A question about what is measurable.

[Opus111]

Yes, the artifacts have zero acoustic cues associated with them, hence zero depth and other spatial information - you "hear" the driver. A 'forward' system is inaccurate, because the ear can't discern, unravel the acoustic information - when the playback is very poor, it devolves to a raucous mess, the brain is overloaded with "corrupted data" - and it becomes impossible to decipher anything, almost.

You've listened to BE's files for noise modulation - did you notice the ones you preferred had more depth? Would that be one thing to listen out for - that more depth corresponds to more accuracy? Certainly I normally follow that guideline myself - its quite possible to estimate the soundstage proficiency of electronics with a single speaker in mono.

 

[Blumlein 88]

Sure, I don't disbelieve your auditory perceptions of adding the Triode but without measurements this would appear to contravene #1 of Amir's forum guidelines?

Well I don't have measured data for what caused the triode coloration. The SS amp had DC to 1 MHz bandwidth. Distortion below .05% at all power levels to 20khz up to 80 wpc. Snr better than 90 dB.

There triode had bandwidth of 2hz to 35 kHz at 1 watt. 2 hz to 20 kHz at 35 wpc. Near max power was 2% thd and varied considerably with output power. SNR I believe was 75 dB.

 

[Opus111]

My initial tentative hypothesis for the alleged 'triode coloration' would be its a source loading effect. The tube amp having a much higher input impedance than the SS. Do you have figures for the Zins of your respective amps?

 

[fas42]

You've listened to BE's files for noise modulation - did you notice the ones you preferred had more depth? Would that be one thing to listen out for - that more depth corresponds to more accuracy? Certainly I normally follow that guideline myself - its quite possible to estimate the soundstage proficiency of electronics with a single speaker in mono.

No, I wasn't specifically looking at that, the laptop speakers can develop a bit of depth if I pull out more stops, but I didn't do that with any of the listening of BE's stuff. What was distinctive was the difference in "jump factor", "musicality" - there's a very soft piano intro, which has a dullness about it in the 'inferior' version, and then an abrupt intro by full bore solo violin - which literally makes one jump a bit more in the 'better' one.

The key thing was that one lost interest in listening to the lesser one - the "this is a bit boring" factor starts to rear its ugly head ...

 

[Opus111]

What was distinctive was the difference in "jump factor", "musicality" - there's a very soft piano intro, which has a dullness about it in the 'inferior' version, and then an abrupt intro by full bore solo violin - which literally makes one jump a bit more in the 'better' one.

Interesting observations - that 'jump factor' is what I'd call 'dynamics' (and objectivists often think I'm talking about 'dynamic range') but I really mean perceived dynamics.

The key thing was that one lost interest in listening to the lesser one - the "this is a bit boring" factor starts to rear its ugly head ...

Yeah I would agree - reducing noise modulation increases the excitement factor. But could it be that there's a pathway internally between ears and emotion which doesn't pass through conscious awareness? So curiosity is stimulated even before we're consciously aware of what's doing the stimulation? This might (my memory is poor here) relate to what Walter Freeman talks about in his excellent book 'How brains make up their minds' - 'preafference'.

Doing a quick search on preafference turns up this summary (those not interested in pragmatic models of brain function look away now) - http://link.springer.com/chapter/10.1007/978-90-481-9695-1_5

 

[fas42]

Richard, I would strongly agree. When perceived dynamics or however one wishes to characterise it are strong, then I keep getting an adrenalin hit so to speak, from the listening. I can listen to the most bizarre, completely unknown piece, and I'm "grooving" to it - I instinctively feel that what's happening is musically right, is what's intended by the creators and musicians.

The other thing is that one can casually relax in its presence - I've had the current unit pounding away on some high energy piece, and she's sitting, doing Facebook on her PC, about 3 feet in front of one speaker - and doesn't ask for it to be turned down, she's just soaking it up unconsciously, without effort.

 

[Opus111]

Richard, I would strongly agree. When perceived dynamics or however one wishes to characterise it are strong, then I keep getting an adrenalin hit so to speak, from the listening.

This is crucial, and emotional impact very much is measurable as emotions are molecules in the body (see Candace Pert's excellent little tome 'Molecules of emotion'). Skin resistance presumably changes, heartbeat may well be increased. So perhaps monitoring these effects would be likely to give more sensitivity and more objectivity to listening tests?

 

[fas42]

That sounds like it could be a very interesting direction to go in ...

 

[Blumlein 88]

My initial tentative hypothesis for the alleged 'triode coloration' would be its a source loading effect. The tube amp having a much higher input impedance than the SS. Do you have figures for the Zins of your respective amps?

137k for the triode and 47kohm for the SS. I don't think my experience supports your hypothesis. I also placed the SS between source and triode. You couldn't tell it was there or not. The SS feeding the triode wasn't loaded enough to matter. The triode feeding the SS wasn't loaded enough to matter either. Unless I have misunderstood what you have in mind in that regard.

Also at the time my source component was a Wadia 25 DAC. Those had high voltage and low output impedance output circuitry. I don't think the loading would bother it either.

 

[Opus111]

137k for the triode and 47kohm for the SS. I don't think my experience supports your hypothesis. I also placed the SS between source and triode. You couldn't tell it was there or not.

Your answer contains quite a number of 'I think's - science isn't about opinion rather observations. Could you re-formulate to make it clearer what's observation and what's simply conjecture?

 

[Blumlein 88]

Your answer contains quite a number of 'I think's - science isn't about opinion rather observations. Could you re-formulate to make it clearer what's observation and what's simply conjecture?

Well the "I think" is because I am not certain of your hypothesis. You give at best a teaser. My thinking since you said source loading was partly responsible was that higher input impedances like on tube gear is easier on the source component. Lower input impedance like on some SS gear of only 10 k ohm is requiring more from the source component. Your hypethesis I guessed was along those lines.

My observations don't have "I thinks" they are quite clear. In this case, a very low output impedance connection to a power amp output stage feeding another power amp is un-fazed by the load. In either direction. My source was beefier than many sources are yet it sounded no 3d, no layered sound stage when feeding just the SS amp. Yet feeding the SS amp which buffered the tube amp showed all those traits. If the loading on the SS amp obscured layered depth it would have been gone before getting to the triode amp. Instead there was no effect.

 

[Opus111]

Well the "I think" is because I am not certain of your hypothesis. You give at best a teaser. My thinking since you said source loading was partly responsible was that higher input impedances like on tube gear is easier on the source component. Lower input impedance like on some SS gear of only 10 k ohm is requiring more from the source component. Your hypethesis I guessed was along those lines.

OK - thanks for the clarification. The hypothesis is about how the power supplies in the source component create more noise when the loading's heavier - and that doesn't need to be very 'heavy' in a traditional engineering sense. In order to make a stab at how much noise is created I'd need to know more than the input resistance, also the capacitance. The 'C' element isn't often given in the manufacturer's blurb, it normally needs the schematic. SS amps often have an RF filter (series R, shunt cap) across their input and the C value is often in the hundreds of pF. So it becomes significant at higher freqs.

My observations don't have "I thinks" they are quite clear. In this case, a very low output impedance connection to a power amp output stage feeding another power amp is un-fazed by the load.

Is the 'unfazing' an observation or conjecture?

In either direction. My source was beefier than many sources are yet it sounded no 3d, no layered sound stage when feeding just the SS amp.

Is 'beefier' observation or conjecture? Output impedance isn't an issue for this particular hypothesis, assuming its not pathological.

Yet feeding the SS amp which buffered the tube amp showed all those traits. If the loading on the SS amp obscured layered depth it would have been gone before getting to the triode amp. Instead there was no effect.

I'm not following this bit.

 

[Blumlein 88]

OK - thanks for the clarification. The hypothesis is about how the power supplies in the source component create more noise when the loading's heavier - and that doesn't need to be very 'heavy' in a traditional engineering sense. In order to make a stab at how much noise is created I'd need to know more than the input resistance, also the capacitance. The 'C' element isn't often given in the manufacturer's blurb, it normally needs the schematic. SS amps often have an RF filter (series R, shunt cap) across their input and the C value is often in the hundreds of pF. So it becomes significant at higher freqs.

I knew those numbers at one time, but it was too long ago.

Is the 'unfazing' an observation or conjecture?

it is an understanding of how those connections work.

Is 'beefier' observation or conjecture? Output impedance isn't an issue for this particular hypothesis, assuming its not pathological.

Measurement. The Wadia would put out 9 volts from a 15 ohm output impedance. Many DACs of the time would not do such. It would do that even into low input impedances. So it was what I would colloquially call beefier.

I'm not following this bit.

Be more forthcoming and complete about your hypothesis and I could respond in some manner that makes more sense.

 

[Opus111]

it is an understanding of how those connections work.

And can you set out this understanding in verbal terms? If not, best consigned to the conjecture pile.

Measurement. The Wadia would put out 9 volts from a 15 ohm output impedance. Many DACs of the time would not do such. It would do that even into low input impedances. So it was what I would colloquially call beefier.

Perfectly valid measurement but irrelevant in this context.

Be more forthcoming and complete about your hypothesis and I could respond in some manner that makes more sense.

Already did that in the post you're replying to. If you find something unclear in that, point it out. Just as I did - I mentioned the specific sentences which weren't clear to me.

 

[Blumlein 88]

And can you set out this understanding in verbal terms? If not, best consigned to the conjecture pile.



Perfectly valid measurement but irrelevant in this context.



Already did that in the post you're replying to. If you find something unclear in that, point it out. Just as I did - I mentioned the specific sentences which weren't clear to me.

Give me an example. Use the info I gave you, assume some reasonable values for the capacitances you don't know. Walk me through how this would make the two amps sound different in the way being discussed based upon what your hypothesis consists of in principle.

I'll give you my honest opinion, that my observations and info might fit your hypothesis, don't fit or that we don't know enough to determine it. That opinion based upon reasonable understandings of how such things work.

 

[fas42]

Another possibility is that the combinations of all the components, and how they were plugged into the mains altered the level and type of interference in the scenario - hence altering the distortion characteristics. From experience, absolutely everything becomes relevant, and has impact when worrying about these behaviours - unless one compares scrupulously equivalent situations, where one and one only parameter is altered between options, it is very easy to have confounders.

Fussiness about the comparison approach is essential - done halfheartedly, it is almost guaranteed to deliver spurious results - I have got myself in trouble many times by missing a key factor, in rushing to try something.

 

[Opus111]

Give me an example. Use the info I gave you, assume some reasonable values for the capacitances you don't know.

I like the approach. I'll probably run out of time in this post before I have an appointment but I'll come back later and fill in.

Firstly the SS has 47k input impedance but most likely it'll be shunted by considerable capacitance. Anything from 220pF to 1nF is possible, let's go in the middle and say 500pF.

The cable will add to this, so let's assume its decent low capacitance ones and not longer than 1m, then something like 100pF might be reasonable, I'm less sure of this figure though. Feel free to contribute data here.

As I'm unfamiliar with valves the input capacitance I don't know about, but I am assuming its quite a lot lower than for the SS amp. Would 50pF be a reasonable stab?

I'll go on to estimate the impedances at audio frequencies later.

 

[Blumlein 88]

Already did that in the post you're replying to. If you find something unclear in that, point it out. Just as I did - I mentioned the specific sentences which weren't clear to me.

Sorry, I detest the pedantic rhetorical method of discussing such things.

I will await your example to be provided rather than retorting about minute meaning differences in sentence structure.

In the mean time using what I have read from you, you guess somehow the SS amp will load the source in a way that will cause it to create some noise which obscures 3 D soundstaging. The triode amp will load it less so and with less noise from the laboring source more of such layered soundstage comes through to the speakers via the triode amp.

One connection was Wadia to triodes to speakers. Big expansive soundstage.

Another was Wadia to SS amp (loaded at output with power resistors similar to speaker loading) and dropped to unity gain(via resistor network). If the source were loaded by the SS amp, and noise from the source covered up soundstaging cues then it would be missing from the SS amp. The following triode, were it accurate, would not be able to play music with that expansive soundstage. Yet in this connection it did do that. Removing and inserting the SS amp resulted in no difference either way. This result doesn't fit with your hypothesis as I understand it.

 

[Blumlein 88]

I like the approach. I'll probably run out of time in this post before I have an appointment but I'll come back later and fill in.

Firstly the SS has 47k input impedance but most likely it'll be shunted by considerable capacitance. Anything from 220pF to 1nF is possible, let's go in the middle and say 500pF.

The cable will add to this, so let's assume its decent low capacitance ones and not longer than 1m, then something like 100pF might be reasonable, I'm less sure of this figure though. Feel free to contribute data here.

As I'm unfamiliar with valves the input capacitance I don't know about, but I am assuming its quite a lot lower than for the SS amp. Would 50pF be a reasonable stab?

I'll go on to estimate the impedances at audio frequencies later.

I don't have my tube data books handy. The input tubes were 12AT7's which probably run in the vicinity of 2 pF input capacitance.

 

[Opus111]

Another was Wadia to SS amp (loaded at output with power resistors similar to speaker loading) and dropped to unity gain(via resistor network). If the source were loaded by the SS amp, and noise from the source covered up soundstaging cues then it would be missing from the SS amp. The following triode, were it accurate, would not be able to play music with that expansive soundstage. Yet in this connection it did do that. Removing and inserting the SS amp resulted in no difference either way. This result doesn't fit with your hypothesis as I understand it.

I agree the hypothesis doesn't fit. So I'll omit updating my earlier post where I was in the process of estimating the magnitude of the loading effects.