Btw, what is even reference level?  I have some issue with this.  You're supposedly trying to reproduce the sound of an instrument at a concert.  First, most concerts/music have amplified sound.  In those cases you're leaving all judgement to the sound engineers or performers or bartender on what constitutes 'reference', not to mention it's going to largely depend on the size of the venue and where you're sitting.  In the case that you're using unamplified acoustic instruments as your reference, you still haven't overcome the variable of distance.  Everybody already knows volume will depend on how far away you are from the source.  Are you standing in back of the venue?  Are you in the middle of the room?  Maybe you got good tickets and are in the first row?  Are you in the band and on the stage?  Perhaps you're the drummer/pianist/whatever yourself?  There are even more factors than just distance but I'm just trying to illustrate that there are major issues with claiming "reference" levels - and this is only going by mic, not by ear (not all ears are equal as I said earlier).

This isn't even getting into things like venue acoustics, listening angle to soundwave origination, instrument models, playing style, etc.  The "reproducing live music" thing isn't a particularly bad goal to aim for, I just find that most people who use this catchphrase aren't really aware of just how variable it is.  It could mean almost anything.

I think I basically agree with everything Mav is saying, including HPs being unrealistic due to soundstaging, even for the best stuff.  The only stuff that has been way more convincing in terms of spacial reproduction is binaural/Smyth.  Even then, you still run into severe issues with bass reproduction on headphones, be it a Stax Lambda or Abyss.

Wow, so much obfuscation and Mickey Mouse-ing in here. Let's answer the question simply. Actually, you kinda answered it yourself in the first statement. Reference level is the level at which the instrument was recorded. That way it is a useful "reference" for playback, i.e. the reproduced instrument sounds as close as possible in SPL to the recording. It really doesn't get much simpler than that.

And, again, reference levels are important because they let the audio engineers at different stages of the production process have that reference to fall back on. I'm sure most engineers will use calibrated mics and play with their preamp and monitor gain until they tune in things perfectly. I bet a lot of others can just step in the live room, then back into the control room and already know where they need to set things in order to monitor accurately. I find that the sound of music is rather intuitive like that.

I'm sure you can come up with a lot of variables regarding the recording and production process. But don't bother if your aim is to somehow justify your own listening preferences. Because the bottom line is that music doesn't sound like that. It needs more range to breath. If it's any consolation, I'm sure in 60 years your hearing will be pristine while Anax and I are shouting at each other from 2 meters away.

I know all that but since I misunderstood Anax's figure as a peak relating bass frequencies instead of average I picked the weighting that was more accurate in the bass.

Wrong, you're not following. You playback white noise (or pink noise) at a certain level and measure it with C weighting, the value will always be higher because it doesn't roll off the bass. You measure it with A weighting and it does do so and you get a lower reading. The A weighting is more in line with what you hear, and it's what everyone refers to when they talk about SPL. Anax's specific test has nothing to do with it. In summary, that means you're playing back at a fair bit lower level than you think you are. How much lower will depend on the energy distribution of the given signal.

I trust you already understand why white noise is a poor indicator of perceived sound level than pink noise....

One album isn't a very long time and most of my favorite music/albums are casualties in the loudness war.  Even with properly mastered tracks my ears tell me to turn it down before an album is finished which is why...

Obviously you should compensate for recordings that are not competently recorded/produced. Discretion as always.

I'm gonna reply to wnmnkh's questions. I disclaim I am a novice.


Dynamic range concerns the dB level at all frequencies. The heK/1000 is relatively flat (very flat 1k and down). The benefit of flatness is to maximize audio reproduction throughout the range, so a single tone of an instrument can continuously evolve and devolve through its sound envelope. This comes across as natural, organic, analogue; we almost never hear discrete (non continuous) decay in nature. I mean by discrete a recession in FR, because, compared to the rest of the envelope, there is less sound there and I feel jarred, much like someone cutting themselves off in the middle of a speech.

Flatness's affiliates are smoothness (continuity in slope, as close to zero slope as possible, and also smallest lengths of hills/valley and smallest amplitude). It is preferable to have a FR of continuous wiggles half an octave apart if they are +/- 2 dB than the LCD2 super flat then roll off.

Maybe a certain dB range of the FR then, over say 9 octaves (the last octave being the very lowest and very most top).

But then . . .ughh. . . I mean certain FR ranges can convey certain dynamic information. A hi-fi headphone is exciting b/c it has a huge dynamic range as the music plays because the headphone can give you a dB rush in the bass when the cymbals are not playing, or have finished their initial spike and are now in decay.

So THD is important so the actual frequency plays, and thus that information in each pitch (say by the measure of cents and fractions of it), so the fundamental of an instrument can rightfully evolve through its envelope.

The noise is similar to the dB FR in that it increases the audible dynamic range.

The CSD is like room coloration of the headphone. Faster is better. So is impulse response.

But impulse response must be balanced with ability to pick up low level info. Plankton extraction is related but not mutual with FR, noise, THD. The driver must be able to physically move in response to a minute signal. The driver tension, electromotive method (ortho, coil, stat), internal damping/resonances caused by these (say with magnet structure, or stators bouncing) matter--

Orthos can sound dynamic because the diaphragm where the traces physically are immediately move and thus you get a full sound. The traces run aside another across almost the whole diaphragm. Bass frequencies require excursion of the entire diaphragm. Traces run the entire diaphragm. And thus the full, extended ortho bass. The sounds radiate from the driver traces. So, the sound reproduction, I conjecture, works by sound radiating from not a planar surface but from the traces, with the diaphragm base simply provide tensegrity and multiply the sound. But the actual signal is purest at the traces. In between the tone is since the traces are long and winding, I imagine they can distort as the waves emitted from the traces smack into another.

A whole bunch of traces solves this but it makes the headphone damn slow, partly negating its benefit, due to weight and relative lack of flexibility of the metal compared to bare diaphragm base material, like polyamide or mylar. LCD 3. As opposed to the better done heK.

Stats have limited excursion, just like orthos. The orthos have the magnet array, the stats have stators. But the orthos force increases significantly as the diaphragm nears the magnet, whereas with stats, it does not. That causes bouncing with huge dynamic swings, especially in the bass, or simply a flattened amplitude. 007 w/ Sheffield.

The HD800 has shown itself able to pick up information and thus recorded dB lower than anything else. But the FR information (and lowest bass is lacking. 500-5k Hz is a rolloff before the spike up to the plateau. And bass rolloff starting at 300, gradual, then an increase in slope at 200, and more at 100.

People can prefer more dynamic range a t certain frequency bands (say the HD 800, which has great bass, but needs the amplitude of a flat FR), or give or take as a whole (in the case of the HeK). I prefer the latter. But if I had flatness with the HD800. Best headphone ever.



Compression: The loudest tone of the smallest detail is louder. But the details of the details can be removed, truncating it.

I don't really know which is the absolute best technology. They have their own damping schemes, driver construction, response to outside equipment/sources.

Wow, so much obfuscation and Mickey Mouse-ing in here. Let's answer the question simply. Actually, you kinda answered it yourself in the first statement. Reference level is the level at which the instrument was recorded. That way it is a useful "reference" for playback, i.e. the reproduced instrument sounds as close as possible in SPL to the recording. It really doesn't get much simpler than that.

And, again, reference levels are important because they let the audio engineers at different stages of the production process have that reference to fall back on. I'm sure most engineers will use calibrated mics and play with their preamp and monitor gain until they tune in things perfectly. I bet a lot of others can just step in the live room, then back into the control room and already know where they need to set things in order to monitor accurately. I find that the sound of music is rather intuitive like that.

If I'm intentionally trying to make things harder to understand then I guess I did my job, as you're still not taking distance into account.  You say you want to replicate the dB level as the instrument was recorded, but the mic was almost certainly WAY closer than you'll ever be at a live event.  Look up the relationship between distance and SPL.  Instruments are played loudly to compensate for how far away the listener is.  If the listener was at the mic (your 'reference'), the musicians would be playing much more quietly. 

Even after this you're still chasing the subjective opinion of the producers who almost certainly pumped the volume level way up on modern stuff.  This is why classical makes way more sense for determining what a true reference volume level should be during playback: the composers fucking say it themselves (ppp, fff, etc).    The last time I heard the creators say what level people should listen at was Nintendo:

But don't bother if your aim is to somehow justify your own listening preferences. Because the bottom line is that music doesn't sound like that. It needs more range to breath.

I'm not trying to justify anything.  I've already stated that I listen at a variety of levels and that a lot of the music I listen to actually has a lot of dynamic range, yet you're still trying to pigeonhole me.  I don't care what level people listen at; you're the one going on about what the only proper listening levels are without taking into account mic placement, composer intent, mastering, etc.

IMO people who are truly concerned about dynamics should be using vinyl rips at every possible chance before they complain about the small stuff, especially for modern rock/pop/etc.

Do people agree with what I said earlier about perceived 'macrodynamics' correlating more with 'impact' than with SPL?

Do people agree with what I said earlier about perceived 'macrodynamics' correlating more with 'impact' than with SPL?

You're probably right. at least it matches my own opinion .

As you do, I struggle to understand how people can complain about dynamics while the listen mostly to heavily compressed music.