3e Audio A5 Stereo Amplifier Review

[UnderpathBum]

Can I just say I find both extremes in this hobby equally funny?

Extreme A: "I swapped one op-amp, night and day! Soundstage now infinite! My wife heard it from the kitchen!"
Extreme B: "Without double-blind, everything you so clearly heard simply... evaporates."

One side thinks an op-amp swap is a miracle, the other thinks any difference you hear with your eyes open is all in your head. Both sound like audiophile ghost stories to me.

Back to my case — after swapping back and forth multiple times, I still think discrete Class-A buffers right before the TPA3251 PFFB input get a bit closer to what I consider real instruments to sound like, compared to IC op-amps. This is only with my equipment, in my room, with my ears, and with my dog wandering in the yard. lol

This will be my last post on this Class-A discrete vs IC op-amp topic. I'm going back to music, which is the whole point.

 

[RandomEar]

Thanks for the links and for explaining your view - I did read Sean Olive's blog before, and I agree sighted bias is real. I didn't do a level-matched blind ABX, it was just sighted listening in my room, so your caution is fair.

But I think there are three issues with your conclusion:

1. Topology, TIM, thermal memory are not meaningless buzzwords. They are real engineering concepts documented by Self, Groner and others. Whether they are audible at -110dB is debatable, but to say they have no relation to performance at all is incorrect. ICs use extensive internal current mirrors and heavy compensation to get that 0.000015% number - that architecture behaves differently from a discrete Class-A circuit that is linear open-loop first.

They are meaningless in this context, because you try to attribute differences which have not been established to a handful of abstract concepts or terms which are loosely connected to the tech being discussed. From a readers perspective, your post tries very hard to sound technical and complex while obscuring the problems that

1. Audible differences for sure have not been established and likely don't exist
2. All the mentioned abstract concepts are assumed to be important for these assumed differences without expanding in any form on how or why they would be
3. All the mentioned abstract concepts are assumed to have an audible effect on the output which - again - is assumed to somehow be important in this case

None of this is a given, though. I'm always weary of such posts, because in essence, they are close to technobabble. Just saying the word "topology" without any context doesn't imply that it is relevant for the problem at hand. It is nothing more than a concept and - sadly - I have seen it used a hundred times by audiophiles here on ASR as kind of a "shield" to defend their position: "But the topology is different!" - I don't care. What matters aren't fancy words, what matters is the measured performance at the speaker terminals. If that is identical, you can put jello-based PCBs in the little black box for all I care.

There's ample posts about hard measured data for op amp swapping here on ASR and based on that data, it is prudent to asume that most op amp swaps change nothing audible and some degrade the performance massively. An example of a very good discrete but still worse-than-integrated op amp would be this one. Being discrete is not an advantage for op amps. Quite the contrary. It's just one of those words audiophiles like so much again.

2. The A5 is well measured, but not with my op-amps in it. Amir's test shows the stock OPA1612 unit is excellent, much more than just SINAD. But it didn't test IMD vs frequency, multitone, or 19+20k with discrete op-amps installed. You can't conclude that an unmeasured modification is inaudible from the stock measurements. That's absence of evidence, not evidence of absence.

Naturally, when we are discussing the measured performance as shown in the OP, we are discussing the stock op amps. In any case, I can't conclude that your modifications are inaudible but I'm also not the one making the claims here.

3. "Below threshold therefore magic fairies" is a strawman. I never said the THD is big and audible. I said two circuits can both be below 0.001% at 1kHz yet have different harmonic profile and loop-gain at high frequencies, which we weight differently than a single 1kHz number. The thresholds you quote are for detecting an added tone, not for timbre. Saying low THD automatically equals correct timbre is itself an assumption.

You were listing thresholds. I quoted you. Maybe this is a mixup. In any case, as I said in my last post: There's more than 1 kHz THD / SINAD in the OP. The data is clear. If the FR is flat, the distortion is below any relevant level, the noise is inaudible and the power is as specified, what could you possibly expect to improve? If the signal is amplified without relevant audible errors, why would [insert any audible quality you would like here] not be correct? Call it timbre or sound stage or whatever.

So I'm not claiming this as scientific proof, just sharing my experience as an owner: same detail, less grain, fuller mids. If it's placebo, it's a cheap, hot-running, enjoyable placebo. No hard feelings, I'm happy to agree to disagree and just enjoy the music.

It's fine to enjoy the tinkering and the music. But on a science-minded forum, posting subjective opinions is kinda expected to generate some blowback. I mean, if you do post, at least offer some proper test results. Anything to substantiate the claims. What is there to discuss otherwise?

 

[RandomEar]

[...]
A question:
Beyond any additional distortion, in part, is he describing a simple shift in tonality from a modest change in the linear output from the dac? Do we have any data from the dac that might show us a slight top end rolloff or a midrange boost? If it does do this it could easily be described as presenting an "improved" midrange. This would be such as with that lowered, or "softened" top end, the mids would be RELATIVELY more elevated/"improved"/"present" etc. Often folks describe hotter mids as more forward or more detailed or some such thing. In general this is of course at the expense of accurate sound reproduction. It is a shockingly common practice with speaker demos to boost the mids or lower treble wherein you will hear statements from an average shopper such as "I've never heard such incredibly detailed mids"; of course you haven't. Because they are artificially boosted by 3 dB from 3 kHz to 6 kHz or something similar.

We're talking about amplifiers, not DACs In any case, since there is no established difference here, I'm hesitant to speculate on what change in which frequency region would cause which change in the impression. In general, I would agree that "showroom sound" is a thing and that sellers do have their tricks to make hardware sound more impressive for a moment - but possibly annoying if listened to for a longer time. I guess this is more of a discussion for one of the speaker threads.

I appreciate you bearing with me here. I aid folks on another forum and in my efforts to understand what they are looking for, what they prefer and do and do not like and desire, I find that I spend a tremendous amount of time trying to understand what they are actually saying. It is made more difficult by nebulous audiofool words presented to me that may or may not mean what I think that they mean. It is NOT easy to figure out...
I do this in an effort to ascertain if the sound of a particular speaker will suit them. If they like some particular sound profile, such as a typical B&W sound for example, then I can not suggest speaker XXX. I have no interest in forcing someone to like a neutral accurate speaker, as I do, if that is not what brings them joy. Of course I always look to show someone what accurate sound reproduction is, as in uncolored and adding little or nothing to the source signal. I come from the school of thought where my goal is to hear what the studio engineer heard that day when he sat behind that console.
Give me a straight wire with gain any day. If a particular mood hits me I can always add a simple PEQ to suit that mood then back to clean, uncolored sound it is.

Thanks again for bearing with me here, off I go on a short 450 mile drive to pick up some speakers. Looking forward to any informative replies later upon my return.

I think ASR tends to be on the "reproduce sound accurately or go home" train for the most part. At least when it comes to electronics (DACs, pre-amps, amps). Nobody here will dispute that speaker selection is somewhat subjective due to interactions with your room and different preferences in qualities like directivity. I am just mostly annoyed when people claim to hear major improvements from tinkering with the elements in the audio chain which are the least problematic. Bonus points if they didn't test the supposed differences correctly.

 

[RandomEar]

Can I just say I find both extremes in this hobby equally funny?

Extreme A: "I swapped one op-amp, night and day! Soundstage now infinite! My wife heard it from the kitchen!"
Extreme B: "Without double-blind, everything you so clearly heard simply... evaporates."

One side thinks an op-amp swap is a miracle, the other thinks any difference you hear with your eyes open is all in your head. Both sound like audiophile ghost stories to me.

While I don't disagree with that observation, the reality is that one side has loooots of evidence which supports their general point of view:

Does Op-amp Rolling Work?

When I post my last video on review of Douk A5 amplifier, I mentioned that rolling (changing) op-amp ICs in there is fruitless. I got have a dozen comments under that video in youtube asking why so I thought I experiment again with the A5. Note that I have done the same testing with both DACs...

www.audiosciencereview.com

3e Audio TPA3251 TPA3255 Finished Amplifier with PFFB is coming!

No additional features, no sub out, no nothing. No power supply options to save 10$ - just offer the biggest one that really supports the specifications - and these should be honest. Overall, a simple product that focuses on quality and power for an affordable price, and just works Fully agree...

www.audiosciencereview.com

Op-amp Rolling Using AIYIMA A07

I have made my homework and re-measured AIYIMA A07 with NE5532, LME49720 and OPA2134. My measuring system has been improved since the time of my first measurements of the A07. However, no breakthrough findings, but good enough to debunk some strange myths. I measured the A07 supplied from SMPS...

www.audiosciencereview.com

Fosi Audio V3 op-amp rolling, has anyone tried it? Snake oil? Or are there actual differences?

A plastic cup holds water. A glass looks better, has better specs in certain areas... but still it just needs to hold water, which it does fine be it crystal, Ikea glass or plastic. And sometimes the "worse" component, a plastic cup, is just what was specified for the job and lasts longer and...

www.audiosciencereview.com

Sparkos SS3602 Opamp Rolling In Fosi P4

This is a detailed measurements of effect of swapping the stock NE5532P op-amp in Fosi P4 preamplifier with Sparkos SS3602 "discrete" opamp. The SS3602 costs US $79.80 and the NE5532P, US $0.47. Swapping opamps in P4 is more challenging due to back connectors/switches that need to be...

www.audiosciencereview.com

Sparkos SS3602 Opamp Rolling In Fosi P4

1% is four orders of magnitude, even less audible. You also need to take the Klippel distortion test to inform yourself how much distortion is actually audible. You can't pass 80dB even with a pure tone. If you can, please post your results. Someone pointed this out to you earlier when you...

www.audiosciencereview.com

 

[Ropeburn]

Extreme A: "I swapped one op-amp, night and day! Soundstage now infinite! My wife heard it from the kitchen!"
Extreme B: "Without double-blind, everything you so clearly heard simply... evaporates."

One side thinks an op-amp swap is a miracle, the other thinks any difference you hear with your eyes open is all in your head. Both sound like audiophile ghost stories to me.

Typical centrist fallacy. Those two are anything but equal. Case A is a bold claim that by 90% or more really evaporates under controlled testing. There's shittons of reasearch and data supporting it. While Case B is a simple exaggeration, and not a big one. As an absolute statement it's false, then again those making it usually don't mean it as an absolute but leave room for exceptions. By 90% or more it's true though. Sighted bias is a big thing influencing our brains significantly, effortlessly so to the point of illusion and hallucination. That's normal and an integral part of how the brain works. It is always worthwhile to eliminate that as far as you can. And only then you even have a chance to get to the physical truth. Without eliminating that bias you will never get reliable data, and even in the rare case of there really being audible differences, you can't know whether it is true or you just imagined it - because the testing method logically doesn't allow for that!

Saying these two statements and attitudes are equal in any way is simply false. They so aren't. You can't just say: 90% true and 90% untrue is the same as 50/50.

 

[Almdudler]

I am working in audio as engineering professional since 25+years with a degree in Acoustics/Psychoacoustics. @RandomEar: yes, those extremes exist, they are "stereotypes", so let me attempt to further detail:
Extreme A:
Loves listening to (or creating) music but is often victim to Dunning-Kruger type of audio marketing. Has two ears they trust, little they understand that these ears are connected to a brain making perceptual decisions that can depend on outside environmental factors and mood... Sometimes those folks are not aware of their ear/hearing deficiencies either. Quite often those folks have no scientific education and come from the artistic and creative side of music.
Extreme B:
Loves data and science but often has non-audio specific background, so audio is a hobby. Those folks apply those scientific methods, but not always in a meaningful way. Quite often convinced that if data does not provide evidence, certain perceptual effects do not exist. Interestingly that type of extreme/stereotype loves to elaborate about amp SINAD of 120dB vs 114dB (inaudible) while their speakers make typically 1% or more distortion (can be audible).

 

[antcollinet]

As an absolute statement it's false,

And it is not the statement I made. @UnderpathBum constructed it as a straw-man by joining up two parts of two different sentences into one. This is the sort of disingenuous style we are dealing with here.

See:

3e Audio A5 Stereo Amplifier Review

Yeah. It's exactly what you add distortion in music for. If the sound is changed it has been distorted. Tube amps add distortion, for example. Not all distortion sounds "bad". But if an amp measures very similar to another it is almost certain to sound the same. Certainly you wouldn't hear the...

www.audiosciencereview.com

1 - Use of the qualifier "most likely"

and 2 - "It has been demonstrated over and over again" - it has.


@UnderpathBum - if you want to quote me - please use the quoting system and quote exactly what i said word for word complete with context. And without joining up two parts of two sentences as though they are one thus changing the meaning. Please also attribute the quote - which the quoting system does for you. What you have just done is rude. And dishonest.

(Edited to remove profanity)

 

[testp]

after swapping back and forth multiple times, I still think discrete Class-A buffers right before the TPA3251 PFFB input get a bit closer to what I consider real instruments to sound like, compared to IC op-amps. This is only with my equipment, in my room, with my ears, and with my dog wandering in the yard. lol

can you mask them identically, maybe try to do some AB testing that way, 10..20 swaps marked your guess on paper, gives some good data

 

[MaxwellsEq]

My point was a bit different, and actually your graph shows it: THD rises significantly with frequency. At 1kHz it's -100dB, but at 15kHz it is already around -65 to -70dB at higher power.

OK, some basics about understanding total harmonic distortion and frequency:

(1). With a pure 100Hz tone I can measure potentially audible harmonics which should not be there at 200Hz, 300Hz, 400Hz etc. in fact, I can measure many harmonics of the fundamental. If I total* them all up within a measurement window of, say 22kHz, that's a lot of values within the audible range.
* which is where TOTAL Harmonic Distortion comes from

(2). With a pure 1kHz tone I can measure potentially audible harmonics which should not be there at 2kHz, 3kHz, 4kHz etc. Although I can measure several harmonics of the fundamental, THERE ARE FAR FEWER OF THEM IN THE AUDIBLE RANGE. If I total them all up within the 22kHz measurement window I would need to realize there are fewer values in the total than at 100Hz

(3). With a pure 6kHz tone I can ONLY measure potentially audible harmonics which should not be there at 12kHz and 18kHz. Now if I total them all up within the 22kHz window, there are only 2 terms of interest!

(4). With a pure 12kHz tone I can measure ZERO potentially audible harmonics which should not be there because the first harmonic is 24kHz, which is inaudible AND outside my measuring window! What is THD in this circumstance? Zero?

(5). Now supposing instead I had used a measuring window out to, say 50kHz. For the 12 kHz fundamental, I would have total-ed up harmonics at 24kHz (inaudible), 36kHz (inaudible) and 48kHz (inaudible). So my THD result would appear worse than if my window was 22kHz, but this number would be meaningless.

Consequently, rising total harmonic distortion at high frequencies is largely irrelevant and sometimes inaudible. In many ways, 7kHz is the highest you should measure 3rd harmonic distortion and 10kHz the highest you should measure 2nd harmonic distortion. At 100Hz you can measure 10th harmonic distortion if you fancy.

Sometimes, you need to understand the graphs and the numbers! For high frequency THD you need to know the measurement window to understand if the distortion is even audible.

Oh, and your hearing sensitivity is very low at 10kHz and masked by pretty much any sound at 2kHz.

 

[Ropeburn]

OK, some basics about understanding total harmonic distortion and frequency:

(1). With a pure 100Hz tone I can measure potentially audible harmonics which should not be there at 200Hz, 300Hz, 400Hz etc. in fact, I can measure many harmonics of the fundamental. If I total* them all up within a measurement window of, say 22kHz, that's a lot of values within the audible range.
* which is where TOTAL Harmonic Distortion comes from

(2). With a pure 1kHz tone I can measure potentially audible harmonics which should not be there at 2kHz, 3kHz, 4kHz etc. Although I can measure several harmonics of the fundamental, THERE ARE FAR FEWER OF THEM IN THE AUDIBLE RANGE. If I total them all up within the 22kHz measurement window I would need to realize there are fewer values in the total than at 100Hz

(3). With a pure 6kHz tone I can ONLY measure potentially audible harmonics which should not be there at 12kHz and 18kHz. Now if I total them all up within the 22kHz window, there are only 2 terms of interest!

(4). With a pure 12kHz tone I can measure ZERO potentially audible harmonics which should not be there because the first harmonic is 24kHz, which is inaudible AND outside my measuring window! What is THD in this circumstance? Zero?

(5). Now supposing instead I had used a measuring window out to, say 50kHz. For the 12 kHz fundamental, I would have total-ed up harmonics at 24kHz (inaudible), 36kHz (inaudible) and 48kHz (inaudible). So my THD result would appear worse than if my window was 22kHz, but this number would be meaningless.

Consequently, rising total harmonic distortion at high frequencies is largely irrelevant and sometimes inaudible. In many ways, 7kHz is the highest you should measure 3rd harmonic distortion and 10kHz the highest you should measure 2nd harmonic distortion. At 100Hz you can measure 10th harmonic distortion if you fancy.

Sometimes, you need to understand the graphs and the numbers! For high frequency THD you need to know the measurement window to understand if the distortion is even audible.

Oh, and your hearing sensitivity is very low at 10kHz and masked by pretty much any sound at 2kHz.

This needs to be pinned somewhere. It's fundamentally important to understand, especially the part about certain harmonics being outside hearing range. When we look at distortion graphs and are like "oh, the distortion rises badly at 10kHz and above" we need to realise this is entirely inconsequential for everyone (save a few exceptions) over 20 or so.

 

[jlx]

We're talking about amplifiers, not DACs

quite right sir, I should have said "is he describing a simple shift in tonality from the amp caused by a modest change in the linear output of the op amp? Do we have any data from the amp that might show us a slight top end rolloff or a midrange boost caused by the op amp swap?
My bad. Sorry for that dopey error.
It was early and my morning coffee had not fully reached by brain

 

[peng]

1. Topology, TIM, thermal memory are not meaningless buzzwords. They are real engineering concepts documented by Self, Groner and others.

I don't think anyone on ASR would say TIM, is not engineering concepts so not surprised if Self had "documented..." but did he say that the effects would be discernible/audible above a certain threshold?

 

[peng]

This needs to be pinned somewhere. It's fundamentally important to understand, especially the part about certain harmonics being outside hearing range. When we look at distortion graphs and are like "oh, the distortion rises badly at 10kHz and above" we need to realise this is entirely inconsequential for everyone (save a few exceptions) over 20 or so.

I would caution that while in cases where harmonics may be outside the hearing range, the non harmonics, such as those resulted from IMD may well be within the hearing range and in that case, whether such distortions is consequential or not would depend on factors including the magnitude, frequencies, ones' hearing ability/loss etc. My point it, when THD is really low, such as below -85 dB, all might be great, but when it gets down to higher than say -75 dB/0.03% then we might have to consider more than effects of the "harmonics".

 

[MaxwellsEq]

I would caution that while in cases where harmonics may be outside the hearing range, the non harmonics, such as those resulted from IMD may well be within the hearing range and in that case, whether such distortions is consequential or not would depend on factors including the magnitude, frequencies, ones' hearing ability/loss etc. My point it, when THD is really low, such as below -85 dB, all might be great, but when it gets down to higher than say -75 dB/0.03% then we might have to consider more than effects of the "harmonics".

My post 1069 only covered unwanted harmonics of pure sine waves, and essential caution about how they are added to get TOTAL Harmonic Distortion. As frequency increases, THD makes progressively less sense and is is essentially audibly meaningless beyond 10kHz, especially if you don't know the measurement window.

I was responding to post 1054 from UnderpathBum which was in itself a comment on one of Amir's THD+N graphs in post 1053 from Peng. The graph shows increasing THD+N with increasing frequency from the amplifier with which this thread started. UnderpathBum thought this showed something significant. I was helping them understand that the significance was less relevant than they thought.

I made no observation about Intermodulation Distortion. It is a different beast having a complicated relationship to a pure frequency tone. Most importantly, Amir's graph, which I was giving advice about, only shows THD+N, not IMD.

Can you recommend any peer-reviewed research demonstrating audibility of IMD compared to audibility of THD?

 

[peng]

My post 1069 only covered unwanted harmonics of pure sine waves, and essential caution about how they are added to get TOTAL Harmonic Distortion. As frequency increases, THD makes progressively less sense and is is essentially audibly meaningless beyond 10kHz, especially if you don't know the measurement window.

I was responding to post 1054 from UnderpathBum which was in itself a comment on one of Amir's THD+N graphs in post 1053 from Peng. The graph shows increasing THD+N with increasing frequency from the amplifier with which this thread started. UnderpathBum thought this showed something significant. I was helping them understand that the significance was less relevant than they thought.

I made no observation about Intermodulation Distortion. It is a different beast having a complicated relationship to a pure frequency tone. Most importantly, Amir's graph, which I was giving advice about, only shows THD+N, not IMD.

Can you recommend any peer-reviewed research demonstrating audibility of IMD compared to audibility of THD?

I always wonder if IMD would become a potential issue if THD gets higher than say -70 dB (just a number I pick). I don’t recall reading peer reviewed papers on audibility of IMD but any of those I read about tis topic, they all pointed to the fact that it is worse than harmonic distortions and the threshold would be much lower. I am interested to read some of those papers too and will post link if I find an interesting one.

Asked Gemini the question and got the following:

Research into human perception shows that intermodulation distortion (IMD) is generally more audible and annoying than harmonic distortion (THD). This happens because harmonic distortion creates new frequencies that are musically related to the original sound, which our brains easily hide or blend in. IMD, however, creates unnatural, disharmonious tones that stick out. [1, 2, 3, 4]
Several peer-reviewed papers explore these audibility thresholds and how they compare.

Key Peer-Reviewed Papers

• "Audibility of Transient Intermodulation Distortion"
  • Authors: Petri K. Gabrielson et al. (Published via IEEE Xplore).
  • What it found: This paper specifically looked at transient intermodulation distortion (TIM). It tested how well everyday listeners and audio engineers could spot it in music. The most sensitive listeners could hear as little as 0.5% momentary IMD, proving that our ears are incredibly sharp at catching non-harmonic changes. [1]
• "Is High-Frequency Intermodulation Distortion a Significant Factor in High-Resolution Audio?"
  • Authors: J. Robert Stuart, Richard Hollinshead, and Michael Capp (Published in the Audio Engineering Society E-Library).
  • What it found: This study investigated whether ultrasonic frequencies (sounds above 20 kHz that we cannot normally hear) interact inside high-quality speakers to create lower-frequency IMD that drops into our hearing range. They confirmed that while this distortion can sometimes affect very precise hearing tests, it does not ruin typical high-resolution music playback. [1, 2]
• "Perceptual Assessment of Distortion in Low-Frequency Loudspeakers"
  • Authors: Louis Fielder (Published in the Audio Engineering Society E-Library).
  • What it found: Fielder is famous for mapping out exactly how human hearing "masks" (hides) distortion. This paper uses complex math to show why harmonic distortion components are easily covered up by the music itself, whereas certain intermodulation frequencies escape this masking and become highly noticeable. [1, 2]
• "In-Vehicle Audio System Distortion Audibility versus Harmonic Distortion and Intermodulation Distortion"
  • Authors: Various researchers (Published in the Audio Engineering Society E-Library).
  • What it found: This practical study compared standard %THD tests directly against IMD tests inside real listening environments. It highlights why old-school THD numbers do not always match what people actually hear, and how IMD provides a much more accurate picture of perceived sound quality. [1, 3]

Summary Comparison of Audibility

Distortion Type	Sound Behavior	Audibility Threshold	Brain Reaction
Harmonic (THD)	Multiplies the original tone (e.g., 100Hz creates 200Hz, 300Hz).	High (1% to 10% depending on the sound).	Pleasing or easily ignored because it mimics natural instruments.
Intermodulation (IMD)	Mixes multiple tones together (e.g., 100Hz and 500Hz create 400Hz and 600Hz).	Low (Often under 0.5% in sensitive ranges).	Harsh, muddy, and easily noticed because the tones do not belong musically.

And Wiki cited a couple of reads, no time to search/read those yet:

Intermodulation - Wikipedia

en.wikipedia.org

Due to the characteristics of the human auditory system, the same percentage of IMD is perceived as more bothersome when compared to the same amount of harmonic distortion.[2][3][dubious – discuss]

 

[antcollinet]

the non harmonics, such as those resulted from IMD may well be within the hearing range and in that case, whether such distortions is consequential or not would depend on factors including the magnitude, frequencies, ones' hearing ability/loss etc. My point it, when THD is really low, such as below -85 dB, all might be great, but when it gets down to higher than say -75 dB/0.03% then we might have to consider more than effects of the "harmonics".

The impact of IMD though is shown in the multitone test...

Impact of rising distortion at high frequencies is evident - but still remaining below a nicely inaudible -100dB

 

[peng]

The impact of IMD though is shown in the multitone test...

Impact of rising distortion at high frequencies is evident - but still remaining below a nicely inaudible -100dB


View attachment 553909

Thanks for pointing that out, so far, as I said before, just wondering about the potential issue but don't really know much about it in terms of quantifying the issue.
Obviously as many have mentioned before, both are related to non linearity so if there is harmonic distortions, there are going to be intermodulation distortions and if THD is high, IMD will likely be high. So I want to know if THD, such as shown in the THD vs Frequency vs Output level, and if that is low enough, would IMD be not low enough? I have to wonder because while both distortions are correlated, IMD apparently is worse sounding at the same level, say 1%.

I intend to search and read more, now that I got interested in the topic again, but from what I read so far, it is not likely going to be an issue, for amps measured as good as that 3 e Audio A5.

 

[antcollinet]

Thanks for pointing that out, so far, as I said before, just wondering about the potential issue but don't really know much about it in terms of quantifying the issue.
Obviously as many have mentioned before, both are related to non linearity so if there is harmonic distortions, there are going to be intermodulation distortions and if THD is high, IMD will likely be high. So I want to know if THD, such as shown in the THD vs Frequency vs Output level, and if that is low enough, would IMD be not low enough? I have to wonder because while both distortions are correlated, IMD apparently is worse sounding at the same level, say 1%.

I intend to search and read more, now that I got interested in the topic again, but from what I read so far, it is not likely going to be an issue, for amps measured as good as that 3 e Audio A5.

The important point with respect to HD tones intermodulating back into the audible spectrum, the HD tones are already tiny. Even in your example, at -70dB. So when they intermodulate the IMD tones are going to be similarly tiny relative to the already tiny HD tones.

Take the IMD test result below from this review: The highest IMD tone is at -80dB below the two tones that are inter-modulating (and bear in mind these are at the top end of the frequency range where distortion is known to be worse). So even if you have two HD tones at -70dB inter-modulating, the result of that is going to be most likely down in the region of -120+ dB. Basically irrelevant.

In the multitone test the "grass" we see, is the combined result of HD of each individual test tone and IMD of every pair of test tones.

IMD resulting from the HD tones will be well down, lost in the "noise" of that chart.

 

[peng]

The important point with respect to HD tones intermodulating back into the audible spectrum, the HD tones are already tiny. Even in your example, at -70dB. So when they intermodulate the IMD tones are going to be similarly tiny relative to the already tiny HD tones.

Take the IMD test result below from this review: The highest IMD tone is at -80dB below the two tones that are inter-modulating (and bear in mind these are at the top end of the frequency range where distortion is known to be worse). So even if you have two HD tones at -70dB inter-modulating, the result of that is going to be most likely down in the region of -120+ dB. Basically irrelevant.

In the multitone test the "grass" we see, is the combined result of HD of each individual test tone and IMD of every pair of test tones.

IMD resulting from the HD tones will be well down, lost in the "noise" of that chart.






View attachment 553941

My original point to Ropeburn was of cautionary nature as I stated clearly, and also reacting to MaxwellsEq's post that seemed to emphasize THD only, such as "Sometimes, you need to understand the graphs and the numbers! For high frequency THD you need to know the measurement window to understand if the distortion is even audible." I was thinking at the time, when it comes to distortions, it is also important to understand IMD and we should not just assume harmonic frequencies above 20 kHz can be ignored because of the resulting IMD that may, or may not be audible to some music listeners.

Your posted graph shows the side bands below 20 kHz are at near -90 dB, that's very good, but it is the 3e A5 we are talking about, no doubt this amp is one of the better one that ranks quite high even among the amps Amir placed in the blue zone. I have nothing bad to say about the A5. My point about the potential IMD concerns perhaps should have been made with reference to another amp to avoid being misunderstood it was about the A5. Likewise, if MaxwellEq and Reopburn's posts were in a thread of a lesser amp, my cautionary note that is about the general concepts regarding THD and IMD in general, might have been better made in a separate thread, instead of reacting to Maxwell EQ and Ropeburn's that obviously were also about the concepts, not the specific amp.

Just for example, if we look at another relatively well measured, and Amir recommended Fosi V3amp:

If we focus on THD, we could also say something like, don't worry about the blue curve because even the 2nd order is 20 kHz, that most of us over 30 could not likely hear... yet if we look at the 19&20 kHz curve, we can see that the magnitudes of those "grass" are approaching my randomly picked -70 dB comfort zone that the THD vs frequency show. I actually have no concerns about the IMD of the V3 either, and I listen to it daily with my LS50 desktop speakers using DL ART, very nice sounding amps! But they are in Amir's recommended list and did also land in the green zone on the SINAD ranked chart.

We should keep in mind that Amir did those 19&20 kHz measurements at Full scale @5 W, that is good for practical reasons but I think most would agree that it would not likely represent the worse or near worse conditions even when listening to music. The graphs does not show the total IMD level either, and the total naturally would push the number a little higher.

Back to my original point, sorry if anyone thought I was cautioning about the A5 amp, no I have no concerns there. In fact I own a pair of A7 mono amps, and also a few Fosi V3 amps, and enjoy both. I do believe that to some people who have excellent hearing, especially those who happen to be of the near ocd type or SINAD chasers, probably would be interested in focussing also on IMD measurements, and if measurements are not available, try to estimate/project from the THD vs frequency measurements that Amir usually include. If on that graph, the black and blue ones shows THD around -70 dB, I would probably be slightly alerted, but that's just me, practically speaking I would take a guess that even -60 dB might be fine, in terms of audibility anyway. I would concede that the concerns about IMD vs THD might seem overrated, but so is SINAD, still better to know about it than to ignore it, imo.

I am still trying to read up on the thresholds for THD vs IMD topic, until then, my understanding is that for the same distortion level, say at around -60 dB (to be more conservative) THD would not be as bad as -60 dB for IMD.

Lastly, about people who potentially might have a higher threshold of audibility, here's one example, @SampleSizeOne , Lol:

3e Audio A5 Stereo Amplifier Review

Yeah. It's exactly what you add distortion in music for. If the sound is changed it has been distorted. Tube amps add distortion, for example. Not all distortion sounds "bad". But if an amp measures very similar to another it is almost certain to sound the same. Certainly you wouldn't hear the...

www.audiosciencereview.com

 

[MaxwellsEq]

My original point to Ropeburn was of cautionary nature as I stated clearly, and also reacting to MaxwellsEq's post that seemed to emphasize THD only, such as "Sometimes, you need to understand the graphs and the numbers! For high frequency THD you need to know the measurement window to understand if the distortion is even audible." I was thinking at the time, when it comes to distortions, it is also important to understand IMD and we should not just assume harmonic frequencies above 20 kHz can be ignored because of the resulting IMD that may, or may not be audible to some music listeners.

Your posted graph shows the side bands below 20 kHz are at near -90 dB, that's very good, but it is the 3e A5 we are talking about, no doubt this amp is one of the better one that ranks quite high even among the amps Amir placed in the blue zone. I have nothing bad to say about the A5. My point about the potential IMD concerns perhaps should have been made with reference to another amp to avoid being misunderstood it was about the A5. Likewise, if MaxwellEq and Reopburn's posts were in a thread of a lesser amp, my cautionary note that is about the general concepts regarding THD and IMD in general, might have been better made in a separate thread, instead of reacting to Maxwell EQ and Ropeburn's that obviously were also about the concepts, not the specific amp.

Just for example, if we look at another relatively well measured, and Amir recommended Fosi V3amp:

If we focus on THD, we could also say something like, don't worry about the blue curve because even the 2nd order is 20 kHz, that most of us over 30 could not likely hear... yet if we look at the 19&20 kHz curve, we can see that the magnitudes of those "grass" are approaching my randomly picked -70 dB comfort zone that the THD vs frequency show. I actually have no concerns about the IMD of the V3 either, and I listen to it daily with my LS50 desktop speakers using DL ART, very nice sounding amps! But they are in Amir's recommended list and did also land in the green zone on the SINAD ranked chart.

We should keep in mind that Amir did those 19&20 kHz measurements at Full scale @5 W, that is good for practical reasons but I think most would agree that it would not likely represent the worse or near worse conditions even when listening to music. The graphs does not show the total IMD level either, and the total naturally would push the number a little higher.

Back to my original point, sorry if anyone thought I was cautioning about the A5 amp, no I have no concerns there. In fact I own a pair of A7 mono amps, and also a few Fosi V3 amps, and enjoy both. I do believe that to some people who have excellent hearing, especially those who happen to be of the near ocd type or SINAD chasers, probably would be interested in focussing also on IMD measurements, and if measurements are not available, try to estimate/project from the THD vs frequency measurements that Amir usually include. If on that graph, the black and blue ones shows THD around -70 dB, I would probably be slightly alerted, but that's just me, practically speaking I would take a guess that even -60 dB might be fine, in terms of audibility anyway. I would concede that the concerns about IMD vs THD might seem overrated, but so is SINAD, still better to know about it than to ignore it, imo.

I am still trying to read up on the thresholds for THD vs IMD topic, until then, my understanding is that for the same distortion level, say at around -60 dB (to be more conservative) THD would not be as bad as -60 dB for IMD.

Lastly, about people who potentially might have a higher threshold of audibility, here's one example, @SampleSizeOne , Lol:

3e Audio A5 Stereo Amplifier Review

Yeah. It's exactly what you add distortion in music for. If the sound is changed it has been distorted. Tube amps add distortion, for example. Not all distortion sounds "bad". But if an amp measures very similar to another it is almost certain to sound the same. Certainly you wouldn't hear the...

www.audiosciencereview.com

View attachment 554021

View attachment 554022

I think it's still valuable to consider what is happening with the IMD chart above. You get about a -75dB spike at about 17.5 kHz by feeding 0dBA signals at 19 and 20kHz. Because of human hearing sensitivity, we cannot hear 17.5kHz at -75dBA.