[lamode]

https://community.klipsch.com/forums/storage/3/1107724/01162904.pdf
 
This article is now almost 40 years old, so this issue is hardly new, yet amplifier reviews still ignore TIM in measurements and insist on measuring performance of steady sine wave signals.
 
Given that the attack envelope of an instrument plays an important role in psychoacoustics (perhaps more than harmonics) the dynamic performance of an amplifier can't be ignored.
 
Here is an article about measuring TIM: http://hifisonix.com/wordpress/wp-content/uploads/2010/10/Measuring-TIM.pdf.pdf
 
(interestingly, the article is also from 1977!)
 
Then there is the "Leach" amp, a low-TIM design, which has been updated over the years, now in version 4.5: http://users.ece.gatech.edu/mleach/lowtim/
 
Has anyone heard the Leach? I wonder if anyone ever tried this design approach with a headphone amp.

 

[Speedskater]

That horse died decades ago.
 
Most reviewers are not interested in serious testing. Neither in measurements nor real listening tests.

 

[prot]

That horse died decades ago.

Most reviewers are not interested in serious testing. Neither in measurements nor real listening tests.

When you say "most" does it mean that you know some magazines/reviewers who actually do meaningful measurements? Used to check HIFICritic (no ads, measurements, some DBTs) but a few months ago some genius there decided to investigate 'storage sound' and ended up with this magical quote:

"the SSD sounds cool, sophisticated, pure, smooth, vital, almost crystal clear .."
―― Martin Colloms

I'm fresh out of magazines to read after that catastrophe and asking for help.

P.S.
apologies to the OP for derailing the thread a bit. Never saw that measurement but it surely looks like something I would like to have in a review.

 

[stv014]

A relevant article that might be worth reading: "The F-word" by Bruno Putzeys.
 
It is actually not a problem if sine waves are used for testing amplifiers, as long as there is a measurement with the combination of the highest frequency and output level for which the amplifier is designed. Therefore, a 20 kHz sine wave or a mix of 19 and 20 kHz tones at just below clipping level can be sufficient to show if there is an issue with signals that have a high slew rate, as long as the input can be expected not to have significant content above 20 kHz (otherwise, the frequency of the tone(s) needs to be adjusted accordingly).
 
A test with high bandwidth square waves may show higher distortion, but it is not a realistic input because of the high amount of ultrasonic and even RF harmonics. On the other hand, for this reason a non-oversampling DAC with no reconstruction filter can be a problematic source for some amplifiers. A well designed amplifier uses a lowpass filter on its input to minimize issues related to unwanted RF content.

 

[bigshot]

  TIM occurs with dynamic changes in the signal, so no it is not measured at all using standard test tones. There is a good article about it here: https://community.klipsch.com/forums/storage/3/1107724/01162904.pdf
Here is an article about how to measure TIM: http://hifisonix.com/wordpress/wp-content/uploads/2010/10/Measuring-TIM.pdf.pdf

 
Oh man! I wish I could get something out of those, but they are WAY too dense for me.
 
I'm curious about it though. Does the distortion just occupy a short time before the amp recovers? How long typically? Does this only affect specific frequencies? If so, which ones? How loud does the distortion get? I saw in that first article -40dB being mentioned before it got beyond me. Is that typical? What kind of a transient are we talking about to cause this? A snare drum hit? Cymbals? Why have I never been able to hear this? Is it something that is much smaller than the typical distortion in speakers in a room?
 
I am specifically interested in the affect that TIM has on listening to music on home equipment. Unfortunately, these papers focus on the why, not the what.

 

[Speedskater]

In audio circuits, TIM is a poorly defined (or incorrectly described) sub-set of SID (Slew Induced Distortion).  There are good tests for SID.

 

[bigshot]

  A relevant article that might be worth reading: "The F-word" by Bruno Putzeys.

 
That article was a little more helpful. But I wish people would set aside some space in their article for "real world" examples.
 
When I was investigating jitter, I hit the same brick walls all the time. Articles dished out tons of complex theory about how jitter was caused, but no one ever pointed to an audio component and said, "Here is an example of a unit with high jitter, let's do a listening test and prove jitter is a problem for music lovers." All they could do was heap on more theory to try to discredit the listening tests that proved that proved jitter was inaudible. At the end of a week of dense reading and research, I realized the whole thing was a lot of hot air. I hope that isn't the case here.
 
OK. Assuming that TIM actually is a problem... Every solid state amp I've ever owned (three in the past fifteen years or so) has sounded *exactly* the same with two channel playback. If TIM was a problem with all three of them, is it likely that an integrated amp by Sanyo, an AV receiver by Sony and an AV receiver by Yamaha all have the exact same effects of TIM? Or am I not hearing the effects because my speakers and room, or the music I am listening to aren't able to accurately reproduce the problem?
 
And if TIM is a problem, what does it sound like in music? Does it affect some frequencies more than others? Does the distortion only occupy a short period of time?

 

[lamode]

   
Why have I never been able to hear this? Is it something that is much smaller than the typical distortion in speakers in a room?

 
Well, I'm sure you have

 

[lamode]

  In audio circuits, TIM is a poorly defined (or incorrectly described) sub-set of SID (Slew Induced Distortion).  There are good tests for SID.

It's not just about slew rate but the interaction between the slew rates of various components in the negative feedback loop (afaik)

 

[Speedskater]

It's not just about slew rate but the interaction between the slew rates of various components in the negative feedback loop (afaik)

Did you read and do you understand the Bruno Putzeys paper?
This whole TIM thing is ancient history and so is the discussion about errors in the TIM theory.

 

[bigshot]

 
Well, I'm sure you have

 
Then it always sounds exactly the same, regardless of the design, brand or model of the amp?

 

[lamode]

Quote:

  A relevant article that might be worth reading: "The F-word" by Bruno Putzeys.

 
It seems that his maths is oversimplified. It doesn't take into account any phase shift occurring due to capacitance in the transistors, for example.
 
If you use negative feedback which has been phase shifted, this won't cause a big problem for constant amplitude sine waves. The output will still be a sinewave, even if the amplitude and phase will be affected slightly. There is no evidence of distortion on the scope. However the first part the sinewave will be distorted before the system achieves stability. With a chaotic input (music) the system will never achieve stability and there will always be distortion.
 
This distortion is what I mean by TIM (and perhaps that's not the right name).

 

[bigshot]

What time frame does the phase shift reach? And how much time are we talking about for the few cycles? Are we talking mileseconds, nanoseconds?

 

[lamode]

  What time frame does the phase shift reach? And how much time are we talking about for the few cycles? Are we talking mileseconds, nanoseconds?

 
Cycles was the wrong word. I edited my post. Its not important for music anyway, as we don't tend to sit at home listening to sine waves. The point is that this is a form of distortion which is not measured if you use steady sine waves.
 
I do believe that an amplifier's SQ is capable of being expressed with numbers, but our currently measurement methodology falls far short.

 

[stv014]

   
It seems that his maths is oversimplified. It doesn't take into account any phase shift occurring due to capacitance in the transistors, for example.
 
If you use negative feedback which has been phase shifted, this won't cause a big problem for constant amplitude sine waves. The output will still be a sinewave, even if the amplitude and phase will be affected slightly. There is no evidence of distortion on the scope. However the first part the sinewave will be distorted before the system achieves stability. With a chaotic input (music) the system will never achieve stability and there will always be distortion.

 
The article does discuss the "slow amplifier with a lot of feedback" myth. Also, difference extraction can be used to test amplifiers and other devices with music or any other complex signal. There are some examples here (pages 2 and 3; do you hear anything that you think is TIM ? Note that the difference has been amplified by a factor of 200). Ultimately, the feedback loop does not know or care about the difference between music and sine waves. What matters the most is the input level and how fast it changes. As long as these are within reasonable limits, the distortion can be kept at an acceptable level.