When this album is played with an average SPL of 90dB you will need to reach 106dB SPL.
10dB more = twice as loud peaks = 10x more power reserve is needed while listening with the same average SPL.


indeed but what IF you use that car to drive your grandmother but also use it to drag a heavy trailer now and then ?
What if you live far away from a hospital and need to get there really fast ?

if we play the math game:

• 90db with 106db peaks is incredibly loud.
• most headphones require little power (or not as much power as people think) to reach 106db. At 106db, there's going to be more distortion from the headphone than from most amps.

Hence we are not pulling any tractor trailers (although the HE-6 could be considered one) or speeding to the hospital. It's really a matter of a hearing safety issue.

That is if we just do the math thing.

The hearing is quite insensitive to bass (examine the phon curves) which explains why a higher voltage impacts bass relatively the most.

Quote from: Solderdude on March 15, 2014, 07:31:07 AM

When this album is played with an average SPL of 90dB you will need to reach 106dB SPL.
10dB more = twice as loud peaks = 10x more power reserve is needed while listening with the same average SPL.


indeed but what IF you use that car to drive your grandmother but also use it to drag a heavy trailer now and then ?
What if you live far away from a hospital and need to get there really fast ?

if we play the math game:

• 90db with 106db peaks is incredibly loud.
• most headphones require little power (or not as much power as people think) to reach 106db. At 106db, there's going to be more distortion from the headphone than from most amps.

Hence we are not pulling any tractor trailers (although the HE-6 could be considered one) or speeding to the hospital. It's really a matter of a hearing safety issue.

That is if we just do the math thing.

To illustrate Marv's point some math:

Here is the not too scientific approach.
Let's assume one listens to high quality recordings with a DR around 15.
Let's assume the listener is average (consider myself average) in sensitivity to sound.
Using a cheap (not calibrated) SPL meter on slow A-weighted setting.
Used Norah Jones - sunrise (live performance), Patricia Barber-touch of trash , Diana Krall-boulevard of broken dreams
used a -1dB 1kHz sine to find the peak that could (and does) occur.
a capable amp was used and HD650

Played on a soft level.
By that I mean one can listen to that comfortably all evening.
average levels were measured between 60-70dB and peak levels of 72(NJ) and 87(PB) were measured.

Played on a 'normal' level which is loud enough to play for at least an hour but would leave you with an headache if used the whole evening.
Measured average values between 70 and 80dB (depending on content)
Peak values were measured 87dB(NJ) and 98dB(PB)

Played on a loud level which would be used to scrutinise a recording or to hear 'everything'. This was loud but NOT uncomfortable and can easily be handled for 1 or even a few songs.
Measured average values between 80 and 90dB (depending on content)
Peak values were measured 97dB(NJ) and 106dB(PB)

Played on a VERY loud level, a level that can ONLY be sustained VERY shortly and is uncomfortable to say the least.
Measured average values between 90 and 105dB (depending on content)
Peak value was measured 118dB(PB)

Now let's assume we do not want to play that loud and have no need to play louder than say 105dB peak levels.
This is loud but NOT uncomfortable with high DR recordings which is what most audiophiles like to hear at above average levels.
This would be around 90dB average.
this is VERY loud when listening to low DR songs with piercing treble !

The numbers:
used HD650 (300 Ohm)
90dB = 0.16mW  0.225V   0.7mA
105dB = 5mW   1.26V   4mA

If you want to be able to drive all headphones to 105dB (peak) but with NO headroom to spare which means an average level between 80 and 90dB for high DR and 90 to 100dB (VERY loud) for low DR recordings.

most sensitive headphones around: 0.2mW,  0.1V,  2mA 
average sensitive headphones: 1.6mW,  0.3V,&n bsp; 5.5mA 
somewhat less sensitive headphones: 29mW, 1.7V,  17mA   
difficult to drive ones: 270mW, 4.5V, 60mA   
the VERY few insensitive ones: 1.5W, 13.5V, 110mA 

some practical examples:

very sens. low imp headphone: (ATH-ES10) = 0.25mW, 0.1V,  2.5mA   
very sens. high imp headphone:  (GMP 435S) = 0.25mW, 0.25V,  1mA 

average sens low impheadphone: (Beyer T70P) = 1.26mW, 0.2V,  6mA   
average sens high imp headphone:  (Beyer T70) = 1.26mW, 0.6V,  2mA   

average less sens headphone: (K550) = 6mW, 0.5V, 13mA   
average less sens headphone: (DT770/600) = 8mW,  2V,  4mA 

difficult ones:
K172 = 160mW, 3V, 50mA   
K240 sextet (600 Ohm) 40mW, 4.9V, 8mA
HE6 = 630mW,  5.2V, 120mA   
K1000  1.3W, 13V, 91mA

So if you want to accommodate for the very rare 'truck' you need an amp that can supply
120mA x 13V = 1.5W  (driving trucks no headroom to spare)
BUT if we leave out the K1000 but include HE-6:  5.2V x 120mA  = 0.63W

For driving most headphones to 105dB peak one ONLY needs
2V x 15mA  = 0.03W  (C'Moy on 9V should not be a problem)
I am willing to take a little bet MOST pirates here have amps that are MUCH more powerful though 

driving ONLY sensitive ones:
0.3V x 3mA = 0.1mW which can be handles with even the smallest portable players/mobile phones.

 
Lets assume you want a beast of an amp that has no problem playing peaks of 120dB things change radically as currents and voltage levels increase by factors 5.5 and power by 30

So IF you want to listen to an HE6 REALLY loud on high DR music:
HE-6:  29V x 0.6A  = 17W (voice coil would burn out !)

most sensitive headphones around: 6mW,  0.6V,  10mA 
average sensitive headphones: 50mW,  1.7V,  30mA 
somewhat less sensitive headphones: 1W, 9.5V,  100mA  (damage may occur to some)
difficult to drive ones: 8W, 24V, 0.33A   (some will never reach it and burn out !)
the VERY few insensitive ones: 45W, 75V, 0.6A  (those will never reach it and burn out !)

So IF you like to play high DR really loud for very short periods you need high powered amps.
If you like to play high DR recordings comfortably loud, but with no headroom to spare you don't need nearly that much.

Not counting the amount of distortion as this is highly dependent on the used headphone.
Also not including the fact that if we want to keep distortion introduced by non feedback designs below audible levels even to reach 105dB peaks on comfortably loud levels we will need amplifiers that can supply MUCH higher voltages (and thus output power assuming current is no problem) to make such an amp sound GOOD at these levels.

I also wonder what it is about some amps that make me want to turn the volume up while others don't. The first time I went mid-fi, I was astounded at how good the amp sounded compared to without it. Then I realized I was cranking the volume on it which of course gave me more listening depth automatically. Once I tamed that it wasn't such a massive leap in sound quality.

The other thing I've noticed is that a headphone changes FR with volume so I suspect that an amp's grip will be different from another's and the user will notice that just from how the particular headphone reacts. I know that at low volume straight from my laptop one of my headphones sounds 'buzzy' but the other two don't. On my integrated amp there's no buzzing/rattling at any volume. So there is something about amp-to-phones pairing as well.

The problem is you can't say for sure what is causing those effects without actually having those things at hand and do some measurements.

As said before most 'normal' headphones SHOULD do well on a simple 9V C'Moy and a lot of them do, certainly up to a certain level. This is because it can only reach 3V (on high impedance headphones) and around 30mA.
This means that while you can play pretty loud it clips as soon as you go above its limits.
That clipping is a form of compression that only affects peaks.. those are bass notes with other signals superimposed. You can't reach full bass (phon curves) and clipping isn't making it sound nicer.

The same reason can be given for battery fed devices and some outputs of DAC's laptops that are limited to just a few V or less. When you amp it properly you can play louder (cleaner, no clipping) and louder means the bass could kick in harder.

Another reason could be the output capacitance of a lot of players (including i-crap) which rolls off the lows with low impedance headphones. As soon as you put an amp in the chain this problem (+ lack of output voltage to reach dynamic levels) is solved and SQ improves. Also you can go louder = better sounding.

Another reason can be crappy opamps used in the output of a device that start to distort at 'medium levels' already.
This could be caused by power supply problems, PCB layout problems, wiring problems. Most laptop PC board designers know a LOT about layout for digital stuff (trace lengths which is sometimes used in their advantage) and other tricks to get stuff working 'up to specs' but really may not have a clue about return currents in PCB's for audio.
They often don't care and slap an opamp on there because it is in the schematics that were provided (more often than not these are NOT the same persons)

There may also be other reasons of course.

Really hard to tell without having the used stuff at hand so the effect may be there and may be explainable but it is more fun for most to treat is as voodoo or as unexplained (yet).
Well that's my view on it based on experience and can be full of BS

I haven't heard the scalability phenomenon even in spot listening (i.e. level-unmatched, uncontrolled), which is why I've conditioned to the belief that the "full potential" mark is relatively attainable.

What's your chain and tubes used in your BA again?  We've done the same test many times in the past w/ the O2 and BA and clearly not the differences.  Either there is a choke in the chain, discrepancy in our methodologies or different metrics being used.  If all you listen for is aberrations/discrepancies in FR then I could see how you could come to that conclusion.