Head-Fi.org › Forums › Equipment Forums › Sound Science › Testing audiophile claims and myths
New Posts  All Forums:Forum Nav:

Testing audiophile claims and myths - Page 199

post #2971 of 3674
Quote:
Originally Posted by Mezzo View Post
 

Sorry, I should have been more clear. I am talking about speakers, not headphones. But I think I understand what you are saying and the 1 kHz test signal makes sense.

 

In any case, you need to measure the voltage on the speakers or headphones. This may require DIY cables or connectors, if you do not have a connector (like the splitter recommended for headphones) that makes it possible to access the terminals with a multimeter.

 

Quote:

Originally Posted by Mezzo View Post

 

So if one measures the left and right channels and there is imbalance then basically there is no way to level-match properly? So the test is basically worthless?

 

Not necessarily if at least one of your speaker amplifiers or receivers has a balance control. Then you can adjust it to match the imbalance of the other device. Another solution that may work is to use different volume settings where the difference is smaller, and change the source volume as well to compensate. With devices that do not use potentiometers, there might not be much imbalance to begin with. It also may not be necessary to match the channel balance as accurately as the overall levels (which can be calculated as the square root of the sum of the squared left and right voltages), but I am not sure what the maximum acceptable amount of error is.

post #2972 of 3674

If it's difficult to go to all that trouble, just level match by ear and take your time about it. Then set up a blind test with a friend. A "close as dammit" test is better than no test at all. Everyone should do comparison tests of their equipment to find out what the differences are. Not many people do that, and that is a big reason audiophiles waste money on things that don't make a lick of difference.

post #2973 of 3674
Quote:
Originally Posted by bigshot View Post

If it's difficult to go to all that trouble, just level match by ear and take your time about it. Then set up a blind test with a friend. A "close as dammit" test is better than no test at all. Everyone should do comparison tests of their equipment to find out what the differences are. Not many people do that, and that is a big reason audiophiles waste money on things that don't make a lick of difference.

Did this with my dac1 USB vs odac / O2. They sound identical. Both fantastic, transparent amp and dacs. Just shows transparent sound really is measurable . It's nothing to do with who made the components and how expensive they are. They either pass the objective criteria for transparency or they don't. If they do they sound identical. It's good to know that amps and dacs have reached this point and they can never sound any better . My wallet is very happy!
post #2974 of 3674
Now we just need the damn sound engineers to make great recordings and so their job properly.
post #2975 of 3674

they just do what the guy with the money asks for - even if they would rather not

 

Quote:
Originally Posted by jcx View Post

Perceived loudness wins over musical quality

at a recent talk at a mastering studio we were told that several candidate mixes with differing levels of compression were provided for their projects and the most heavily compressed was always selected by the clients - even after it was explained what the compression was doing to the music and that the world renown mix engineer recommended the lesser compressed mix


I'm also surprised there still aren't alterative binaural mixes of new releases optimized for headphone listening given the ubiquity of iem/bud/headphone portable audio today
post #2976 of 3674
Quote:
Originally Posted by bigshot View Post
 

If it's difficult to go to all that trouble, just level match by ear and take your time about it. Then set up a blind test with a friend. A "close as dammit" test is better than no test at all. Everyone should do comparison tests of their equipment to find out what the differences are. Not many people do that, and that is a big reason audiophiles waste money on things that don't make a lick of difference.

As has been mentioned to you before, the problem with level matching by ear is that two sources that differ slightly in level can be perceived as having the same exact level, but with the slightly louder one having a better quality. This can be discerned in a double blind test too. If you want to properly compare two components, level matching by ear is not good enough.

post #2977 of 3674
Quote:
Originally Posted by cjl View Post
 

As has been mentioned to you before, the problem with level matching by ear is that two sources that differ slightly in level can be perceived as having the same exact level, but with the slightly louder one having a better quality. This can be discerned in a double blind test too. If you want to properly compare two components, level matching by ear is not good enough.


+1.  It is better than nothing, but not really good enough.  Besides using speakers it is simply too simple to do with a voltmeter.

 

And how closely should it be matched?  Well usually .2 db is considered audible.  Making things simple a 2% voltage difference is just under .2 db (.172 in fact).  So match voltage to within 2% or a touch less. 

post #2978 of 3674
Quote:
Originally Posted by cjl View Post
 

As has been mentioned to you before, the problem with level matching by ear is that two sources that differ slightly in level can be perceived as having the same exact level, but with the slightly louder one having a better quality. This can be discerned in a double blind test too. If you want to properly compare two components, level matching by ear is not good enough.

Unfortunately true.

 

Why unfortunately ? Volume control, be it digital or analog, gets troublesome below 0.5 dB "steps". And , again unfortunately, requires quite a lot of $. Most of the time

one has to work without a reference tone for level ( like 1 kHz sine wave at specified level ) - and even if this reference tone is available, matching within 1% or its equivalent of 0.1 dB requires an oscilloscope ( 1% requires periodically calibrated/certified scope - or else scope inaccuracies will lead to incorrect matching ) or a good RMS voltage meter.

 

Nothing impossible, but pricey - to the point that even rather big studios do not use level matching to within 0.1 dB or less. 

 

Level matching by ear is sensitive down to approx 0.2 dB - which is also the amount of error in level caused by various conversions among different digital forms of recording. On one set of digital hardware it is possiblle to make/normalize a  0 dB peak recording - that same file will produce clipping by approx 0.2 dB with another software and/or hardware. That 0.2 dB is also audible - but only if you work with these low level differences for extended periods of time, like during editing.

 

So the trouble is that both hearing and measuring becomes iffy below 0.2 dB difference - possible, but costly. 

post #2979 of 3674
Quote:
Originally Posted by analogsurviver View Post
 

Unfortunately true.

 

Why unfortunately ? Volume control, be it digital or analog, gets troublesome below 0.5 dB "steps". And , again unfortunately, requires quite a lot of $. Most of the time

one has to work without a reference tone for level ( like 1 kHz sine wave at specified level ) - and even if this reference tone is available, matching within 1% or its equivalent of 0.1 dB requires an oscilloscope ( 1% requires periodically calibrated/certified scope - or else scope inaccuracies will lead to incorrect matching ) or a good RMS voltage meter.

 

Nothing impossible, but pricey - to the point that even rather big studios do not use level matching to within 0.1 dB or less. 

 

Level matching by ear is sensitive down to approx 0.2 dB - which is also the amount of error in level caused by various conversions among different digital forms of recording. On one set of digital hardware it is possiblle to make/normalize a  0 dB peak recording - that same file will produce clipping by approx 0.2 dB with another software and/or hardware. That 0.2 dB is also audible - but only if you work with these low level differences for extended periods of time, like during editing.

 

So the trouble is that both hearing and measuring becomes iffy below 0.2 dB difference - possible, but costly. 

Uhhhhh.....no.  Not really.

 

If it works out that .5 db is close as you can get, then get that close.  Depends on the particulars of what you are comparing and how of course.  But with analog volume controls it is touchy, but possible to get very close.  Much digital gear is also cable of .1 db adjustments these days.  But even take a hypothetical like comparing two things and your digital volume control is only good in .5 db steps.  The two items aren't exactly .5 db different (if they were you could match them).  Being less than .5 db different or more than .5 db different means you will have the chance to match volume between the two at something for sure less than .5 db.  On average you might expect .25 db though it could be anywhere between 0 and .5 db.  Most test files can be done with at least 1 db of headroom so the little processing artifacts of DSP causing a level to clip aren't an issue.  So I am not sure why you would push the idea such matching will usually cost lots of money.  If worse comes to worse there are free digital sound editors that can adjust volume of digital files in a thousandth of a db or less with no artifacts worth worrying about.  Nor will you need more than a reliable AC voltmeter.  Plenty of them under $50 are good enough for matching a 1 khz tone.   Sure it may be 5% off in the reading, but they are usually a consistent 5% and consistency is the main thing for this sort of matching.

 

So very possible and inexpensive. 

post #2980 of 3674
Quote:
Originally Posted by cjl View Post
 

As has been mentioned to you before, the problem with level matching by ear is that two sources that differ slightly in level can be perceived as having the same exact level, but with the slightly louder one having a better quality. This can be discerned in a double blind test too. If you want to properly compare two components, level matching by ear is not good enough.

 

It doesn't apply if you're looking for very subtle differences, but if you are looking for clear differences, it's good enough. I don't sweat the little stuff. I'm looking for significant improvements. There are better uses for my time than worrying about things that I can barely hear in a rough comparison.

post #2981 of 3674
Quote:
Originally Posted by bigshot View Post

It doesn't apply if you're looking for very subtle differences, but if you are looking for clear differences, it's good enough. I don't sweat the little stuff. I'm looking for significant improvements. There are better uses for my time than worrying about things that I can barely hear in a rough comparison.

It's great to read rational posts like this.
post #2982 of 3674
Quote:
Originally Posted by esldude View Post
 

Uhhhhh.....no.  Not really.

 

If it works out that .5 db is close as you can get, then get that close.  Depends on the particulars of what you are comparing and how of course.  But with analog volume controls it is touchy, but possible to get very close.  Much digital gear is also cable of .1 db adjustments these days.  But even take a hypothetical like comparing two things and your digital volume control is only good in .5 db steps.  The two items aren't exactly .5 db different (if they were you could match them).  Being less than .5 db different or more than .5 db different means you will have the chance to match volume between the two at something for sure less than .5 db.  On average you might expect .25 db though it could be anywhere between 0 and .5 db.  Most test files can be done with at least 1 db of headroom so the little processing artifacts of DSP causing a level to clip aren't an issue.  So I am not sure why you would push the idea such matching will usually cost lots of money.  If worse comes to worse there are free digital sound editors that can adjust volume of digital files in a thousandth of a db or less with no artifacts worth worrying about.  Nor will you need more than a reliable AC voltmeter.  Plenty of them under $50 are good enough for matching a 1 khz tone.   Sure it may be 5% off in the reading, but they are usually a consistent 5% and consistency is the main thing for this sort of matching.

 

So very possible and inexpensive. 

I do not agree with that inexpensive part. Simply look at the prices for analog potentiometers that are specified at anything approaching 0.2 dB across at least 60 dB

range. I agree that reliable AC voltmeter can be had for 50$ and that its absolute accuracy is not required for consistent result of matching we need.

 

Problem is that for the most part I DO NOT WANT digital volume control - as it means DSP and that means PCM. Even if it is accurate to within 0.000000.....1 dB.

Not on analog sources and not on DSD. Which brings us back to quality analog potentiometer$ / attenuator$. And THESE can be source of trouble - by the time they are "good enough", that means money. 

 

The first stupid thing any analog volume control device must avoid like a pleague is - microphonics. Unfortunately anything but trivial and taken for granted to be free from it.

 

It is fair to say that multichannel ( > 2 ) equipment can not be built with analog potentiometers ( differences become practically unmenageable, with the po$$ible exception of attenuator$$$ ) - but for all practical purposes, multichannel is already in PCM and yet another PCM stage will do no additional audible harm in such a system.

post #2983 of 3674
Quote:
Originally Posted by analogsurviver View Post
 

I do not agree with that inexpensive part. Simply look at the prices for analog potentiometers that are specified at anything approaching 0.2 dB across at least 60 dB

range. I agree that reliable AC voltmeter can be had for 50$ and that its absolute accuracy is not required for consistent result of matching we need.

 

Problem is that for the most part I DO NOT WANT digital volume control - as it means DSP and that means PCM. Even if it is accurate to within 0.000000.....1 dB.

Not on analog sources and not on DSD. Which brings us back to quality analog potentiometer$ / attenuator$. And THESE can be source of trouble - by the time they are "good enough", that means money. 

 

The first stupid thing any analog volume control device must avoid like a pleague is - microphonics. Unfortunately anything but trivial and taken for granted to be free from it.

 

It is fair to say that multichannel ( > 2 ) equipment can not be built with analog potentiometers ( differences become practically unmenageable, with the po$$ible exception of attenuator$$$ ) - but for all practical purposes, multichannel is already in PCM and yet another PCM stage will do no additional audible harm in such a system.


Well I can't do anything about an irrational fear of DSP or PCM processing.  I would imagine that is an issue for a small minority of people.  Most sources these days are going to be digital.  By your own description digital volume bypasses issues of consistency, microphonics and fineness of control. 

 

If you wanted something better in analog control you can go with switched resistors.  It is a problem created by eschewing digital processes. 


Edited by esldude - 8/15/14 at 2:29pm
post #2984 of 3674
Quote:
Originally Posted by esldude View Post
 


Well I can't do anything about an irrational fear of DSP or PCM processing.  I would imagine that is an issue for a small minority of people.  Most sources these days are going to be digital. 

 

If you wanted something better in analog control you can go with switched resistors.  It is a problem created by eschewing digital processes. 

Unfortunately, it is not an irrational fear.  Otherwise, all analog recorders and tapes to feed them would have long ago become relatives of the Dodo. Yet at least rell to reel tapes made comeback - not en masse as in the past, but there was sufficient demand to warrant current production of tapes. Slowly digital is approaching the overall quality of analog tape - and it eventually should become able of eclipsing it. Yet analog recordings  to sound as original would require perfect (digital) recorder - no such thing yet; almost, but almost never catches a rabbit.

 

Switched resistors are unfortunately quite prone to microphonics - I had more than one rude awakening due to this in the past. What good is level perfectly matched if volume control "sings" along with the real tune ?

 

There is a reason why some manufacturers insist on carbon potentiometers for audio - although plastic on paper looks better and is guaranteed to last much longer. Even if it means replacing the carbon potentiometer(s) every now an then.

 

The reason is simple - sound.

post #2985 of 3674

None of that matters. Listen with your ears with your brain turned on. Those who are too crazy to do that can feel free to disregard this advice.

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Testing audiophile claims and myths