Sound Quality Vs. Measurements

Status
Not open for further replies.
So what is the usefulness and what would be the most appropriate measurements?

Is this useful?
 

Attachments

  • Best stereo amplifier tested.png
    Best stereo amplifier tested.png
    53.3 KB · Views: 129
Last edited:
Member
Joined 2009
Paid Member
Indra maybe you are a bit harsh sounding: these kind of (test) problem/approach is difficult to approach even for PhD or students in the field.

For example a 0,1db difference in output volume will make you prefer the louder DUT each time...

That said i agree with you that the acoustic output have to be taken into account: in the chain the loudspeakers are very likely the worst place in the chain regarding artefact ( distortion, whatever,...) so it should the place to emphasis in my view ( electronic circuits which are 'blameless' can easily be found for converters, amplifiers, preamp,...).
But even then some loudspeakers will interact 'better' with a tube amp or a transistor amp or a classD,... and i don't even talk about multiamp/active or dsp driven filters.

This is'nt easy as there is interactions at play and without a clear view of many 'details' it could be misleading.
 
Administrator
Joined 2004
Paid Member
Hi Mark,
If you are well versed in engineering, then it does make common sense. To others it may appear as a complete mystery. The amount of experience and training most older and well versed members have here with setting up experiments that make sense is beyond understanding for a layperson. There are just so many small things that matter and we have learned this over many years. I'm not saying this group is any more intelligent than other folks, but the amount of experience is greater than one might imagine.

Andrea and all,
Interpreting measurements (not single numbers) takes a great deal of practice. I think we are all still learning, but the groundwork is there. It is easier to see what and why a great technical amplifier sounds good. Harder to look at a compromised amplifier and figure out why it sounds good sometimes. Tube products are the greatest challenge here. But a SET amplifier, expensive or not, is extremely coloured by any metric. They do not do well with complicated music, but some people seem to like quartets and vocals with them. They are in the minority by far. Compare a good amplifier (p-p tube or solid state) instead of an unknown LM3886 and you may well have a different outcome unless peer pressure is very active. (People don't like to appear stupid or unknowledgeable). Use a good speaker system of course, something along the lines of a current or near current Klipsch system where efficiency is close to 98 dB/watt or better. That levels the playing field and does not attenuate the highs where many nasties live. SET amps can't hide behind a speaker that doesn't cover the highs properly. So that is a level playing field. Now, play some piano and full orchestral music. You will find the better (non-SET amp) sounds good with everything but poorly recorded material. It will ruthlessly reveal recording problems and that is after all what you want in a system.

So that would make a more valid test between amplifiers. Don't load the deck with a known bad sounding solid state amplifier and draw conclusions with that. Same for the speakers, use something that is flat and good sounding, but efficient so the SET can perform to the best of its ability.

Done this test many times. The SET amplifier always loses badly except to fanatics married to the idea, but not actually listening. Want a realistic presentation? Use a technically accurate system. If the recording is bad, it should sound bad. A good recording should be a very pleasurable experience. I have run this test several times with various "good amplifiers", both solid state and tube. The SET amplifiers are always too coloured to render the sound accurately and they are not very pleasing with complicated music. You will find SET amplifier system owners play very simple music and not typically very loud.

-Chris
 
"For example a 0,1db difference in output volume will make you prefer the louder DUT each time..."

Probably so if they are otherwise identical. If they vary more in other ways besides volume, the quieter one may be preferred.

Personally, I suggest that the volume of each device be adjusted up and down by the test subject. Sometimes characteristics of the sound change depending on the DUT's operating point. Its also one way to reduce the perceptual effects of small volume level differences.

Another way to measure is the have the test subject adjust the volume level of one device compared to a reference device at a fixed volume. The test subject adjusts the volume level until the devices are equally preferred. The volume adjustment can be taken as a figure of merit for the perceptual preference difference between devices.
 
Last edited:
Member
Joined 2009
Paid Member
"He missed the 'body' the distortion ...."
Is this wrong?
It does not matter.
Would you impose your vision on other people?
Do you want to make the choice for them?

No it is not wrong.
Does it matter? To me yes as it was my job and i'm used to a kind of presentation which try to not colour the message reproduced.

I make a point not to impose my choice to anyone but as i have a science driven approach and as i'm like everyone looking for some kind of trustfull parameters to make choice i prefer to have objective point to make subjective choices.

You seems to be 'subjectively' driven about all this, i am not. Tools are tools ( remember it was my tools at work) and either they are ok to me either they are not. I don't have feelings involved into this.

Tbh i've got nothing against tubes and when i track i usually try my own tube driven mic first ( my own design, as most of my tube gear), i've got tube preamp, varimu compressors and 'tube'eq,... which i love! But for playback my preferences are for something much more transparent. It's only preference thing.

I want objective criteria to make choice, not feeling driven.
 
Last edited:
Member
Joined 2009
Paid Member
Markw4,
Not my experience at all when i perform mastering tasks: clients always choose the louder one even if it is clipped!
We are done like that, our brain beg/crave for high rms output level ( the reason the loudness war started comes from study/observation during 50's/60's that the higher volume printed vinyls in jukebox was the most played, then radio station applied this to commercials first then overall, then in the 90's the war begun with the apparition of multiband dynamic treatments and digital tools ( vinyl didn't allow such high levels so kind of protected from excess at the time...)).
 
Last edited:
There are people who prefer loudness above all else. Not me. I can't stand clipped.

However, my kids like the sound of the clipped-sound popular music recordings they grew up with; they liked it more even with its volume lowered compared to an unclipped reference. Also, a lot project studio recordings contain so much digititus from cheap converters and plugins that a little clipping can make them sound better. It can serve as a gluing effect.

Musicians, producers, and record companies often want to compete in the loudness wars, but there are some who prefer clean recordings with space for room sound to be part of the listening experience.
 
Last edited:
Hi Chris,

Ok, then in the meantime let's see if I used poorly recorded material:

Mary Black - Columbus (16 bit/44.1kHz)
Eric Bibb - I Want Jesus to Walk with Me (16 bit/44.1kHz)
Enya - Fairy Tale (16 bit/44.1kHz)
The Eric Westberg Vocal Ensemble - Sanctus (16 bit/44.1kHz)
G.Karr - Gavotte in D Major (16 bit/44.1kHz)
Archie Shepp Quartet - More than you know (16 bit/44.1kHz)
Stefan Milenkovich - Violin Concerto no.2 in B Minor (16 bit/44.1kHz)
Carol Kidd - When I dream (16 bit/44.1kHz)
Daniel Baremboim Beethoven - Piano Sonata No.23 Appassionata (16 bit/44.1kHz)
Trondheimsolistene - Britten Simple Symphony Op 4 (24 bit/192kHz)
Marianne Beate Kielland & Sergej Osadchuk - From Let Us Garlands Bring, Op.18: Come Away, Death (24 bit/192kHz)
Christian Grøvlen - Chromatic Fantasia and Fugue in D minor, BWV 903: Fantasia (24 bit/176.4kHz)
HOFF ENSEMBLE - Blågutten (24 bit/192kHz)
Hoff Ensemble: Jan Gunnar Hoff, Audun Kleive & Anders Jormin - Innocence (24 bit/176.4kHz)
Nidarosdomens jentekor & TrondheimSolistene - Arnesen: MAGNIFICAT 4. Et misericordia (24 bit/192kHz)
The Staff Band of the Norwegian Armed Forces / Ole Kristian Ruud - Bozza: Children’s Overture (24 bit/192kHz)
Engegård Quartet - Haydn String Quartet In D, Op. 76, No. 5 - Finale - Presto (24 bit/192kHz)
Tone Wik & Barokkanerne - VIVALDI Cantata Rv 679: Che Giova Il Sospirar, Povero Core - Aria: Cupido, Tu Vedi (24 bit/192kHz)
 
No it is not wrong.
Does it matter? To me yes as it was my job and i'm used to a kind of presentation which try to not colour the message reproduced.

I make a point not to impose my choice to anyone but as i have a science driven approach and as i'm like everyone looking for some kind of trustfull parameters to make choice i prefer to have objective point to make subjective choices.

You seems to be 'subjectively' driven about all this, i am not. Tools are tools ( remember it was my tools at work) and either they are ok to me either they are not. I don't have feelings involved into this.

Tbh i've got nothing against tubes and when i track i usually try my own tube driven mic first ( my own design, as most of my tube gear), i've got tube preamp, varimu compressors and 'tube'eq,... which i love! But for playback my preferences are for something much more transparent. It's only preference thing.

I want objective criteria to make choice, not feeling driven.

I believe a piano teacher knows how a piano should sound.
Don't you believe?

Do you think a select few who write on an audio forum are more capable?
 
Member
Joined 2009
Paid Member
Mark,
All you described was my day to day reality.
I'm with you about the excess ( clipping and close to 6db dynamic range...) and can't stand it on everything i listen ( for some genre/style you can't do without high rms level, good or bad).
I'm to the point to only buy vinyl which i digitalize to have 'reasonnable' rms level... crazy situation! ( but i will stop about it as it is too much off topic and keep on bring it back...).

Andrea,
I'm not aggressive to you so please reformulate your wording as it seems to be to me.

I explained my view on all this, if you think what 'you like' ( what please you) is the reality that is fine to you. It is not my view or experience so please respect it as i respect your own, but please don't try to put words into my mouth or try to implies things on me.

If you want to convince me that is fine but please bring back some arguments, not anecdotal evidence or pure subjectivist views.
 
Last edited:
So what is the usefulness and what would be the most appropriate measurements?
Good question. I had been on the quest for years just like you do, and the end of the road is still nowhere in sight yet. Just like you, most of us here comes to learn. There is simply overwhelmingly too many qualities a reproduced sound have. These days we kept reading that sound at a level no human can hear is within capability of measurement. But it still needs a lot of work to correlate how a normal listening level sound to perceived qualities. I really hate comments implying no more research would bring any benefit. Measurable means a possibility to understand but does not imply complete understanding.
Indra maybe you are a bit harsh sounding: these kind of (test) problem/approach is difficult to approach even for PhD or students in the field...
Apologies, the harsh part was meant to be directed only to the comments benefitting shills and snake oil peddlers, regardless if those who made the comments were aware of this bit or not. And agreed, if it were easy then we would have had the complete picture a long time ago.
 
Mark,
All you described was my day to day reality.
I'm with you about the excess ( clipping and close to 6db dynamic range...) and can't stand it on everything i listen ( for some genre/style you can't do without high rms level, good or bad).
I'm to the point to only buy vinyl which i digitalize to have 'reasonnable' rms level... crazy situation! ( but i will stop about it as it is too much off topic and keep on bring it back...).

Andrea,
I'm not aggressive to you so please reformulate your wording as it seems to be to me.

I explained my view on all this, if you think what 'you like' ( what please you) is the reality that is fine to you. It is not my view so please respect it as i respect your own, but please don't try to put words into my mouth or try to implies things on me.

If you want to convince me that is fine but please bring back some arguments, not anecdotal evidence or pure subjectivist views.

I wanted to be anything but aggressive, I'm sorry if you sensed it.

I meant I believe that a piano teacher is sufficiently prepared to evaluate its reproduction.
How could I tell her she was wrong because the best reproduction was that of the other amplifier?
 
But it still needs a lot of work to correlate how a normal listening level sound to perceived qualities.

I totally agree on this.

And I'm also here to learn.
In fact, I have given my full availability to carry out all the measurements deemed necessary.

Just to clarify, the tube amp was the Music Angel XDSE 845, not a $$$ amp, it was bought for EUR 375 in 2015.
The LM3886 amp was the Mauro Penasa's My_Ref Evolution.
 

Attachments

  • Amplifier.jpg
    Amplifier.jpg
    359.8 KB · Views: 118
Last edited:
Member
Joined 2009
Paid Member
Ok fine. We are all here to learn so we have many things in common.
The point is not who is right or wrong to me: (edit: how i see) the reality is once a message have been modified ( by an amp, a mic, a loudspeaker,...) it is not the reality anymore.

Is it pleasing or not is a different matter.

The fact that a musician prefer something isn't to me a proove of anything: i can bring you guitarist which saturate/clip their amplifier because this is the sound they like or what is needed for their style.

I don't list the thing i listen to because you would probably think this is just ... noise ( which in some case really is!).

It's all about preference and subjectivism.

Measurements quantify things and doesn't impose preferences.

I don't mind to use a mic which color the source if it is what i think is needed from an aesthetic pov.

But it is not 'reality' of sound source anymore. I or the artist may prefer it but this is not 'true'.

But to be sure i'm not mistaken the ( reproduction) chain have to be clean enough and only by measurements you can have an objective view of it ( if you are examinating the correct parameters of course).

Indra no need to apologize. I understand and share your point about snake oil but... i can't help myself and bought some russian teflon caps for re evaluation in some of my tube circuits... but please hush hush i don't want to be busted by my objectivists friends! ;)
 
Last edited:
Andrea_mori said:
If this is accepted as the fact I go beyond, otherwise I stop here.

It is obvious Andrea.

More than 20 years ago I used tube amps (and Lowther etc.). Any of my friends or family (call them newbies) that heard any of my tube amps (especially those not involved with audio) will always be amazed by tube amp and prefer that over my other amps. It is a guaranteed phenomenon.

But if I were to ask experienced audiophile, I will not bet on it...

There is something that is called "acquired taste". Only after we are exposed to something better we may change our taste or preference. You can't expect newbies to be exposed to this kind of experience.
 
Last edited:
Ok fine. We are all here to learn so we have many things in common.
The point is not who is right or wrong to me: (edit: how i see) the reality is once a message have been modified ( by an amp, a mic, a loudspeaker,...) it is not the reality anymore.

Is it pleasing or not is a different matter.

The fact that a musician prefer something isn't to me a proove of anything: i can bring you guitarist which saturate/clip their amplifier because this is the sound they like or what is needed for their style.

I don't list the thing i listen to because you would probably think this is just ... noise ( which in some case really is!).

It's all about preference and subjectivism.

Measurements quantify things and doesn't impose preferences.

I don't mind to use a mic which color the source if it is what i think is needed from an aesthetic pov.

But it is not 'reality' of sound source anymore. I or the artist may prefer it but this is not 'true'.

But to be sure i'm not mistaken the ( reproduction) chain have to be clean enough and only by measurements you can have an objective view of it ( if you are examinating the correct parameters of course).

Indra no need to apologize. I understand and share your point about snake oil but... i can't help myself and bought some russian teflon caps for re evaluation in some of my tube circuits... but please hush hush i don't want to be busted by my objectivists friends! ;)

It was not the piano teacher only, it was 4/4 people (5 including me).

But OK, my mind is open.
Now I would like to know what measurements I have to perform, but if no one is interested I stop here.
 
It is obvious Andrea.

More than 20 years ago I used tube amps (and Lowther etc.). Any of my friends or family (call them newbies) that heard any of my tube amps (especially those not involved with audio) will always be amazed by tube amp and prefer that over my other amps. It is a guaranteed phenomenon.

But if I were to ask experienced audiophile, I will not bet on it...

There is something that is called "acquired taste". Only after we are exposed to something better we may change our taste or preference. You can't expect newbies to be exposed to this kind of experience.

I don't think audio is something for a select few.

And I didn't even think I was a newbie since I had the same preference of the other people.
But evidently since '80s I still haven't understood anything.
And to think that the first amps I designed were just solid state and used feedback.

I'm really a wrong guy, I did the reverse.
 
Administrator
Joined 2004
Paid Member
Hi Andrea,
Not my musical tastes, but I can't help but notice it is simple music. So how it is recorded I don't know. I'll assume they are good recordings.

You probably don't listen to more complicated music because that or similar systems devolve into a nasty mess. You should try an orchestra playing spirited music. I tend to listen to everything but opera, country or rap. So for me, something accurate and clean is pretty critical. The harmonic structure of a piece goes all to heck with SET type equipment.

The SINAD readings you posted aren't useful for determining the quality of anything much. It is commonly used to gauge the intelligibility of speech channels in communications equipment. I don't understand why anyone would try to use it as a metric for a music reproduction system.

-Chris
 
Member
Joined 2009
Paid Member
Tbh i don't know what you want to measure: a sound quality index that's it?

For loudspeakers i can tell that 'spinorama' is close to corelate to a quality index ( i'm more into loudspeakers than electronics for the last ten years).

Spinorama spreadsheet | AVForums

That said this is statistical and your own preference may vary.


For an electronic circuit pov i think Anatech or Syn08 could tell you more about it that i could.

Anyway my own experience with my own tube circuits tell me that distortion ( it's harmonic profile) is of importance. Signal to noise ratio ( or said differently DDR in the Allen Wright's own language), the capability of the psu to not sag under load ( i like parallel regulated psu), the dynamic behavior in general ( transients integrity)...
But this is limited and i don't have the gear to test all this so i 'trust' the designers or simulations i do of circuits then real life test ...


I think the first book Abraxalito linked some page back could help you evaluate all this within your own concern or interest. Then you test and see by yourself if it is meaningfull or not.

Put theorical hypothesis first then verify it in real life. If it doesn't bring you something then re evaluate the hypothesis, rince and repeat... until you reach your own conclusion.

I think of an other thing you could do: take a signal, a DAW and a plug like Ozone then try things with distortion you can add by bands, eq, dynamics,... it'll give you another ( interesting) insight.

Edit: The first link could be of interest:

https://www.diyaudio.com/forums/the...-dr-earl-geddes-gedlee-audio.html#post6636818
 
Last edited:
Status
Not open for further replies.