[killkli]

Quote:

It seems to be mainly distortion, rather than EQ. If you subtract the signal of one file from the other, the difference sounds shifted up by one octave, which suggests 2nd harmonics. But it is a small difference, so it is not surprising many people voted that the files sound the same.
 

Wow@@
You got my works right  ^_^.
 
However, the difference is quite big but you can't really recover the distortion by subtracting, since distortion mixing changed the original signal.
 
One distortion % is about 10%, the other is around 1%.
 
Reasonable it should be the lesser distortion, the better. But there are also a frequency distributing pattern involved.

 

[killkli]

Quote:

 
That does not simulate intermodulation distortion.
 

As I've stated before, I want the problem to be simpler. Solve one before another.
So I need to know how the simple harmonic distortion pattern effect the listener.
 
btw, IMD is much harder to simulate. To do that would require some programming, which is beyond my power .
 
 
 

 

[stv014]

Quote:

Originally Posted by killkli /img/forum/go_quote.gif
 
One distortion % is about 10%, the other is around 1%.

 
The difference between the two files looks more like 1%, although that can already be audible.
 
Quote:

Originally Posted by killkli /img/forum/go_quote.gif
 
btw, IMD is much harder to simulate. To do that would require some programming, which is beyond my power .

You can simulate both by applying simple non-linear distortion. For second harmonics, you need to square the original signal (this can be combined with filtering for a more complex frequency dependent effect).
 

 

[firev1]

Quote:

 
The current result is quite illustrating.
 
 

Humans like distortion? There is one other study on the effects of distortion on recorded music and the study suggest that even the artist themselves feel that distortion expresses their work in a more musical manner. 

 

[Deadeight]

I went with the option I did because it didn't sound as cold, not sure how to describe it. 

 

[2000impreza]

honestly, I could not hear a difference after listening to each file twice. I listened with my Audeze LCD-2 rev-2. 

 

[killkli]

Quote:

honestly, I could not hear a difference after listening to each file twice. I listened with my Audeze LCD-2 rev-2. 

Yep, indeed it may really be not such a big difference.
Therefore I put the third option there ^^.
 
 

 

[morks]

I think I know which one was tubed. Let's see if my hearing/sound memory betrays me...

 

[Magick Man]

Personally, it seems a bit obvious, but then I really enjoy tube sound.

 

[killkli]

I'm sorry for the delay in putting the answer up.
Here it comes:
 
A file got more distortion.
The distortion rate is about 20% for A file, and around 1% for B file.
All distortions are 2nd distortion.
 
However, there is another difference here:
A file's distortion sound has an EQing as shown in the picture below:

 
B file also got an EQ, it's here:

 
 
My theory is quite simple, while in actual listen measurement, SS amp got very low distortion rate, but it's unstable. because it's low distortion, and prone to be effected by small factors such as power & etc..
 
While it's very low in simple loop-back measurement (without headphone or speaker on,) it can be perceivable in actual listening.
Actual distortion rate of a normal headphone is around 0.5%, lower in flagship such as HD800, but generally it's about 0.5% or higher.
For me, my usual listening level is around 70~80dBL, so the distortion is about 70 + 20(log0.005) = 23dBL loud. It's not loud but you can tell it's there.
And with the instability mention above, the sound would be a small but much distorted sound.
 
With tube amplifier, it's different since it was designed with higher distortion rate, with 1% distortion in loop back measurement and around 5% in headphone measurement, it's so loud that the small factors such as unstable power, etc., 
is of no much consequences in these amplifier's design.
Therefore, we here a distortion sound level of 70 + 20(log0.05) = 44dBL distortion sound, but it's more consistent with the original signal.
In the case of 2nd harmonic distortion, it's actually one octave above the original music, and which can make kind of harmony with the melody.
 
So here's my guess: in higher designed distortion amp such as tube amplifier, the distortion is produced as a stable constant, so it actually augmented the original sounds, but surely distorted more.
In SS amplifier, we can lower the distortion to such a low level that it may not be perceivable. But it's hard to have THIS low distortion, and with most headphone & amplifier combination,
the distortion is low be still heard, and due to it's low level, it's very difficult to keep it stable, so the distortion is not melodically accorded to the original music.
Therefore, with not so great headphone & not so great SS amp, it's more probable to have a clear but somehow grainy and harsh sound come out.
As for most high distortion designed tube amp, it's quite easy to come forward with pleasant sound, but it may be difficult to have the resolution of the well matched SS system.
 
 
If the distortion rate of an SS amp & it's headphone could be kept under 0.05%, then there can really no fault in my listening.
Which is the case with my current system of HE500.
I really love the sound of HE500 with Benchmark DAC1 & blackcube combination. The measured THD rate across all frequencies are almost all under 0.05% under my listening level (70 or 80dBL.)
Which means that the distortion sound is about 3~13dB, one can hardly heard this kind of sound.
 
 
Well..... it's still much of my own opinion, but the test did show a much higher inclination toward the constant higher distortion character of tube amp to the character of SS amp.
Hope this could be as illustrative as it can be.
 
 

 

[Tilpo]

As far as I know most tube amps still have quite low distortion figures. It is true that some tube amps intentionally add distortion to create a specific sound signature, but definitely not all tube amps do this.

Furthermore you seem to be implying that the higher distortion somehow masks other bad characteristics such as noise. I think this is a wrong assumption, and it doesn't make any sense whatsoever to me.
In fact, I suspect that amps having very high harmonic distortion will most likely have higher noise levels than designs with low THD.

I think you are misled by the fact that tube amps with high THD have a high amount of even order distortion. The ear usually perceives even order distortions as pleasant, and therefore a high THD amp can sound more pleasant than a low THD one.
Also, IIRC tube amps create far more even order distortion than odd.


Finally, I wonder how you measured THD from your HE-500. Did you use the specifications of the headphone, or did you actually measure it? If the latter is the case, then could you elaborate on your measuring procedure. Just out of pure interest, of course.

 

[killkli]

Quote:

As far as I know most tube amps still have quite low distortion figures. It is true that some tube amps intentionally add distortion to create a specific sound signature, but definitely not all tube amps do this.
Furthermore you seem to be implying that the higher distortion somehow masks other bad characteristics such as noise. I think this is a wrong assumption, and it doesn't make any sense whatsoever to me.
In fact, I suspect that amps having very high harmonic distortion will most likely have higher noise levels than designs with low THD.
I think you are misled by the fact that tube amps with high THD have a high amount of even order distortion. The ear usually perceives even order distortions as pleasant, and therefore a high THD amp can sound more pleasant than a low THD one.
Also, IIRC tube amps create far more even order distortion than odd.
Finally, I wonder how you measured THD from your HE-500. Did you use the specifications of the headphone, or did you actually measure it? If the latter is the case, then could you elaborate on your measuring procedure. Just out of pure interest, of course.

Please don't use suspection but actual measuring.
I measured my system with RME's babyface interface and a pair of binaural microphone.
 
As for your question, I can tell you. Higher distorted system did masked the disturbance of most of the effect of the enviroment.
How do I know this?
Here's the measured result I got from higher distortion rate tube system and low distortion SS system:
SS （Black Cube）

 
tube (quite cheap one, high distortion rate):

 
So it's quite clear that the distortion rate is not very stable under blackcube. If I increase the sampling time to 1 second then it would smooth down very much like this:
 

 
But it's not so stable when the sampling time is only 0.2 second. each time I do the experiment, the result would be slightly different.
But it's not so with Tube system. Tube system is constant, whether I sample it with 0.2 second or 1 second.
 
 
Please don't just suspect..... I put my question forward because I had the experience that somehow I liked lower end system with tube more then lower end ss system.
And while I'm measuring the things, I got this kind of result.
 
You mentioned that SS's distortion is very low, I do admit that. But it's not low enough to be not heard with headphone on.
In loop back recording, it's indeed so low that anyone say they can heard such distortion is nonsense. But with headphone or speaker drivers working with them, the problem is not so simple.
The distortion is perceivable, because it has not gone over the threshold, which would be around -60dB(0.01%) for 70dBL listening level.

 

[killkli]

BTW, I'm not talking about the GREAT tube amps that are so expensive.
I'm talking about east easy designed & you can see everywhere around tube amps. They generally do have much higher distortion rate than the same priced SS amp.

 

[stv014]

Quote:

Originally Posted by killkli /img/forum/go_quote.gif
 
So it's quite clear that the distortion rate is not very stable under blackcube. If I increase the sampling time to 1 second then it would smooth down very much like this:

 
Are you sure you are not just measuring the noise of your microphone, and that is why it does not look stable ?

Of course, it will look more "stable" with low end tube amps, because the actual distortion is so high then that it outweighs the microphone noise. A good, but not expensive solid state amplifier (like the $150 O2) can have <= 0.01% distortion 20 Hz to 20 kHz even with a low impedance load.
 

 

[killkli]

Quote:

 
Are you sure you are not just measuring the noise of your microphone, and that is why it does not look stable ?

Of course, it will look more "stable" with low end tube amps, because the actual distortion is so high then that it outweighs the microphone noise. A good, but not expensive solid state amplifier (like the $150 O2) can have <= 0.01% distortion 20 Hz to 20 kHz even with a low impedance load.
 

 
But you can't see such figure when actually matching with a headphone.
Tell me, did you see ANY headphone stating their THD lower thatn 0.01%?
HD800 perhaps..... which is indeed very best I've ever heard in this case. But I don't have them for measurement.
 
As for the noise problem, measured noise is around -80dB to -120 dB (higher in lower frequency).
While the signal of 80dBL measure with the microphone is around -20dB, which means that -80dB's noise is about 0.01% distortion. So measured distortion higher than 0.01% is not of the effect of the noise.
 
this picture actually shows the noise floor of my microphone system:

D3~D6+ is probably just the noise of the microphone.
 
the lowest I measured with my own system is HE500, but it's still around 0.02~0.03%.