Are tubes more musical?

[atmasphere]

I have offered three possibilities, all of which have met up with gaffaws:

1) model the Tiger amp
2) model the Adcom GFA565 in bridged mode
3) since my main assertion has been and is still that DvsF can cause brightness, I proposed a model where 1KHz was the cutoff frequency and distortion rises 6dB/octave above that frequency as indicated in this paper

 

[GXAlan]

I have offered three possibilities, all of which have met up with gaffaws:

1) model the Tiger amp
2) model the Adcom GFA565 in bridged mode
3) since my main assertion has been and is still that DvsF can cause brightness, I proposed a model where 1KHz was the cutoff frequency and distortion rises 6dB/octave above that frequency as indicated in this paper

It seems like #3 is approaching feasibility since @pkane has done some modification to Distort?

 

[Blumlein 88]

I have offered three possibilities, all of which have met up with gaffaws:

1) model the Tiger amp
2) model the Adcom GFA565 in bridged mode
3) since my main assertion has been and is still that DvsF can cause brightness, I proposed a model where 1KHz was the cutoff frequency and distortion rises 6dB/octave above that frequency as indicated in this paper

You quoting that Putzeys paper is ironically hilarious.

So you are good with any of the above models. Good we are getting somewhere.

So I assume SIY can proceed with the one he wishes to use from these three. @SIY

 

[SIY]

Let’s go with the Tiger since Ralph called it “painful.” And we have no measurements on bridged 565.

 

[pkane]

Maybe we can do Tiger next. For now, here's a simulation of the GFA565 bridged, based on the specs in the manual. Note that I didn't attempt to hit the values for THD exactly, but these are close and the general shape of THD vs Frequency should be similar.

Here are the files, one is the original at 96k, the other sent through DISTORT transfer function:

File A:
https://app.box.com/s/9y8uj60pjxxz7vnexl24ppll810xtl9c

File B:
https://app.box.com/s/kbk4itp9gy4i5rtunbwepau8i9iwzagv

Here are the DISTORT simulated harmonics at 1kHz:

THD is about 18dB higher at 10kHz:

File A and B compared in DeltaWave for any obvious differences:

 

[GXAlan]

File A and B compared in DeltaWave for any obvious differences:

What does the pkMetric show?

 

[pkane]

What does the pkMetric show?

 

[Multicore]

I can't hear a difference between gfa565-a and gfa565-b. I like the graphs.

 

[pkane]

I can't hear a difference between gfa565-a and gfa565-b. I like the graphs.

So in your case, the difference in rising THD at higher frequencies is inaudible. Let's see if @atmasphere finds one of these files "brighter" than the other in a controlled test. Note that there is no difference in frequency response, if you judge by the graphs.

 

[Multicore]

So in your case, the difference in rising THD at higher frequencies is inaudible.

So it seems. Btw, since your message does not say which file is the one and which is the other, I took it as a blind test. Presumably one could unbind by analyzing the files but I didn't.

 

[pkane]

So it seems. Btw, since your message does not say which file is the one and which is the other, I took it as a blind test. Presumably one could unbind by analyzing the files but I didn't.

The idea was to provide an unbiased comparison, so you could test yourself - just like you did. If you were to hear obvious differences, the next step would be to do a proper blind test, ABX or other, to be sure. There’s no point in doing a blind test if you can’t hear any differences sighted .

 

[davmol]

I can't hear a difference between gfa565-a and gfa565-b. I like the graphs.

Here are the results of my golden ears:

Smashed it

 

[Pluto]

What is it that I’m hearing in my tube preamp and tubed headphone amp that is so romantic, warm, soft around the edges, and smooth, even syrupy sometimes?

It is this…

 

[pkane]

Here are the results of my golden ears:
View attachment 398693
Smashed it

Almost a perfect score!

 

[Blumlein 88]

It is this…

I think this is closer to the truth. I don't find most tube pres to have much if any sound. Tube amps yes, and I do think it is a number of things, but compression is some of it.

You probably know this, but for anyone who doesn't a nice little article about the Fairchild 660 and 670 compressors.

Fairchild 660 & 670

These 1950s compressors are probably the most revered — and imitated — processors in the world. But what is it that makes them so special?

www.soundonsound.com

 

[atmasphere]

You quoting that Putzeys paper is ironically hilarious.

So you are good with any of the above models. Good we are getting somewhere.

So I assume SIY can proceed with the one he wishes to use from these three. @SIY

I had a weird day yesterday- a reaction to a bee sting. They don't get easier as you get older!

If you think its ironic, my surmise is you only skimmed my posts, you didn't read Bruno's actual article at the link or both. The mechanism for distortion rise with frequency is clearly explained in that paper.

I figured out how to install Lacinato on my desktop - now to get it running on the shop machine we use to play tunes.

 

[Blumlein 88]

I had a weird day yesterday- a reaction to a bee sting. They don't get easier as you get older!

If you think its ironic, my surmise is you only skimmed my posts, you didn't read Bruno's actual article at the link or both. The mechanism for distortion rise with frequency is clearly explained in that paper.

I figured out how to install Lacinato on my desktop - now to get it running on the shop machine we use to play tunes.

And you'd be wrong. I read Bruno's article carefully when it first came out. I re-read it again a few times. I also read your posts. I even skimmed thru the article to see if I remembered it wrongly after your posts.

Condolences on the bee sting. Hope you are well from it and continue to be so.

 

[atmasphere]

And you'd be wrong. I read Bruno's article carefully when it first came out. I re-read it again a few times. I also read your posts. I even skimmed thru the article to see if I remembered it wrongly after your posts.

Condolences on the bee sting. Hope you are well from it and continue to be so.

Thanks!

The issue is in the chart on page 3. When the loop gain slope intersects the STF, feedback decreases so distortion goes up. One solution is greater Gain Bandwidth Product so the loop gain can be pushed up further. This puts the intersection with STF out to a higher frequency. Or do I have this wrong?

 

[Blumlein 88]

Thanks!

The issue is in the chart on page 3. When the loop gain slope intersects the STF, feedback decreases so distortion goes up. One solution is greater Gain Bandwidth Product so the loop gain can be pushed up further. This puts the intersection with STF out to a higher frequency. Or do I have this wrong?

No I agree. I agree that is why some SS amps have rising distortion with frequency. I agree that is why the effect is more common and larger with most of the commercially available tube amps. I'd think not yours, but I don't recall seeing measurements of yours this way.

 

[atmasphere]

No I agree. I agree that is why some SS amps have rising distortion with frequency. I agree that is why the effect is more common and larger with most of the commercially available tube amps. I'd think not yours, but I don't recall seeing measurements of yours this way.

Good- that was really the point of my linking Bruno's paper. Since our tube amps are zero feedback they don't have rising distortion (also since they are OTLs). Our class D has plenty of gain bandwidth product so it doesn't have rising distortion with frequency either.