[castleofargh]

I'm enjoyin all the hi-fi talk.

To go back to some of the OP's original questions though... In case it wasn't apparent from some of my previous posts, I do also think that a neutral response should include some type of a bass boost. Because I subscribe to the "room sound" theory. When you put speakers that measure flat in an echo-free chamber into a room with some reflectivity, they acquire a darker tilt. Because they do not (usually) have the same amount of dispersion or directivity at all frequencies. This is what Floyd Toole talks about in the video I posted earlier.

The shelving question is a bit harder to answer. Speakers are imperfect instruments though. And eventually start to falloff at some point in the lower frequencies. And if you are using that as your model for the sound of a headphone, then it may make some sense to either level or roll off the sub-bass. At least a bit, to better approximate that. Exactly where and how much is sort of up for grabs though.

It would not be that hard for the manufacturers of the head and torso rigs which are used to measure headphones to include a compensation curve for room effects like this. They could very easily set up a pair of flat speakers in an average sized room, with average reflectance. And measure what the in-ear response is on their rig from a typical listening position. Or averaged over several positions.

Doing the measurements for a free field and diffuse field are actually alot harder, because those require either an echo-free chamber. Or a totally reflective room. Neither of which is especially easy to create. A room that replicates the average size and reflectance of a typical listening space in a home would probably be a trivial task to setup by comparison. So I don't know why it has not been done before.

If they wanted to, they could even do this for a variety of different spaces, and speaker configurations. So you could take your pick and choose the one you like the best. That sort of information would probably also add some add'l value to their products.

I believe the very first FR that Harman gave to listeners was what you describe, while following their first assumption; That if we present the FR preferred on speakers at the eardrum, it would also be preferred on headphones.
But rapidly they got a trend in how listeners were EQing it in the general direction of a warmer FR.
My educated guess as to why that approach doesn't work is this:
-the matter of missing tactile bass.
-the various issues coming from the massive differences in presentation between stereo album on speakers and stereo on headphones(without crossfeed, reverb, or our own HRTF cue).

Basically, same FR alone does not give us the same experience or even same sound in this context.

 

[sander99]

It is my understanding though that the reference listening level standards described in that THX article are designed to accurately replicate the conditions in which the movie's soundtrack was mastered. So the medium shouldn't really have any effect on the final result that you hear in that regard.

Exactly, the end goal is that if a track was mastered under these precisely described conditions, and somewhere else presented under these same precisely described conditions then this presentation will be the same as in the mastering room (at least with respect to the levels and overall balance). That effectively means that the the transfer from mastering room to presentation room is flat! But apparently they master tracks in such a way that the result contains (or can contain) low frequencies at much higher levels than the maximum levels for higher frequencies. I am guessing the reason for this is that strong low frequency effects can add very spectacular thrills to the movie experience, while at the same time they are less dangerous for your ears compared to other frequencies at those levels(?).

And if you get a bigger or 2nd sub to cover the cross-over information (like the article suggests), then the sub(s) should still have the add'l 10 dB on top of what it needs with the cross-over active. No?

Yes, but this is about maximum SPLs (that are needed for the actual content as it is - or sometimes can be - including very strong low frequencies), not about changing the frequency response. Oh, and of course the cross-over information - as you call it - will not be boosted 10 dB when played over the subwoofers instead of the other speakers, otherwise changing from non-crossovered to crossovered would effectively change the frequency response indeed!

 

[ADUHF]

Fwiw, I found a link to this PDF in another topic on reference levels...

http://www.atsc.org/wp-content/uplo...tablishing-and-maintaining-audio-loudness.pdf

And in the very last chapter on "Audio Monitoring Setup", in tiny print on pages 34 & 35, just before the Annexes, it says this...

Attention is also drawn to the level alignment of the Low Frequency Effects or 0.1 channel. Some confusion occurs due to its requirement to have 10 dB of in band gain compared to the main channels. This does not mean that it should measure 10 dB greater in sound pressure level than the main channels when calibrating. Because the LFE channel’s bandwidth is more limited than that of the main channels, its level playing a pink noise source of the correct spectrum and electrical level will be approximately 4 dB greater than one of the main channels.

If this is correct, then it means pretty much anyone who's doing their own measurements, and setting the sub-woofer/LFE 10 dB higher is actually doing it wrong... Good grief.

And a 6 dB difference in the sub-bass setup (ie the difference between 10 and 4) would make the LFE effectively about twice the volume that it should be, perceptually. That's a big difference.

So much for ensuring consistency I guess.

There are links in the same chapter to some band-limited pink-noise WAV files for setting the correct levels in both the sub-bass and midrange as well... Supposedly.

 

[ADUHF]

He’s right to point out that Harman target is not flat. They clearly tested for preferred FR. It could be argued that at large, perceived flat or a slight tilt toward warmth tend to be preferred by most, but Olive&co did not say anything about seeking neutral.

Back to bass, ambient sounds have tactile bass impact that strongly increases and changes how we perceive the low end. It's been hypothesized by many that the desire for more bass on headphones and IEMs, might be our way to compensate for the missing tactile feeling. Is it the reason? IDK. But the trend of wanting more bass than on speakers or acoustic events is long going. As it is only a substitute, it makes sense to me that different people might interpret it differently. Some could feel like they are getting back some of those visceral sub feelings, while maybe all you feel is too much low end disturbing the tonal balance? I only experienced me, so I couldn't say how most other listeners feel.

Personally, I have another issue where the measurement tool might give me similar low end when I very clearly feel them as different(even after applying a solid low pass to remove differences in the upper range. After matching the SPL output and after doing my very best to find tips that would seal my ear canal).
This is an example of what troubled me with bass measurements(first paragraph after the graphs):
https://www.head-fi.org/threads/hea...setups-different-results.751100/post-13933374
And this is all done by me so I could control the measurements and make adjustments at will. And repeat such measurements as many times as I wanted ever since. It's just an example with IEMs, but I felt the same way with various headphones, where similar low end(isolated) felt clearly more ”tactile” on certain headphones. IDK why. Maybe distortions, maybe the amount of skin receiving the wave. I also considered that some headphones might get shaken more than others, or at least transmit more of the vibrations through the pads, giving us some actual tactile bass, albeit only localy.
It's a trend for me those days, but I don't have an answer.
I just know that at large, people seem to enjoy a little bass boost on headphones, and that as you pointed out, they might like to have more boost on IEMs. That seems to be preferred for reasons. Will it feel neutral? IDK. Probably not.

Fwiw, Jude addresses the question of the higher bass measurements on IEMs near the end of this video. And suggests that it could have something to do with the inaccuracies in the simulated ear canals in current 711 HATS measurement rigs...



...So I guess that means that the more accurately modeled ear canal on the new HB&K 5128 rigs could potentially deliver more reliable measurements on this, that might possibly be more consistent or in line with over and on-ear headphones. Guess we'll have to wait and see on that though.

 

[ADUHF]

Oh, no wonder I have been so lost in this tread. Harman Target Curve is a thing apparently and it simulates how loudspeakers sound in a room. I need to study this subject before I can understand the content of this thread properly I think. I assumed this was simply about "bassy" and "neutral" sound. Sorry about my ignorance.

It is a thing, as you say. But not everyone agrees with it. This is from Dr. Olive's twitter...

This is the same over-ear curve shown above in blue, with diffuse field compensation...



Note that the bass response below about 160-200 Hz appears roughly the same on both curves. That's because the diffuse field curve is essentially flat below that frequency range. And the compensation for the ear-related gain effects is primarily in the higher frequencies centered around the peak at approximately 3 kHz on the blue curve.

With the gain at 3k from the shape of the ear removed, the Harman curve starts to look alot like a loudspeaker's room or sound power curve.

 

[ADUHF]

I believe the very first FR that Harman gave to listeners was what you describe, while following their first assumption; That if we present the FR preferred on speakers at the eardrum, it would also be preferred on headphones.
But rapidly they got a trend in how listeners were EQing it in the general direction of a warmer FR.
My educated guess as to why that approach doesn't work is this:
-the matter of missing tactile bass.
-the various issues coming from the massive differences in presentation between stereo album on speakers and stereo on headphones(without crossfeed, reverb, or our own HRTF cue).

Basically, same FR alone does not give us the same experience or even same sound in this context.

This is going back a bit, but I think the original Olive-Welti curve presented in a 2013 AES paper was intended to depict the approximate response of a loudspeaker at the ear-drum. And in raw form that curve looks something like this...



I haven't bothered to try applying diffuse field compensation to this curve, btw, because it appears to have been made on a different type of rig (possibly one similar to the Head Acoustics rigs used by Rtings and Inner Fidelity) than the later Harman target plots, and Oratory's graphs. So it would probably require a different curve for diffuse field compensation than the one used in Oratory's graphing tool.

I won't bother going into more detail on this here, but each different manufacturer's measurement rig will use a slightly different curve for diffuse field compensation to account for the small variations in their anatomical design, and simulated ear shapes.

The first versions of the Harman target that I can recall seeing (after the Olive-Welti curve) were in Tyll's articles on Inner Fidelity, like this one...

https://www.stereophile.com/content/innerfidelity-headphone-measurements-explained

As explained in Tyll's article, and in the Harman video that NA Blur posted earlier in this topic, the green curve above represents the sound of a loudspeaker electronically calibrated to have a totally flat response at the listener's position. So basically with all the room-related gain effects removed.

The black curve is a combination of the green curve (so, flat response at listener position with no room gain). And Harman's original preferred response curve, based on their subjective tests.

The video that NA Blur posted explains some of how that curve was derived...



Subjects were essentially given a bass and treble control, and asked to adjust the sound of the green curve until it seemed more or less correct to them. And the black curve represents a rough average of those first tests. A little more about this here...

https://www.headphonesty.com/2020/04/harman-target-curves-part-3/

This was still early in the curve's development though. And the Harman target has been revised several times since then, to also include IEMs and a broader cross-section of listeners. So it's now basically a family of several different curves for different demographics, and headphone types. Which Harman uses to design products for different types of sound consumers. And the blue curve shown in the previous post above is a rough across the board average of the preferences for an over-ear headphone's response.

https://www.soundstagesolo.com/index.php/features/217-where-are-we-at-with-the-harman-curve

 

[ADUHF]

Here's a link to the original AES white paper that is the source of the above curve. It was actually from 2013 (so not quite as far back as I originally thought)...

https://www.aes.org/e-lib/browse.cfm?elib=16768

 

[ADUHF]

This is the original plot that my drawing of the Olive-Welti curve (which was actually made in Windows Paint ) comes from...

https://www.reddit.com/r/headphones...nears_compensation_curve_vs_olivewelti_curve/

It compares the Olive-Welti curve to another target response curve that was previously used for compensating raw headphone response measurements by a Korean headphone review site called Golden Ears. Which is now defunct.

 

[ADUHF]

I believe the very first FR that Harman gave to listeners was what you describe, while following their first assumption; That if we present the FR preferred on speakers at the eardrum, it would also be preferred on headphones.
But rapidly they got a trend in how listeners were EQing it in the general direction of a warmer FR.
My educated guess as to why that approach doesn't work is this:
-the matter of missing tactile bass.
-the various issues coming from the massive differences in presentation between stereo album on speakers and stereo on headphones(without crossfeed, reverb, or our own HRTF cue).

Basically, same FR alone does not give us the same experience or even same sound in this context.

It would still be interesting to see the actual in-ear measurements of loudspeakers in a room for comparison though, imo, based on an average human anatomy. That would at least give us a better starting point for understanding some of this HRTF-related stuff a little better.

And perhaps when our ability to measure and understand these things is a bit better, the differences between the headphone vs. speaker in room might turn out to be quite small, if not insignificant.

Variations in human anatomy will always have some effect in the final measured sound at the eardrum though. So that will probably always have to be taken into account, to some degree, in designing different types headphones for different types of users. IEMs are likely the most problematic in that respect, because they have to take into account both the differences in the ear anatomy, and also the head and body-related effects. While over-ear headphones only have to deal with the two latter things. So there's potentially a little less variation from individual to individual.

 

[bigshot]

But apparently they master tracks in such a way that the result contains (or can contain) low frequencies at much higher levels than the maximum levels for higher frequencies. I am guessing the reason for this is that strong low frequency effects can add very spectacular thrills to the movie experience, while at the same time they are less dangerous for your ears compared to other frequencies at those levels(?).

The philosophy of mixing for movies is different than mixing for music. With music you want ebb and flow to create an overall balance. With movies, you need to emphasize contrasts to support the images on the screen. If you're seeing a bomb go off and scatter rubble everywhere, you need to *feel* that energy in the sound. In general, movies have a much wider dynamic range from quiet to loud. The one philosophy that is the same with music and movies is that you aren't creating a realistic sounding experience. You are creating a hyper-realistic sound. It isn't just about the way the musicians sound in the studio, or capturing the sound of the place where the movie is shot. You are building an optimized sound layer by layer that is more real than real.

 

[bigshot]

I think you're probably still getting something roughly on the order of an extra 10 dB in the sub-bass frequencies at the listening position, whether it's handling the bass for the other speakers or not. No?

I was confused about this at one time as well, and asked Gregorio about it. Apparently the 10db bass difference is just for theaters. When they master for home video, they correct to make it match home audio calibration. So the bass produced by a movie is the same level as bass produced by music when you play a blu-ray in your home.

Speakers are imperfect instruments though. And eventually start to falloff at some point in the lower frequencies. And if you are using that as your model for the sound of a headphone, then it may make some sense to either level or roll off the sub-bass.

Good speakers are much better at producing sub bass than good headphones are. My sub produces sound down to 14Hz at a remarkably high level. It can rattle the walls without you even hearing a sound.

It compares the Olive-Welti curve to another target response curve that was previously used for compensating raw headphone response measurements by a Korean headphone review site called Golden Ears. Which is now defunct.

I doubt you could hear much of a difference between those two response curves. They appear to be no more than a couple of dB apart. Maybe just barely noticeable in the bass, but I tend to doubt you could hear that in music, only with tones.

 

[ThanatosVI]

The harman curve bass boost isn't nearly enough for a basshead.

I get that this could have been an April fools post, but on the off chance it is not, it's terribly inaccurate

 

[bigshot]

There are two Harman curves with different bass responses. Take your pick. The Harman Curve isn't a target curve. It is an *averaged* response of people's polled *preference*. This makes it a good baseline to begin EQing to your own taste. Odd are, your own preference might be different, but it won't be too far off the Harman Curve.

 

[ADUHF]

I was confused about this at one time as well, and asked Gregorio about it. Apparently the 10db bass difference is just for theaters. When they master for home video, they correct to make it match home audio calibration. So the bass produced by a movie is the same level as bass produced by music when you play a blu-ray in your home.

Not sure about this. Here's what the following link says...

https://www.thx.com/questions/what-is-the-reference-level/

THX Certified Audio products are designed to let you experience movies the way they were meant to be heard. On every THX Certified AV Receiver the “0” on the volume dial represents THX Reference Level, the exact volume level used by the moviemakers and sound artists in the studio...

Experience Studio Clarity: THX Certified Receivers reproduce studio Reference Level, 85dB SPL with 20dB of headroom.

It makes no reference to the higher LFE levels. But it seems to be implicitly understood that goes with the "reference levels" being referring to. Based on looking at a number of topics in different forums on this subject. Including AVS Forum.

The ATSC paper that I referenced above also makes a point of commenting on the 10 dB increase for the 0.1 LFE channel in their audio setup recommendations for authoring this type of content. The only thing it seems to call into question is the proper way of measuring it.

My guess is that most receivers will take care of this automatically and transparently for you. So you don't have to think about the different measurements for the full range vs. LFE speakers. And it will all just come out in the wash correctly. (One would hope anyway.)

My main interest in the reference levels was not the difference between the LFE and other channels though. That was mostly an afterthough. It was primarily to get a general idea of the maximum decibels needed to approximate something close to a similar result in a pair of headphones. (Should I ever want to do something like that.)

Based on the THX link above that would be 85 dB, +20 dB for the additional headroom to reach 0 dBFS for the full range channels. Or about 105 dB in total. The ATSC paper also has some recommendations though on how this starting 85 dB (with a -20 dBFS pink noise signal) reference level should be adjusted for different sized rooms. And those recommendations are shown in Table 10.2 on page 36 of the above link. Fwiw, here's how those seem to break down, based on the cubic footage of the room...

Categories:	Room Size in Cubic Feet:	-20 dBFS SPL in dB re 20 µN/m2:	Peak 0 dBFS SPL:
I,II	> 20,000	85	105
I,II	10,000 < 19,999	82	102
I,II	5,000 < 9,999	80	100
I,II	1,500 < 4,999	78	98
I,II	< 1,499	76	96

I added the 4th column based on the original 20 dB of overhead for the 85 dB reference level. And I don't know what the categories in the first column refer to.

This does not take into account the additional 4 or 10 dB potentially needed for the LFE.

In a home setting, the above levels can probably be achieved by just lowering the master volume control by a comparable amount below the unity setting, after the system has already been correctly calibrated to the proper reference levels (based on 85 dB at -20 dBFS).

Good speakers are much better at producing sub bass than good headphones are. My sub produces sound down to 14Hz at a remarkably high level. It can rattle the walls without you even hearing a sound.

Sensurround fun!

I doubt you could hear much of a difference between those two response curves. They appear to be no more than a couple of dB apart. Maybe just barely noticeable in the bass, but I tend to doubt you could hear that in music, only with tones.

The Golden Ears website that used that other curve for compensation also made a couple different adjustments in their target response curve. So I'm not sure that was the final one that they ended up with.

 

[bigshot]

I'm talking about the 10dB difference in the LFE channel. It is two different exports of the mix- one for theaters at -10dB, and one for home viewing without the added headroom. It's the volume level of the track itself. If an amp has headroom, that is completely separate from the track itself. They're just saying the amp is powerful enough to produce sound up to 105dB in theory. That would depend on the speakers though. That is probably just hooey. A lot of consumer THX "specs" is just sales pitch. They had strict standards for theaters, but for home theater, it is pretty much just a brand name you pay extra for with no real benefit. You really can't assume that a statement about theatrical THX applies to home THX.