Can amplifier speed and resolution be measured?

[SIY]

My question is what amplifier measurements can predict such behavior difference?

If it's real, then frequency response. But first, you have to determine if this is an artifact of what you're doing with the software (especially in level-matching and time-alignment), thus my suggestion to use a more appropriate tool to examine the data.

edit: Something I should have asked but neglected to (I blame early morning and insufficient coffee): can you explain exactly how you acquired the data you're looking at?

 

[chebum]

If it's real, then frequency response.

Frequency response is almost identical for both configurations. I use REW to prepare equalizer settings. For both configurations I used the same target curve. Equalizer settings were generated based on 3 averaged measurements across listening position.
These are control measurements of right channel. After applying an eqalization, I do a control measurement to check it was applied correctly. These are measurements for denon and denon+carver.

Denon is red, carver+denon is blue. Compression driver powered by Carver is crossed at 600Hz with progressive, almost brickwall "Reject Low" filter. There is basically no signal from compression driver at 400Hz.

can you explain exactly how you acquired the data you're looking at?

I have a PC that acts as a UPnP renderer, DSP and crossover. It sends signal to an av receiver via HDMI cable (6 channels in total, 3 per speaker). AV receiver powers two speakers made of RCF compression driver, Beyma midbass and Peerless sub.
I played a track in Tidal and streamed it to my DSP-PC via UPnP.
I placed a Tascam handheld recorder on a tripod in my preferred listening position. I didn't move the recorder around when switching amps. I adjusted recording volume a couple of decibels though. Recording settings was 48/24 WAV. I normalized volume of the recorded files in Audacity to -0.1 dB peak. Compared and presented the difference.

 

[SIY]

Ah, OK, this is a mike measurement? If so, don't expect much in the way of repeatability. Sorry, I should have asked sooner but my brain wasn't in gear. What you need to compare is the electrical signals being delivered to the speakers, which can be measured with far better reliability.

 

[chebum]

Ah, OK, this is a mike measurement?

Yes, I made them with a mike from LP. The difference we observe at recorded music are much more prominent than differences in frequency response at sinus sweep.

What's the cheapest device I can use to record signal from the amp output? I should connect it in parallel with the speakers, right?

 

[SIY]

A sound card will work, though depending on the voltage drive to the speakers, you'll probably want to put a voltage divider (two resistors) across the soundcard input so that you don't overload/damage it. Aim for a peak level of 1-2 V when choosing the resistor values.

 

[617]

For some reason I've never heard of people recording different amps into an ADC but it seems like a pretty good way to find audible differences.

I've never really understood the 'speed' descriptor, but suspect it's what happens when percussive rhythm is produced cleanly. I personally disagree with the notion that subjective descriptions are useless - they may not be useful in the way measurements are, and they certainly shouldn't can't be trusted for establishing cause and effect, they do give us a window into what audiophiles like and don't like in sound reproduction. Like it or not, there are still subjective decisions to be made in any system; wide vs narrow dispersion, bass level and top octave level are among them.

As one's technical knowledge builds, I think you rely less and less on subjective description; as you build a system that is more and more technically correct, it generally does sound better, or as I like to say, it has a bigger performance envelope - it can do more different things simultaneously. Nevertheless, we still offer technical explanations for the subtleties of what we hear, and these may not be accurate. Let the subjectivists spend $$$ on a new amplifier when what they really needed was a 50 cent resistor on their tweeter; that hobby is separate from this one.

 

[SIY]

No one claims that subjective evaluation is useless, what is useless is sighted subjective evaluation.

 

[ProFan]

Speed of an amplifier = slew rate + analysis of the level of TIM (transient intermodulation distortions), can be measured.
Resolution of an amplifier = dynamic range + analysis of the level of IMD (intermodulation distortions), can be measured either in dB or in bits

There is a good research regarding sound envelope precision and importance of the lowest frequency range for it (performed by Nota Bene). We can not ignore the lows and infra lows, thus the hi res sources with their frequency range starting at 5Hz are technically superior to CD quality (starting at 20Hz), no matter if somebody can not hear the difference.
The most important part of DBT is - listen to the most revealing fragments of music, not the entire songs. Understand what difference you are looking for.
It is like you came with inspection into a clean room, which looks nice and shiny, but if you will look into dark corners using enough light, for sure you will find some dirt.
Such revealing fragments are:
1. orchestra at high volume when all instruments are playing - can you define the location and sound quality for each of them? This transparency defines the total level of IMD in system (or record / format)
2. “Flowing” of lows - are they just “boomy” or surround you like waves in the ocean, flowing around? Especially revealing during big saxophone playing alone, like start of “Us and Them” in Dark Side of the Moon”, if the system is working properly one can hear like “breath” of the saxophone.
3. Decay of highs (cymbals, bells), the sound shall be Crystal clear. It is extremely difficult to get it in CD format, such as it is limited to fundamental frequency of 11kHz signal, second harmonic of frequencies above will not pass through. Thus, the hearable sound envelop will also be distorted.
And how many harmonics do we have for musical instruments above 20kHz? The answer can be found here http://www.tnt-audio.com/casse/life_above_20khz.pdf

Way more revealing fragments can be listed, those above are just examples. If we know what we are looking for in DBT, we can easier hear the difference, otherwise our brain analyzer goes into “enjoying” mode, which masks imperfections. And our musical memory is pretty short somehow.

 

[daftcombo]

It is like you came with inspection into a clean room, which looks nice and shiny, but if you will look into dark corners using enough light, for sure you will find some dirt.

Where is the audio light and the audio magnifier?
You can stay in the room at the same place in the dark for a long time without noticing the dirt. It is not a matter of concentration.

 

[ProFan]

Where is the audio light and the audio magnifier?

Your ears are your audio magnifier and light. If you listen to the music and hear some imperfection, you can hear it in DBT - I believe. Just need to set up the system properly.
The same as in that room - if everything looks nice and shiny, but some smell (or dust if that matters) comes from nowhere, you start looking for a source of it, right?
If you don’t “notice the dirt”, just enjoy your music You don’t need DBT or any test if you are completely satisfied.

 

[chebum]

Speed of an amplifier = slew rate + analysis of the level of TIM (transient intermodulation distortions), can be measured.

Are there any formal proofs of that statement?

 

[ProFan]

https://www.audioholics.com/audio-amplifier/amplifier-slew-rate
http://hifisonix.com/wordpress/wp-content/uploads/2017/10/The-Theory-of-TIM-Matti-Otala.pdf
Also I would like to add that traditionally slew rate is measured by reproduction of a square signal. Exactly like they measured THD some time ago using just one sine signal, while the contamination distortions http://www.aes.org/e-lib/browse.cfm?elib=4719 using multitone at different levels provide way more information.

The same with the slew rate - I believe in addition to traditional method it shall be measured using pulse train test signal, in this case we will also measure (and see) the sound envelope reproduction.

 

[SIY]

Otala became something of a cult despite fundamental errors on his part. Such is "high end" audio.

Transient Intermodulation Distortion Distortion (TIM) was a cleverly-devised and over-used buzz word for high frequency distortion created by amplifiers with inadequate slew rate. Its inventor incorrectly interpreted his math equations and wrote several papers asserting that TIM was caused by the use of large amounts of negative feedback. So emerged the myth that low-feedback/nofeedback amplifiers are superior. Not so, but it lives on to this day. Even a special TIM distortion test was developed, but it turns out that 20 kHz THD and TIM are highly correlated. Any amplifier that has low 20 KHz full-power THD (say, less than 0.02%) has essentially no TIM, whether it uses negative feedback or not.

There are many better measurement methods these days, but swept harmonic distortion covers the bases for any amp that isn't pathological.

 

[ProFan]

Agreed. I would not say that Otala had fundamental errors (his concept regarding the deep negative feedback as a root cause could be considered as a mistake today, but it was some kind of a fix at that time) , and created by him terminology word “TIM” still exists.
Swept harmonic distortion test allows to measure even the pulse response in the system, it is a good tool indeed: https://pdfs.semanticscholar.org/aeab/a8bb33cad2f6914cc02c76c52064d58a8136.pdf
I said “analysis of TIM”, and it can be performed by different methods, including transient process measurements (like pulse response - ringing or delay, time of decoy, etc) just to verify the lack of such distortions.

 

[SIY]

Or, as Cordell pointed out, just run full power testing at 20 kHz. Boring, but completely indicative of whether an amp will have slewing issues.

Farina's method is basic these days for measurement. Great paper.

 

[ProFan]

There is a good newer article on the subject:
https://www.edn.com/design/consumer...fiers--Why-there-is-no-such-thing-as-too-much
It looks like another “cult”, proved by results ...

 

[restorer-john]

thus the hi res sources with their frequency range starting at 5Hz are technically superior to CD quality (starting at 20Hz), no matter if somebody can not hear the difference.

This is incorrect. CD doesn't 'start at 20Hz'. Most players are flat to a few Hz with typical specs of 2-20,000Hz +/-0.5dB. The test discs we have been using for over 35 years have frequency sweeps from 5Hz-22.05KHz for frequency response tests.

 

[ProFan]

Thanks for your input, John. I definitely have to make a correction - technical superiority is related to high frequencies (here is another good example http://www.drummerworld.com/forums/showthread.php?t=66957) while for the low frequencies we have more like “historical superiority”:
“The 20Hz lower limit is likely not a technical limitation of the red book standard, but was simply added because it's the generally-accepted (and often-cited) lower limit of human hearing -- the point below which supposedly 50% of people cannot hear the sound.” Source: https://en.m.wikipedia.org/wiki/Talk:Compact_Disc_Digital_Audio
CD audio format has been created: “The selection of the sample rate was based primarily on the need to reproduce the audible frequency range of 20–20,000 Hz (20 kHz).” Nobody even thought at that time about frequency range below 20 Hz, and accordingly, audio engineers during the mastering process just cut such frequencies. Ref. the attached waterfall of Eagles “Hotel California” song (CD made in 1976). Compare with modern CDs like James Blunt’s.
Next step of cutting “useless” infra low and low frequencies was performed during so called “loudness wars”; in addition to compression majority of audio engineers limited lows again to reach the highest possible volume (lows are higher in level and tend to clip earlier).
So an enormous amount of older CDs have to be remastered (and we know a lot of remasters are in the process) while high res sources do not have that problem.

 

[solderdude]

Not speed related but IMO relevant to the LF extension which is equally important to me in SOME cases.
Anecdotal and subjective ramblings below... ignore when subjective impressions leaves you cold.

While bass response is nearly impossible to get right 'flat' with speakers in normal rooms it is possible to get it flat down to 5Hz with headphones.
I do agree that the difference between a headphone that goes down (flat) to 40Hz sounds correct with most music and most pop recordings simply seem to be cut-off in the sub-lows in the studios for a good reason.

But having some experience with correcting headphones I do find with some better recordings that bass extension (switched in/out) with a filter that only addresses bass extension so it measures down flat enhances the realism of for instance piano.

One can easily say ... BS because the lowest piano notes don't go far down enough. Here is the thing though. There is plenty of Sub-lows present because the piano itself has mechanical parts and the pressing of keys as well as pedals and the mechanisms it drives emit sub lows.
Even though it is 'present' in the better headphones already that gently roll-off below say 50Hz or so the realism that you are hearing the piano for instance simply increases. You immediately detect the difference yet you don't hear more bass nor can detect any differences with most pop recordings.

With speakers (active subwoofer that reached 20Hz flat measured directly in front of it) I also did notice an effect but not as strong.
I lived in a house with a wooden floor and had a very long room (16 meters).
There was one (classical) recording that at 1 specific recording made the whole floor go up and down noticeably when playing loud (CD) and it reproduced every time. Later I learned that the recording was made in a hall where there was an underground metro which produced that sub lows rumble.

All annecdotal of course but my point is that LF response is more important in some cases and CD is perfectly capable of reproducing it but it depends on which engineers were working on the recording and whether or not they decided to cut the sublows.

 

[ProFan]

frequency response range of musical instruments believed to be true in 1967 http://www.rfcafe.com/references/el...usical-instrument-sound-chart-aug-1967-ew.jpg
Started at 40Hz and ended at 16kHz ... with hands clapping ... Cymbals frequency range ended at about 12-15 kHz.
Take a look at the cymbals’s frequency range and compare with JEJ’s measurements ... Compare where bass notes ended on their graphs and for example the attached waterfalls frequency range on James Blunt’s music waterfalls ...
For sure we know that it is just an average example - there are way more sources with real high level of lows, expanded into infralow range - Lorn, for example.