Jump to content
IGNORED

How do you want your treble roll-off served?


Recommended Posts

The measurements of actual concert halls presented here

https://pdfs.semanticscholar.org/e4c9/b3222f26472f6ba12e7fea900c2edb10c14c.pdf

 show a much more drastic drop of the frequencies above 5K than suggested by target curves such as B&K C, Bob Katz's or HK RR1 

The bass roll-off is consistent with the FR of the monitors used while the drastic treble roll off appears to be the hall response.

I recently frowned at the measurements of an expensive Audio Research amp (see Stereophile) but this, and some solutions blamed for rolloff, would it be filters cables or what have you, might well answer a respectable quest. Respectably ? Personally, I simply created new convolution files to mimic SI R2, preferred by assessors cf 

https://users.aalto.fi/~ktlokki/Publs/JASA_lokki2012.pdf

Although I can't say I suffered from brightness with my previous filters, the first impression is very good and paradoxal : more precision in instruments placement (while I had in mind that more treble favours soundstage) 

Link to comment
2 hours ago, Le Concombre Masqué said:

The measurements of actual concert halls presented here

https://pdfs.semanticscholar.org/e4c9/b3222f26472f6ba12e7fea900c2edb10c14c.pdf

 show a much more drastic drop of the frequencies above 5K than suggested by target curves such as B&K C, Bob Katz's or HK RR1 

The bass roll-off is consistent with the FR of the monitors used while the drastic treble roll off appears to be the hall response.

I recently frowned at the measurements of an expensive Audio Research amp (see Stereophile) but this, and some solutions blamed for rolloff, would it be filters cables or what have you, might well answer a respectable quest. Respectably ? Personally, I simply created new convolution files to mimic SI R2, preferred by assessors cf 

https://users.aalto.fi/~ktlokki/Publs/JASA_lokki2012.pdf

Although I can't say I suffered from brightness with my previous filters, the first impression is very good and paradoxal : more precision in instruments placement (while I had in mind that more treble favours soundstage) 

 

Can you share the FR curve or the filters you implemented?

Link to comment

We don't go to live concerts as much as I'd like, but every time we go (always in the nosebleed section, of course), I'm struck by the lack of treble; i.e., "where's the top?".  I've read descriptions of treble rolloff based solely on distance (i.e., attenuation through the atmosphere), fascinating to see it tested and displayed so clearly.

 

 

John Walker - IT Executive

Headphone - SonicTransporter i9 running Roon Server > Netgear Orbi > Blue Jeans Cable Ethernet > mRendu Roon endpoint > Topping D90 > Topping A90d > Dan Clark Expanse / HiFiMan H6SE v2 / HiFiman Arya Stealth

Home Theater / Music -SonicTransporter i9 running Roon Server > Netgear Orbi > Blue Jeans Cable HDMI > Denon X3700h > Anthem Amp for front channels > Revel F208-based 5.2.4 Atmos speaker system

Link to comment
1 hour ago, jhwalker said:

We don't go to live concerts as much as I'd like, but every time we go (always in the nosebleed section, of course), I'm struck by the lack of treble; i.e., "where's the top?".  I've read descriptions of treble rolloff based solely on distance (i.e., attenuation through the atmosphere), fascinating to see it tested and displayed so clearly.

 

 

Actually I attended a concert last night, third row, level with the musicians, but not nosebleed since it did not measure above 85 dBC. Maybe a trick of the mind but the  album with the same musicians playing same program sounded more like the real thing with the new filters

Link to comment
1 hour ago, Le Concombre Masqué said:

I won't be able to do measurements before tomorrow at best but this is what I can show :

 

I went back to the rephase matrix that yielded the attached FR and applied (frequency/dB/Q):

 

Left 3000/0.6/2 ; 6349/0.3/2 ; 10000/-1.35/1 ; 16000/-7.65/1

Right 1000/1.2/2 ; 1500/0.15/1 ; 3000/0.45/1 ; 4000/1.5/2 ; 6350/1.35/2 ; 10000/-4.5/1 ; 16000/-6.75/2

 

Q 1 & 2 @16K : just shows I made those quickly !

 

in short : a -2dB soft shoulder from 3 to 5 K and then straight down -7dB from 5 to 10 and an additional -3 dB down to 16K or so ( as I sais made quickly  and it slightly goes up from 16 to 20 

Proofing B&K C @ -9 copie-Modifier-2.jpg

 

This is close to the standard house curve I use for my system. That does sound best to me, and I’ve tried many hundreds of variations over the years before I settled on this one.

Link to comment

The perception of the sound overall alters dramatically, depending upon the quality of the presentation - this is something that I learnt from actually experiencing it over decades. And it has nothing to do whatsoever with the actual treble content - fiddling with the latter is mere band-aid stuff.

 

The subjective balance matches how one experiences live acoustic sound when jarring, disturbing anomalies in the SQ are sufficiently attenuated - one's brain then does the rest, and automatically adjusts for the actual, measurable SPLs. At this point the treble content fits in "perfectly", no matter what the circumstances are - the tone controls are all inside one's skull, and are far, far more skilful that any external, 'mechanical' compensation.

Link to comment

  I think my current loudspeakers are -6dB at 20kHz. But my ears are worse.

  And not much in fundamental tones above 15kHz. With harmonics there is a little up there.

 

2012 Mac Mini, i5 - 2.5 GHz, 16 GB RAM. SSD,  PM/PV software, Focusrite Clarett 4Pre 4 channel interface. Daysequerra M4.0X Broadcast monitor., My_Ref Evolution rev a , Klipsch La Scala II, Blue Sky Sub 12

Clarett used as ADC for vinyl rips.

Corning Optical Thunderbolt cable used to connect computer to 4Pre. Dac fed by iFi iPower and Noise Trapper isolation transformer. 

Link to comment
2 hours ago, AnotherSpin said:

I have similar observation. Real life concert hall classical or chamber sound slightly "less bright" than in my home system, while rock or popular or have more treble, certainly because sound-engineers add some treble in latter case.

acoustic live gigs don't show off as much airiness and soundstage as (some) audiophile systems...

Link to comment
10 hours ago, fas42 said:

The perception of the sound overall alters dramatically, depending upon the quality of the presentation - this is something that I learnt from actually experiencing it over decades. And it has nothing to do whatsoever with the actual treble content - fiddling with the latter is mere band-aid stuff.

  

The subjective balance matches how one experiences live acoustic sound when jarring, disturbing anomalies in the SQ are sufficiently attenuated - one's brain then does the rest, and automatically adjusts for the actual, measurable SPLs. At this point the treble content fits in "perfectly", no matter what the circumstances are - the tone controls are all inside one's skull, and are far, far more skilful that any external, 'mechanical' compensation.

Yes and I believe much more in state of mind in order to get a cooperative involvement of our brains to listening pleasure than in the state of electricity distribution or RFI. 

 

I have had ecstatic listening sessions in the past while my system was far less optimized ; but I have lesser and lesser bad days now (btw, our ears might fluctuate too and I don’t hesitate to force air with pinched nose and to move my jaws in order to optimize my Eustachian duct).

 

The question that puzzles me at the moment is : are there SQ anomalies that disappear beyond treble content?

I don’t have harshness/brightness to tame issues ; actually, hours after the short listening session I had with the new filters, it came to my mind I did not notice the rear of the stick scratching Elvin Jones performed on a cymbal and that I marveled to the day before… but stereo anomalies did not bother me any longer and Coltrane’s tenor was so real…

So, what kind of SQ anomaly could disappear that does not relate to brightness discomfort?

 

The first quoted paper shows the evolution of the FR over time. In a house direct sound from the speakers reach us @ about 10 ms while I designed my curve looking at what happens after 80 ms in Sibelius Hall. Is mimicking such FR at listening position + reproducing live SPL at Listening Position moot for we can't reproduce @ home the evolution of FR over time in a large studio let alone a large venue?

 

If my new filters prove satisfying while looking so extreme it might be just because of the extreme unnaturalness it confers to the room we chat and make noise walking moving etc in.

Link to comment
9 hours ago, Archimago said:

 

Thanks for the great post and link to the article! Will have to try these values next time I create some room correction filters.

 

For audio playback, we obviously need to also consider how the recording was created, especially placement of the mics. What's captured on the recording might already have applied much of that high frequency roll-off as opposed to a close mic'ed instrument where to sound more "real", we would need to implement more roll-off to simulate the larger space in a concert hall.

 

As such, maybe if we are to implement only one room-correction target frequency response, the B&K, Katz, etc. less steep roll-off at "only" around -6dB by 20kHz would be a fine compromise.

I share your concern and more

 

does it happen that sound engineers of amplified gigs think:” wtf with this venue? I want my 20k @ -6dB at my ears” and eQ?

 

what does mean mastering in the digital age? I hope we'll get interesting insights… let’s rule out mastering for Dre headphones or car mp3 ; let’s take the latest José James, the drums are so gorgeous that I checked : recorded in Capitol studio B. Which representation did the engineers have of the best compromise for customers’ systems/homes? And even if the engineers decided to get the best in a great room with a great system, I guess they still had to decide on a target curve, we don’t have access to, since the great room isn’t the 95m2 x 6 m ceiling studio itself. And I designed my incumbent filters with the idea that the best masterings are done with curves such as B&K C, Bob Katz’s, HK RR1, that all (including Katz’s, taking into account it’s “psychoacoustic”) go, roughly, down 9dB from 30 to 20 K + the most expensive most tweaked system I can think of with published measurements (M Fremer’s ; see https://www.stereophile.com/content/sonus-faber-aida-loudspeaker-measurements)

seem to follow those guidelines.

So, give or take +3dB below 150 I’m with you or was… 

 

Need to do more listening

Link to comment
9 hours ago, Archimago said:

 

Thanks for the great post and link to the article! Will have to try these values next time I create some room correction filters.

 

For audio playback, we obviously need to also consider how the recording was created, especially placement of the mics. What's captured on the recording might already have applied much of that high frequency roll-off as opposed to a close mic'ed instrument where to sound more "real", we would need to implement more roll-off to simulate the larger space in a concert hall.

 

As such, maybe if we are to implement only one room-correction target frequency response, the B&K, Katz, etc. less steep roll-off at "only" around -6dB by 20kHz would be a fine compromise.

 

The response in room as opposed to concert hall is not the same. In concert halls, you are listening to 2 seconds reverberation. The so called slope in concert hall is not intentional but due to the limits of high frequencies roll off due to longer distance they travel. 

 

There are are some trade off involved in concert hall sound. Between rolled off highs and long reverberation enveloping listeners, the preference will always be the reverbs. 

 

Having said that, this only applies to full orchestra music. For other types of genre, the rolls off may not be desirable nor the long RT. 

Link to comment
7 minutes ago, Blackmorec said:

As far as I can tell, this is logically a completely flawed approach. 

 

From a system point of view this is the equivalent of listening to every single recording in the same, highly attenuating venue. Every recording is going to have the same HF cuts applied, regardless of how the original recording was made. 

 

Instead of your system playing back musical performances from a variety of venues,  you are going to stamp every single recording with the self same venue response profile. In the past this was termed coloration in that the hi-fi imposes its own sonic signature on everything it replays. 

 

Of course there’s a very good possibility that I’m missing the point completely as the included academic paper was super-complex and written only for other academic researchers and specialists in the field. The maths alone is beyond 99.9x% of the audiophile population and even looking at the pictures isn’t helpful.. 

 

Personally, I find the idea of adjusting the hi-fi to compensate for shortcomings in a given recording to be pretty much a hopeless cause, since I don’t know how the original sounded, but adjusting the hi-fi to apply the same correction to every recording, whether required or not, just seems like a failure in logic. 

 

Also I would say that the goal of setting up a system to reproduce a particular live sound is also flawed as no 2 sound engineers are going to be working toward that exact same goal. Personally I would want to recreate perfectly what those sound engineers recorded, with the differences and nuances intact. 

 

My car stereo has this control that can make everything sound like it was recorded in several particular types of venue, going from small intimate club to something like a concert hall (or perhaps an aeroplane hangar).  After a few songs it becomes fatiguing then soon after downright annoying. 

On the one hand, I'm actually concerned that such extreme attenuation levels recordings, as you suggest ; on the other hand, I suggest you start there  https://www.innerfidelity.com/content/warren-tenbrooks-summary-head-measurements-harman

 

Link to comment
12 minutes ago, STC said:

 

The response in room as opposed to concert hall is not the same. In concert halls, you are listening to 2 seconds reverberation. The so called slope in concert hall is not intentional but due to the limits of high frequencies roll off due to longer distance they travel. 

 

There are are some trade off involved in concert hall sound. Between rolled off highs and long reverberation enveloping listeners, the preference will always be the reverbs. 

 

Having said that, this only applies to full orchestra music. For other types of genre, the rolls off may not be desirable nor the long RT. 

 I might have to fiddle between 2 convolutions filters?

Link to comment
4 hours ago, Le Concombre Masqué said:

you're a pioneer ; never saw such a bold target curve proposed. btw, you write "house" : great ; I'll comment it below

 

No, not me. I'm following a bunch of others who have done this and studied similar curves. REW calls the desired curve a 'house curve', and I think that's generally the name used for it, but I'll agree that it's not correcting for the house or room, but rather for an overall system response.

 

 

Link to comment

I think in this context "house" has a different meaning -- i.e., their personally-preferred, customized response curve that subjectively sounds best (after room correction, speaker correction, etc), instead of a completely flat response, which a priori one would think should sound the most neutral (but in reality artificially compensates for natural, distance-dependent treble roll-off, etc.).

 

Kind of like "home brewed" doesn't literally mean brewed in your house.

Link to comment

When I go to a live concert - a good seat in a good hall with a world class orchestra -  I’m often struck by the gorgeous sound of the orchestral strings.    A wonderfully smooth silky bloom, but with such inner complexity and life.   

 

Often all we get from (even quite high-end) hifi systems is a harsh, shrill travesty.  Especially fortissimo, above the stave, violin passages.   It’s possible that simple excess brightness is the main culprit here.

Link to comment
36 minutes ago, wgscott said:

I think in this context "house" has a different meaning -- i.e., their personally-preferred, customized response curve that subjectively sounds best (after room correction, speaker correction, etc), instead of a completely flat response, which a priori one would think should sound the most neutral (but in reality artificially compensates for natural, distance-dependent treble roll-off, etc.).

 

Kind of like "home brewed" doesn't literally mean brewed in your house.

How does a completely flat response curve compensate for anything?  Flat surely means that it neither adds nor subtracts anything?  If there’s a natural roll-off of the treble then a flat response curve will reproduce that same roll off.  If the roll-off is say -9dB at 15KHz, then passing that through anything with a flat frequency response will result in the same -9db roll off  i.e nothing added or subtracted.

27 minutes ago, mulberry bush said:

When I go to a live concert - a good seat in a good hall with a world class orchestra -  I’m often struck by the gorgeous sound of the orchestral strings.    A wonderfully smooth silky bloom, but with such inner complexity and life.   

 

Often all we get from (even quite high-end) hifi systems is a harsh, shrill travesty.  Especially fortissimo, above the stave, violin passages.   It’s possible that simple excess brightness is the main culprit here.

Probably the main culprit is lack of resolution which fails to separate ‘the smooth silky bloom’ from the instrument tone. The bloom is still in the signal, as a better system would resolve it; but that unresolved information is included as distortion in the balance of the signal, making it sound hard and shrill.  When upgrading a hi-if system, you’ll generally notice that any increase in air and bloom is always accompanied by a reduction in harshness....as the ‘distortion’ resolves into additional detail, it is by definition removed from the balance of the signal. 

Link to comment
On 10/7/2018 at 8:38 AM, Le Concombre Masqué said:

The measurements of actual concert halls presented here

https://pdfs.semanticscholar.org/e4c9/b3222f26472f6ba12e7fea900c2edb10c14c.pdf

 show a much more drastic drop of the frequencies above 5K than suggested by target curves such as B&K C, Bob Katz's or HK RR1 

The bass roll-off is consistent with the FR of the monitors used while the drastic treble roll off appears to be the hall response.

I recently frowned at the measurements of an expensive Audio Research amp (see Stereophile) but this, and some solutions blamed for rolloff, would it be filters cables or what have you, might well answer a respectable quest. Respectably ? Personally, I simply created new convolution files to mimic SI R2, preferred by assessors cf 

https://users.aalto.fi/~ktlokki/Publs/JASA_lokki2012.pdf

Although I can't say I suffered from brightness with my previous filters, the first impression is very good and paradoxal : more precision in instruments placement (while I had in mind that more treble favours soundstage) 

This tallies with my findings as well. Using a hand-held spectrum analyzer, I found severe roll-off of high frequencies above 7KHz in The following venues: San Jose (CA) Center for the Performing Arts. Flint Center; De Anza College, Cupertino CA, Davis Center for the performing Arts,. San Francisco, CA, (home of the San Francisco Symphony).

In all three cases, high-frequencies rolled off rather precariously after the first couple of rows of seats (these measurement are taken in an empty venue) as shown in the Finnish paper. But when measured from my preferred microphone position in all three halls, I found no appreciable roll-off (mikes hanging from the proscenium arch; 5 ft behind the conductor and to ft over his head). But one has to keep in mind that most modern cardioid-pattern condenser microphones have a slightly rising top-end above 8 KHz. This would compensate, somewhat , for any natural roll-off on the stage itself. 

 

George

Link to comment

I'm (almost) back to the convolution filters that yielded the measured curves I published above : I just kept the slight uplifts between 1 and 6 K that straightened things a bit.

 

Horn here, strings or piano there, the focus put on the fundamentals of, yes, rather low frequency notes, the fact that transients of low key notes are perceived moved down, has its charms but with extreme rolloff I miss the impact of drums and the slam of, yes, some bass notes. Even at realistic SPL ; if it's not loud enough it's even worth.

 

 

 

 

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...