+1.
I also want to add that you can get lots of 'plug-ins' nowadays that you can load in your fav replay system, to get any sound you want. There's plug-ins for tube sound, for guitar sound, even for phonograph-sound. For the guitar plug-ins you can even select the guitar brand, the amplifier type (Marshall, Fender, ...) and even speaker you want to emulate. The phono plug-in makes all your hires digital files sound like LP records, with all the pops and scratches you want.
It'a great world out there ;-)
Jan
Last edited:
Agreed, I don't get the subjective movement that claims that the ears can hear what the instruments can't measure, if it can be heard, it can be measured. However, I must agree that THD is not the entire story, 5% THD due to clipping doesn't sound as bad as 5% THD due to crossover distortion, but the end result is that there is 5% THD and that it can be measured. The consensus should be that any distortion is bad in a hi-fi amp, regardless of its nature, and one should try to minimize it.
I personally have a hard time hearing 1% THD on anything but a sine wave, with a sine wave things are easy, music is more forgiving, either way, music these days is so compressed and hard limited that the inherent distortion of the recording is already high. That being said, I think that an amp with less distortion is inherently a better amp than one with higher THD, since audio and music is so subjective one should try to quantify it, and that is by taking measurements not by how good it sounds to you. I have yet to see an amp with extremely low THD and IMD that sounds 'bad'.
Now, if we are talking guitar amps, then all bets are off, since distortion is part of the guitar sound. In that case I personally do prefer tubes to transistors, and you'll have a hard try convincing a musician otherwise, human perception in that case should be taken into account, just the same way as you would judge a violin of a certain brand vs another brand.
Last edited:
I don't agree with this. If this were true, how does one explain the popularity of the First Watt line of amps. None of them measure particularly well against modern HiFi amps. Yet they are clearly well regarded.
Also look at Sakis's threads (user east electronics) on Rod Elliott's P3a. He has repeatedly said that his customers will favor his loaner P3a over the low-distortion amp that he is repairing for them. The P3a is another amp that doesn't measure well comparably.
This means that people's tastes vary. Some prefer ultra-low distortion. Others like a splash of low order even or odd harmonics. Some want their "added sound" in the line stages... others prefer it in the power amp.
And finally, this is a hobby. It should be enjoyable. So ignore what everyone tells you and do what makes you personally happy.
I like the idea of building things using discrete components and prefer to avoid op amps. Not because I think they're bad or inferior. I just enjoy the DIY aspect of discrete designs.
Also look at Sakis's threads (user east electronics) on Rod Elliott's P3a. He has repeatedly said that his customers will favor his loaner P3a over the low-distortion amp that he is repairing for them. The P3a is another amp that doesn't measure well comparably.
This means that people's tastes vary. Some prefer ultra-low distortion. Others like a splash of low order even or odd harmonics. Some want their "added sound" in the line stages... others prefer it in the power amp.
And finally, this is a hobby. It should be enjoyable. So ignore what everyone tells you and do what makes you personally happy.
I like the idea of building things using discrete components and prefer to avoid op amps. Not because I think they're bad or inferior. I just enjoy the DIY aspect of discrete designs.
Everyone will like whatever Nelson Pass makes just because its made by Nelson Pass, I am not saying that they are bad, just that the Nelson Pass name has a weight on it. As for the rest, I would like to see hard figures, saying "some guy says people like the bad amp more than the good one" is just rumours, under which conditions was the test conducted, was it a double blind test? or just a listening among friends and beers?
Last edited:
You make some good points that touch on important aspects of the subjective vs objective debates that have been hammered at over years. As we touch on some of this subject, there are some important caveats and observations that we should bear in mind.
THD is just a measurement result from which we infer nonlinearity under well-controlled laboratory conditions with sinewaves. We don't necessarily "hear" THD, especially THD at high frequencies. THD is just one of many symptoms, but a pretty good one at that. We also need to bear in mind what the "instruments" DON'T measure. That can be lots of things. Your example of THD resulting from clipping sounding different from THD resulting from crossover distortion is an example, although even it is slightly explained by a careful analysis of THD as a function of power level. This relates somewhat to the "first watt" issue.
THD does not come from a particular kind of nonlinearity. Under most conditions a given nonlinearity (the same nonlinearity) will also create the other kinds of measurable distortion, like SMPTE IM and CCIF IM. With music, IM distortion is more likely what we are hearing. IM products tend to be very dissonant from the source signal, so they tend to be less masked by the source. High-frequency THD is important because it often correlates with high frequency IM distortion, like 19 kHz + 20 kHz CCIF IM. The IM products of interest from HF IM are in-band at lower frequencies, and probably more likely to be heard because they are dissonant and maybe not as well masked. I believe that if you hear a dirty cymbal from distortion, it is likely the IM products that are creating the dirt.
TIM was a type of distortion measurement that was steeped in controversy many years ago. It also was a measurement metric reflective of HF IM. It also was created by the same nonlinearities that create THD and CCIF IM. People heard it. Crappy amplifiers had it. It was no different than slewing-induced distortion. The controversy arose because it was wrongly claimed that large amounts of negative feedback or low open-loop bandwidth caused it. That was easily disproved by properly designing a really good amplifier that had high feedback and low open-loop bandwidth that measured extremely low TIM.
Low THD does not necessarily lead to the best sound, but sometimes the steps taken, design diligence and attention to detail exercised to get low THD make for better sound anyway. Similarly, moderate THD doesn't necessarily mean a bad-sounding amplifier. Valve amplifiers and solid state amplifiers with soft clipping circuits come to mind.
Another thing I like to think is that amplifiers sound different because they misbehave differently. Conventional lab measurements often do not capture misbehavior that may occur in the real world.
Cheers,
Bob
Problem is: headphone amplifiers. They sound different from each other, the ones that sound very good don't always measure very good (and vice versa). Worse yet, they are never driven into clipping, which is the standard misbehavior everyone talks about. A clipped HPA would create earsplitting volume in the headphones; so HPA users turn down the knob and listen at reasonable volume, far far away from the onset of clipping.
If you are driving your amplifier into clipping, you need a bigger amp, as simple as that.
Not exactly the same, but a parallel idea from 1955.
Philips, Rejection Factor Of Difference Amplifiers
Cascodes bootstrapped with a conceptual battery or a gas "Zener".
Cost HAD to be an objection to these schemes; more likely poking at limits than a practical application.
(I now suspect that "M18" is 0.18 Meg or 180K; odd to US eyes but probably common around Philips.)
Attachments
Agreed, its unlikely the headphone amps are clipping unless they are really crappy. However, the load impedance range of headphones themselves is all over the place, ranging from as low as 16 ohms to perhaps 600 ohms. Some may not be able to supply the current for low-Z phones. Others may have poor damping factor (ill-defined for headphones) and may interact with the frequency-dependent impedance of lower-impedance headphones.
Cheers,
Bob
The most comprehensive work I've ever seen on Headphone characterization is this list:
Headphone List - Google Sheets
Don't know who the author is but the link was given to me around 3 years ago, so it may not include the latest models. The list includes circa 500 headphone brands and models, their impedance and the required power/current/voltage to achieve 90/100/105/110/115 dB SPL, its the product of a massive ammount of work and research and its been a real eye opener for me when designing headphone amps.
Headphone List - Google Sheets
Don't know who the author is but the link was given to me around 3 years ago, so it may not include the latest models. The list includes circa 500 headphone brands and models, their impedance and the required power/current/voltage to achieve 90/100/105/110/115 dB SPL, its the product of a massive ammount of work and research and its been a real eye opener for me when designing headphone amps.
Last edited:
I'm designing an amplifier with distortion less than 0.01% at 20kHz. Many people said they can hear the different than other amplifier.
You should learn how to listen. Listening in critical mode is difficult and need practice for long time. May be more difficult than understand how circuit work
People preference is not always less distortion is the better. But I agree with you, it is cheaper to add "sound effect" in pre-amplifier.
I don't agree with this. If this were true, how does one explain the popularity of the First Watt line of amps. None of them measure particularly well against modern HiFi amps. Yet they are clearly well regarded.
Also look at Sakis's threads (user east electronics) on Rod Elliott's P3a. He has repeatedly said that his customers will favor his loaner P3a over the low-distortion amp that he is repairing for them. The P3a is another amp that doesn't measure well comparably.
This means that people's tastes vary. Some prefer ultra-low distortion. Others like a splash of low order even or odd harmonics. Some want their "added sound" in the line stages... others prefer it in the power amp.
And finally, this is a hobby. It should be enjoyable. So ignore what everyone tells you and do what makes you personally happy.
+1000
An audio amplifier designer should able make audio amplifier with different philosophy and try to understand why they are sound different.
Mark is this correct? It seems to indicate that 100% of the headphones need 0V or 0mA.
Jan
+1
I would expect the graph rising fast and then showing that with about 1.5V respectively 30mA about 80% of the Headphones are covered.
3V and 50mA are needed for 95%.
Yes.
100% of the headphones need 0.0V or more, to achieve 110dB SPL
90% of the headphones need 0.4V or more, to achieve 110dB SPL ... and 10% of them don't
80% of the headphones need 0.5V or more, to achieve 110dB SPL ... and 20% of them don't
70% of the headphones need 0.7V or more, to achieve 110dB SPL ... and 30% of them don't
Similarly,
100% of the headphones need 0mA or more, to achieve 110dB SPL
90% of the headphones need 5mA or more, to achieve 110dB SPL
80% of the headphones need 10mA or more, to achieve 110dB SPL
They are plots of the cumulative distribution.