as you do so, especially if you can probe the inside of a feedback loop. Then when you add feedback you are allowing the speaker to control the amp (by connecting it up to the amp's input) - which in my view can give a muddled sound and lack of realism - depending upon the speaker.
As always cost is a big driver of design, rarely do audio companies care about sound above all else, it's deadlines for the 'new' model, press pictures, adverts, lead times and getting engineering to get on with it. Even in companies of over 1000 people it is rare that a single engineer will be allowed to do some pure research or new thinking (in my experience of a non-audio field).
Designing a decent amp is not rocket science, but it is beyond most accounting departments and deadline conscious managers. I'm not a fan of Naim but I like their business model - keep the design the same and just hype up the marketing, even if they start to look rather stupid when asking £500 for a mains cable to 'totally transform' ones system - a tacit admission of hype or poor PSU design.
As for all amps being the same - that's not my experience at all..
You also should not take an amp away from the speaker - it is a symbiotic relationship which affects some speakers and amps more than others.
As for the 100s of opamps being in the recording path, I suspect the real number is somewhat less than 10. I'd guess on about 2 on a modern digital deck - if that.
Despite what goes on in the chain - I feel the amplifier/speaker interface is the one that primarily affects the sound you will get. Many topologies do this well, but my own feeling is that a tube amp with it's transformer is naturally the best (IIRC a McIntish SS amp also has a transformer). This is because I feel the secondary to voicecoil forms a complete and sympathetic circuit that you simply do not get with 'direct drive' transistors.
People who design amps with resistor loads are deluding themselves - connect up a real speaker (A/B switch) and watch how a square wave changes shape as you do so, especially if you can probe the inside of a feedback loop. Then when you add feedback you are allowing the speaker to control the amp (by connecting it up to the amp's input) - which in my view can give a muddled sound and lack of realism - depending upon the speaker.
As always cost is a big driver of design, rarely do audio companies care about sound above all else, it's deadlines for the 'new' model, press pictures, adverts, lead times and getting engineering to get on with it. Even in companies of over 1000 people it is rare that a single engineer will be allowed to do some pure research or new thinking (in my experience of a non-audio field).
Designing a decent amp is not rocket science, but it is beyond most accounting departments and deadline conscious managers. I'm not a fan of Naim but I like their business model - keep the design the same and just hype up the marketing, even if they start to look rather stupid when asking £500 for a mains cable to 'totally transform' ones system - a tacit admission of hype or poor PSU design.
I think you've just outlined a couple of reasons why that may not be such a good approach to a serious "niceness" box.
Serious?...well yes. benb's focus identified a marketer's eye to me and I think most would agree that the weight of opinion (in younger folk particularly) is against earlier market demands for purer, straighter wires with gain. This is at the bottom of some irrational opposition To Self's design criteria and solutions, I suspect. Any designer who even goes some way to accommodating subjective interests is naturally going to get a more enthusiastic reception in this climate.
On forum, many threads and most tube forums by definition seek a miracle distortion mix that seems to enrich the musical experience, as it were.
If you can save yourself thousands of hours experimenting with build after build in a probably fruitless search, why not save the bucks, hours and sweat and spend it on a gadget that may encompass 90% of what you could envisage trying, even zeroing in on what might best suit you?
It seems like a smarter idea to me.
Interesting, but I would have said the complete opposite! The OPT is the biggest weakness of a valve output. Do you use the word "sympathetic" in the sense of emotional support or magic?
I agree that the speaker-amp interface is important. My view is that inherently low impedance outputs are preferable to ones which rely on global NFB. So UL rather than pentode (triode has too low an output, but I'm sure it sounds nice), and emitter follower rather than collector. Then GNFB to finish off.
The McIntosh MA 6800 uses output transformers driven by transistors. Has anyone here heard one? I recall hearing one at a hi-fi show, limited jazzy music was on offer but it did sound very relaxed and realistic to me.
I use the word sympathetic because the joining of one magnetic field to another suggests that to me. The big voltages on the anode map via a magnet-to-magnet circuit to voice coil position - via the leverage of the step-down.
I.e. the process of going from primary voltage to magnetism is coupled with the inverse process of going back from magnetism to voice coil position - with the magnetism -to- magnetism being formed by the independent and sympathetic secondary-voice coil link.
Output transformers rock 😉
Sorry, I'm not following you here. I selected a quote from what you wrote earlier. How was that 'far too selective' ?
You then added something after you read the story on Stereophile, which is:
This quote is uninteresting unless and until you supply a meaning for 'competently designed'. Please illustrate in relation to the Stereophile article - which of the two amps (in their initial states) compared was 'incompetently designed' and how are you able to tell?
Last edited:
Sure, its entirely possible. So I'll make measurements at the end of the process to see whether I've introduced more distortions. And then ponder what it might be about distortions that increase the realism. In my case, I'm designing active speakers - there are no 'straight wires with gain' in the case of active crossovers, rather they're bent by design 😛
I'm not clear what 'sterile sounding' sounds like. I don't so much hear the 'sound' of an amp myself, I hear renditions of acoustic spaces with performers in them, some of which are more convincing than others. While I do from time to time listen to what I would call 'non-acoustic' recordings, I don't use them to evaluate my designs - the ability to render a musical space is most important to me.
In regards to the 'niceness' control, based on my experience it won't fly. This is because 'niceness' isn't something that can be added to a 'nasty' signal to make the 'nastiness' go away. Rather I take it that 'niceness' is the original signal recorded - amplifiers (and DACs) that sound like they're taking out something are in fact adding in 'nastiness' rather than extracting the 'niceness'. Once 'nastiness' has been added, its impossible to take it out again.
This realism is an illusion, and if you were not in the room next to the mic when the music was recorded you are just guessing as to what it really sounded like. Moving a stereo mic a few feet will change the tone, instrument balance and reverberation a lot more than changing one good amp to another. So you are not necessarily increasing realism, you are just changing the sound to suit your taste and your own idea of realism.
Hello Bob
Have you seen this patent that presents the case of applying a DC servo after the first long tail pair ?
Worth a read
Stable distortion amplifier for audio signals - Patent 5635874
Yes, to the same extent that all 'reality' is an illusion, a projection of our minds. Do you really want to debate that - it seems to be a bit beyond your level?😀
Nope, this is nonsense. I don't guess any more than when presented with a photograph I just 'guess' what the landscape really looks like. There is no such thing as 'what it really sounded like' in many recordings anyway. They tend not to be so interesting ones to me though.
Perception is indeed a series of intelligent guesses, but these are unconscious guesses so do not qualify. To learn more you could try books such as 'On Intelligence' (Hawkins) and 'Visual Intelligence' by Hoffman.
Yes, I agree with you. I have recordings (Nimbus mostly) for which I feel the mic is too far from the instrument and I'd have moved in closer if I'd done the recording. The balance of reverb to direct sound is very important to achieve an optimal effect - this is for sure my own personal taste. But realism does not suffer by moving the mic, what changes is perspective, clarity of individual instruments. A zoom lens and a wide angle lens produce similar levels of realism but with different perspectives. Perspective is a matter for personal preference, realism is not.
I can't follow you here. What is just changing the sound - moving the mic? My idea of realism is irrelevant, so again this is nonsense. My perception of real is what matters, not my idea of it.
In regards to the 'niceness' control, based on my experience it won't fly. This is because 'niceness' isn't something that can be added to a 'nasty' signal to make the 'nastiness' go away. Rather I take it that 'niceness' is the original signal recorded - amplifiers (and DACs) that sound like they're taking out something are in fact adding in 'nastiness' rather than extracting the 'niceness'. Once 'nastiness' has been added, its impossible to take it out again.[/QUOTE]
Whilst i can understand the sentiment, aversion and particularly the difficulty in assigning characteristics of additive niceness, it is an overt, everpresent driving force behind many threads on this and particularly "toob" threads here. In the wash up of commercial exposure, your experience says you may well be correct about the viability issue but this is DIY Audio where most guys like to play with these concepts before agreeing to disagree or embark on designer-bashing forays.
My presumption (let me express it if it is not obvious) is that Doug Self's blameless model amplifier is, or is intended, to provide sterile sound. That, I take it, is the actual quality commonly reinterpreted as "nasty" i.e. --the absence of "niceness". That ought not to be an amplifier full of nasties IMO.
(asbestos suit on!)
Ian,
I am not criticising your view as such, but this is something that bothers me. As a designer, I strive to design an amp that is as transparent as possible. An amp that doesn't add or subtract from the signal as well as possible.
Am I to understand that such an amp is considered sterile? Does that mean that the source signal, unaltered, is sterile and that it is REQUIRED to modify the source signal to make it 'sound nice'?
Does that mean that all those recordings are somehow missing something for us to enjoy them?
jan didden
Relating my own experience wasn't intended to discourage others from playing with the idea of adding in 'niceness' 🙂 Its quite possible that others find distortion euphonic - so far I haven't, but then I haven't been in the design by listening game for very long.
Since it seems far from evident that Mr. Self believes in amplifiers having a 'sound' I question this presumption. I'd say he intends to produce a 'straight wire with gain' as far as possible - almost to the point of 'sound be damned'. There's an audio urban legend going around that amps built according to Mr. Self's designs sound sterile but I haven't (so far) noticed evidence for this. Myself I've fairly recently come to the view that its layout and grounding rather than circuit topologies that make an amp sound 'sterile'. So perhaps its the apparent ease with which Doug Self's work allows green hands to build amps that results in relatively poor results? Prior to his book being available, amp building from scratch required arcane knowledge and hence only experts did it.
- Status
- Not open for further replies.