Sound Quality Vs. Measurements

[nigel pearson]

I am DVV . We have live music at Woodstock UK today so my visit here is breif .

TTFN as we said in 1942 ( Ta ta for now , is Welsh for bye bye )

[nigel pearson]

As nobody seems to be doing this thread , how about this . Do valves/tubes add useful colour to the sound and make it nice ? I find they don't if we say to a reasonable hi fi standard ( DIN45500 ) . When they do it , it is not my cup of tea . Some valve amps were very low distortion . Accepted wisdom says we should not be hearing coloration if below 0.1% THD espeacailly when class A amps . Doubtless low damping factor is a coulouration and might sometimes be heard as boom or poor frequency response . Also transformers or circuits with poor frequency response . To me voice quality is better and distortion can be low . Camouflage is not what they do , or do well if they do . Marantz Model 9 followed by the Dynaco ST70 as low colouration designs of note .

[gootee]

Originally Posted by nigel pearson

As nobody seems to be doing this thread , how about this . Do valves/tubes add useful colour to the sound and make it nice ? I find they don't if we say to a reasonable hi fi standard ( DIN45500 ) . When they do it , it is not my cup of tea . Some valve amps were very low distortion . Accepted wisdom says we should not be hearing coloration if below 0.1% THD espeacailly when class A amps . Doubtless low damping factor is a coulouration and might sometimes be heard as boom or poor frequency response . Also transformers or circuits with poor frequency response . To me voice quality is better and distortion can be low . Camouflage is not what they do , or do well if they do . Marantz Model 9 followed by the Dynaco ST70 as low colouration designs of note .

?House of Fire?: Firebottles And Groove Tubes Versus Devices That Find Their Origins in Sand (Part 1)

Transistors, Tubes Sound The Same, It?s The Circuit Topologies That Differ

[gootee]

Here's a crazy idea: What about actually trying to find ways to advance the state of the art, in terms of discovering or developing (or just sharing) measurements or methods that could be used to help predict "sound quality", with the goal being something like "developing some understanding and some methods that could be useful in improving everyone's ability to repeatably generate better audio system designs"?

I "accidentally" posted a rough initial outline of one possible type of scheme for a process for trying to do that, somewhat collectively, earlier tonight, starting at the fourth paragraph of post 878, which is at:

http://www.diyaudio.com/forums/solid...ml#post3145013

I also posted a method for measuring the distortion of only the tops and bottoms of a pulse train, within an LT-Spice simulation, which is illustrated at

http://www.diyaudio.com/forums/solid...ml#post3142618

In that case, it was only for use as a tool to examine the output error levels of automated sequences of simulation runs where PSU reservoir capacitance was varied, in order to more-easily be able to find the largest capacitance value at which the power supply could still "mangle" the output waveform, by any amount.

But I plan to eventually try to transform that type of processing method into something that can be used to help evaluate any simulated system's transient response accuracy (i.e. not just the tops and bottoms of the pulses), especially for automated sequences of simulation runs where various circuit parameters are step-swept through ranges of values.

Measuring Transient Response accuracy in different ways should be important because: a) steady-state performance measurement is already relatively easy and b) low sine distortion is relatively easy to achieve, and c) low steady-state (sine) distortion seems to not be a good predictor of "sound quality", anyway (although I would probably want to assume that having some acceptable steady-state distortion level would be necessary but not sufficient), and (also: ) therefore the Transient Response must be involved.

Before anyone pooh-poohs spice modeling as a tool for such purposes, think about how repeatable it is. People far apart can test and compare with exactly the same circuit, rather than with different implementations of the same circuit. That might be important in speeding up a development process by leveraging through the use of multiple people, connected by the internet (or even simply multiple instances of a circuit, being simulated on multiple PCs at the same location). Simulation is cheaper, faster, and much more repeatable and controllable. I do understand that there are also many limitations.

Chers,

Tom

[dvv]

Originally Posted by nigel pearson

As nobody seems to be doing this thread , how about this . Do valves/tubes add useful colour to the sound and make it nice ? I find they don't if we say to a reasonable hi fi standard ( DIN45500 ) . When they do it , it is not my cup of tea . Some valve amps were very low distortion . Accepted wisdom says we should not be hearing coloration if below 0.1% THD espeacailly when class A amps . Doubtless low damping factor is a coulouration and might sometimes be heard as boom or poor frequency response . Also transformers or circuits with poor frequency response . To me voice quality is better and distortion can be low . Camouflage is not what they do , or do well if they do . Marantz Model 9 followed by the Dynaco ST70 as low colouration designs of note .

Careful with the damping factor, Nige. Accepted wisdom has it that it should be as high as possibel, i.e. the output impedance should be as low as possible, but from personal experience I know that this is not quite so linear.

For a start, there's the question of how was that DF obtained. Could be through multiple output series/parallel devices, through lots of global NFB, etc, and I find that some devices with a nominally smaller damping factor actually sound better than others with a higher damping factor, and in the bass region, too.

As for tubes, and I now risk my life from tube camp hit squads, I find that far too many of them sound like a call girl telling you in the morning that last night you were the world's bets lover ever - you know she is lying, but it's good to hear. They sound mostly "syrupy" to me, sticky, too sweet.

On the other hand, I do not believe this is the way it has to be, rather that this is what has come to be accepted as "tube sound", as what is customer anticipated. After all this time, I feel reasonably certain that say Waveborn could do a nicely neutral tube amp I could really enjoy without the usual "sticky" character of commercial audio. I have heard a tube amp from an Italian company (I forget the name) which was supremely neutral and had an uncanny tendency to disappear sonically, leaving you and the music. Its nominal DF was just 10:1, and let me assure you, it had no problems whatsoever with the bass.

As most things in audio, everything is very much relative.

[dvv]

@gootee

Tom, I haven't read the other thread yet, but I will. That said, I find that a reasonable rule of thumb (no science, just quick'n'dirty) for transformers is to use the number of watts you need from it in any case, add say 15-20% reserve, and multiply the number of watts by sq. root of 2, i.e. 1.41, for the required VA rating. Thus, for 200W/4 Ohms, I'd use (200*1.2)1.41, or 338 VA. Since I cannot have just that locally, only 300 or 400 VA, I'd go for the 400 VA. More never hurts.

As for the capacitors, 10,000 uF is good enough for 100W/8 Ohms, so for 200W/4 Ohms, as I see it, I'd need double that, as the current is double the initial.

What I'd like to know is how actually useful would it be to switch from one bridge rectifier for both + and - sides, as compared to using two full wave bridge rectifiers, one for the + side, and the other for the - side. Any thoughts?

[nigel pearson]

I think damping factor is interesting and undiscovered by most . Add a 0R47 to a modern amp and see what you get . Keep going up to 8R if you like . The amp might be very happy with 3R3 also . If theory is correct ( 1950's ) then that it where we start to notice if an 8R load ( boom ) . Remember most 8 R speakers are not exactly that so 2R2 might be more appropriate . If a speaker sounds a bit lean and dry it might help ?

[dvv]

Originally Posted by nigel pearson

I think damping factor is interesting and undiscovered by most . Add a 0R47 to a modern amp and see what you get . Keep going up to 8R if you like . The amp might be very happy with 3R3 also . If theory is correct ( 1950's ) then that it where we start to notice if an 8R load ( boom ) . Remember most 8 R speakers are not exactly that so 2R2 might be more appropriate . If a speaker sounds a bit lean and dry it might help ?

Yes, but you also get, in the same package tour, bass boom, overhang and whatnot. God knows I love bass dearly, but I can't stand imprecise and uncontrolled bass.

0R47 I can stomach, but 3R3 is, in my view, far too high, it will deteriorate the DF to below 10:1, which is the acceptable minimum for speaker Q control.

Anyone remember the first (Kensonic) Accuphase integrated amp from the early 70ies? Well, it had a damping factor switch contro at the back, allowing for a DF of 66, 12 and 6, if memory serves. I remember the 66 or 12 test was a good one, but if memory serves, nobody was happy with 6.

[tvrgeek]

I thought it was pretty well established damping factor was one of the least important parameters.

[Wavebourn]

Originally Posted by tvrgeek

I thought it was pretty well established damping factor was one of the least important parameters.

...when particular speaker design does not rely on certain electrical damping. Most of modern woofers do, so damping factor for them is needed, however not of crazy numbers, just enough to damp resonances on main resonant frequency.