Separating a tweeter from a midrange driver (including horns, if any) by more than 1/4 wavelength at the center crossover frequency introduces terrible lobing behavior external to the loudspeaker, i.e., nearfield and farfield directivity anomalies. (Hence the invention of concentric dual diaphragm drivers and MEHs to keep the two driver mouths within 1/4 wavelength at crossover.)
Pretty soon, you'll come to realize that you don't want to cross any two drivers above ~1000 Hz--give or take a half octave or so, depending on the radial sizes of the tweeter and midrange driver housings and how close you can get them together on the front baffle or inside a horn.
Below the Schroeder frequency of a listening room, all you have really is modal behavior, so placement of drivers/horn mouths is reduced to "filling up the room modes" at mid-wall locations and room corners, etc. in order to put energy into these separate modes to even the response in-room.
At the transition between the sparse mode region and the dense mode region (the Schroeder transition frequency band -- about 80-210 Hz in most home-sized listening rooms), the human hearing system is distracted by room resonances--such 1/2 wave across the room width, length, and height, and the tendency of loudspeakers to pump a lot of unwanted energy into nearfield reflectors just around the loudspeakers due to the loudspeakers' loss of directivity control. This is the "boominess" region that was described above.
These are the factors of interest--not so much electrical crossover filters, IME. Beyond that, you're now dealing with recording mastering artifacts.
Chris
Last edited:
OK let's digest each of these "mono-ing factors" and ask "what do we have control over?"
We can't make any changes about item 1, our hearing is not adjustable.
We can't make big changes about item 2, not after a room is built and furnished, anyway.
We can make big changes about item 3, crossover tuning and/or just using different hardware are both in play.
Only one of these is something we can actually adjust.
We can't make any changes about item 1, our hearing is not adjustable.
We can't make big changes about item 2, not after a room is built and furnished, anyway.
We can make big changes about item 3, crossover tuning and/or just using different hardware are both in play.
Only one of these is something we can actually adjust.
I don't know where you got that transfer function for an elliptical filter but it doesn't look in any way to what i've seen used on cuting lathe we used to have track s cut ( Neumann VMS most of the time, can't remember the module reference of elliptical filter in Neumann range but itwasn't a Chebyshev or anything exotic at all) neither the one i applied at mixing stage expecting what the Mastering Engineer will have to perform anyway ( which were most of the time an MS matrixing with filtered Side with a 1 or 2 pole hp filter to keep thing simple and under management as well as easy to recall in case the ME aslked for something different).
Undoing mastering is a fantasy in my view. And what you call features are most of the time requirements rather than subjective aesthetical decision.
I think we already had this discussion before Chris, and you do what feels good to you and that is fine, but don't use terms like 'demastering' please, it introduce too much confusion to people not involved into the technical process of musical production:
Given the area of music you mostly listen to Chris, it is possible to give some kind of eq 'corrections' to please anyone taste but it's in no way possible to correct any dynamic treatments applied even more so if multiband processing was involved which is the norm since 00's.
And given the dynamic range we are now facing eq treatments (or expansion techniques to try to compensate for dynamic) ask some skills and understanding of technical limitation most amateurs don't have ( treatments of signals with no headroom left is not for beginners).
I would not state the room is mono at low freq, rather the modal region of the room behave really differently than above schroeder freq. and surimposed by oir brain capability can 'corrupt' hypothetical stereo information carried by initial signal to be reproduced.
Not everyone choose to run mono subs: in studio it is often found to be stereo sub used, especially in mastering facility. And it become trendy in same circles to use cardio subs in nearfield for more lower stereo ( separated from room effects).
That said, for domestic room and use (and theaters) it can make sense to use a number of subs monoed to have a statistically coherent rendering over a wide covered area ( Earl Geddes approach to multi sub, Devantier and al for Theaters).
It's a rendering choice.
With thougthfulls acoustic treatments and use of FIR for room corrections you can make very nice things to correct the room.
My own experiments about xover and Schroeder freq doesn't seems to bring something revolutionary to me. Used through a global approach as E.Geddes's one it could maybe bring something more but with typical stereo approach it didn't brought me anything noticable. I hope it'll be different to you and you'll share results with us if so!
https://inst.eecs.berkeley.edu/~ee247/fa09/files07/lectures/L2_2_f09.pdf
Acknowledged.
Well, you can ask...but the milk's long been spilled, however, i.e., you should have complained 9 years ago.
The term is actually quite useful in my experience. I've spent many thousands of hours doing it (at about 25,000 CD music tracks over 9+ years). Perhaps if you had full directivity control in your three-way loudspeakers (down to the room's Schroeder frequency), you'd hear the benefits, too? I don't know. I hear them clearly, and so do others.
And perhaps your sensitivity is in being so close to the practice of introducing the problems to the produced music tracks. (I don't know, however.) By way of example:
etc. , etc. This is the tip of the tip of the iceberg, of course.
Chris
Last edited:
I agree with everything you say and my last two designs crossed the tweeter at around 1500hz the mid\midbass at 250 in one design and 100 for another. These issues have always plagued designers and DIYers. Concentrics are a great idea but there doesn't seem to be a silver bullet yet there either. Modern 1.5" CD\horn combos look like a good solution if a person can afford that.
Chris, thank you for the link. 5 or 6 pole Elliptical filter is to my knowledge never used. Might be an experimental thing, lab works, whatever, but not what was used in practice, on real life cuting lathe i've seen.
I found some reference to the Neumann modules i've seen used, those were Neumann EE66/EE70/EE77. 6db/octave ( 1pole). First two one used tapped inductors, 77 a switched RC network. Neumann had other ref like VAB84 but i do not know much about them, i would be very surprised if they were different though.
The most steep i've seen or used were ( are) 2poles hp. Not something that would render such anomaly in phase i'm sure you'll agree.
You talked about phase issue in the Schroeder freq range and the culprit being elleptical filters on many recording you treated.
I doubt about your conclusions ( but not your observations): one of the first and foremost role of premastering treatments ( audio treatments realised during a mastering) is to ( try) to counteract the issues related to acoustics of control rooms were mixing took place.
There is often issues in the acoustics in the freq range you mentioned requiring ( sometimes) heavy eq treatments to compensate. I would rather take a look at this as a culprit of what you observed rather than an hypothetical issue with 1pole hp treatments occuring on side signals.
I'm not complaining about what you do, like i already said if it please you and the style of music you listen to, that is all that matter .
I'm complaining about the use of the term, which i repeat is misleading to peoples not into what a 'mastering' is, which is a lot of people.
From previous discussion we had i remember you listen to pre 90's music mainly.
Before loudness war was raging. In those days the use of dynamic processing was mostly aesthetical related as compressors can be a creative esthetical tool when used moderatly. Dynamic range of the music ( style) you listen to enabled this, leaving headroom availlable in the signal which in turn gives possibility to eq the message 'to taste'.
And this is what you do in my view, no 'demastering' as the only way to have access to this once a dynamic treatment had been applied is to have access to raw mixes which is rare for many reasons*, but you are doing your own 'mastering' to... your taste.
The fact you mention gear ( loudspeaker) and room is another clue to me: it doesn't matter as what a premastering does is not dedicated to a kind of gear or room: it's done for an 'average'.
That's why you will see as much mastering loudspeakers systems/rooms as there is engineer: from Magnepan ( dipoles) to horns based systems to nearfield monitors... and there is as much variation in rooms too ( i've seen pristine acoustician designed rooms ( Hidley's or Jouanjean's) to some 'post production booth' with many variation in betweens).
The gear on which it is performed doesn't really matter as long as you know it and what to target for the style/genre you work on and it please your client/ customers as they are the ones which validate the work in the end.
Another clue you don't get some of the reality of the job: all genres aren't the same and some require low dynamic range. It's an aesthetic requirement.
(Extreme) Metal, dance related electronic music, Hip Hop derivative, etc,etc, all require this as without high compression they would not sound the same...
The real issue with low dynamic range is when it's applied to genre unsuited to this.... which is the issue with loudness war: it doesn't make sense to have 6db dynamic range on jazz tracks or classical.
The list you give is an example, but given your taste i doubt you would listen to Babymetal ( i don't!), Die Krupps or Sabbaton anyway.
Baby Metal could have higher dynamic range, the wall of sound of Sabbaton would not sound the same with higher dynamic applied and Die Krupps would probably not benefits of an higher dynamic range either given what industrial is. ( i'm more into industrial than heavy/power metal as a musician but work(ed) in both style as a technician as well as many others requiring different approach to dynamic range applied to them ( including classical related music).
Another thing you don't get in my view ( implying the technicians are responsible of what you point) is we work for customers which are musicians.
They have the 'final cut' on what happen and the outcome of their production: what you hear is what THEY want for the vision of their art ( at least for the one i work(ed) with). I give my point of view about what is needed and the limit i think are wise not to cross but it's their work, their vision. The outcome is what they want... it's rare i have 'carte blanche' ( i can do what i think is wise).
I know it's the case for 99% of technicians implied.
And as i'm involved in the production process i have difficulty to change things on what is given to me as final listener/end user: to give an analogy with another art, i would not change contrast or some colors on a painting because it doesn't suit my taste... i'm not sure Leonardo Da Vinci would have accepted me modyfiing 'La Joconde', as some effect he targeted could have been changed... but that is me.
That said Banksy which is an artist decided he could put a smiley face on La Joconde, giving a different sense to the painting. I liked it and what this implyed in an artistic context. As a technician i'm not asked to do such things, but if i was hired as a musician, maybe i could do things like that too... 😉
*the main one being artist are implied into what they let access to audience: mix and mastering are part of this. I've seen people spend many days on a snare drum which didn't 'fit' a track. I've seen tracks fully re-mixed ( mixed again) during mastering because artists didn't felt it was right into the perspective of an album. There is video on youtube where you see Bob Katz spend time on a breathe at the opening of a track as it sounded unatural to him ( and you can't accuse him to be part of the loudness war as he is one of the few who tryed to go against it...). All this kind of things come from experience into the field which cannot be acquired by most of the final listener/end user... i'm sure you would agree with that.
Last edited:
I can see that I "hit a nerve", if I am to believe that you are not a malignant narcissist--which I'm sure you're not, as I've indicated before in other threads.
However, the fact is that this last discussion is about 95% off-target for the subject of this thread. That's how I know you feel strongly about the subject that you wrote about above (in response to my comments--which actually were ancillary to my point). I'll try to curtail the volume of words in reply, but not ignore what you said (as so many people think is an okay approach. It isn't, BTW).
First, the "on point" topic:
Second, the off-topic stuff:
The notion that you're going to "make the music more commercially attractive" by using bad loudspeakers in too small mastering spaces to dumb down the originally recorded music from the recording studio to find an "average" sound is fundamentally flawed in my view.
I would rather leave this particular discussion alone here, as I'm sure that everyone else reading this already has a firmly established opinion on this subject.
Krivium (I would rather use your real first name here if I knew it), I do value your opinions and judgments.
I do wish that you'd consider that I did already know what you said in your reply. I do highly value hi-fi music production in most all genres, especially ones that do not require electronic or digital manipulation to create or modify. I believe that the performance of human hearing system (my hearing in particular) is descended from millennia of evolutionary adaptation that most current popular music production ignores (especially when considering most clipping and compression use today, etc.). I've found that currently used mastering practices used in most consumer quality popular music production today really do not interest my ears--only how I can typically undo what they've done--hence my use of the term "demastering" since there really isn't any other way for me to say it.
A pertinent example of "less is more" in mastering--which my ears enjoy listening to immensely:
Patricia Barber • Café Blue Un-Mastered
Regards (and sorry for the longer reply than I wanted to make it),
Chris A
However, the fact is that this last discussion is about 95% off-target for the subject of this thread. That's how I know you feel strongly about the subject that you wrote about above (in response to my comments--which actually were ancillary to my point). I'll try to curtail the volume of words in reply, but not ignore what you said (as so many people think is an okay approach. It isn't, BTW).
First, the "on point" topic:
This was my intent--to highlight that I've empirically seen problems (many times) in as-produced music tracks that are correspondingly right on top of the same frequencies that we're talking about. That was in fact the point I was discussing and was attempting to bring into the discussion--that the Schroeder transition band has been a large problem in recorded music--mostly due to the small size of the mastering rooms used, and the monitors used that have no real directivity control below ~500 Hz, and as high as ~1500 Hz. Using a mastering control room that's been completely treated 100% by absorption squares in no way makes up for this deficiency in the monitors themselves, and in fact, displaces your comment about "average", above.
Second, the off-topic stuff:
This is where you have likely missed my point. If you want to produce something that has no real "hi-fi" value (i.e., music that is not produced acoustically--ostensibly because the musicians haven't the skill--but is largely electronically or algorithmically produced via digital signal processing), that is your prerogative. There is a lot of music today that I put into this bucket--which is most "popular music," in fact, as it mostly always has been the case. (My intent is not to offend--but rather to distinguish from hi-fi music.]
The notion that you're going to "make the music more commercially attractive" by using bad loudspeakers in too small mastering spaces to dumb down the originally recorded music from the recording studio to find an "average" sound is fundamentally flawed in my view.
I would rather leave this particular discussion alone here, as I'm sure that everyone else reading this already has a firmly established opinion on this subject.
Krivium (I would rather use your real first name here if I knew it), I do value your opinions and judgments.
I do wish that you'd consider that I did already know what you said in your reply. I do highly value hi-fi music production in most all genres, especially ones that do not require electronic or digital manipulation to create or modify. I believe that the performance of human hearing system (my hearing in particular) is descended from millennia of evolutionary adaptation that most current popular music production ignores (especially when considering most clipping and compression use today, etc.). I've found that currently used mastering practices used in most consumer quality popular music production today really do not interest my ears--only how I can typically undo what they've done--hence my use of the term "demastering" since there really isn't any other way for me to say it.
A pertinent example of "less is more" in mastering--which my ears enjoy listening to immensely:
Patricia Barber • Café Blue Un-Mastered
Regards (and sorry for the longer reply than I wanted to make it),
Chris A
I think describing my system might help. It's 2.8 and is an earnest attempt at a ray generator for the highs and a wave generator for the lows.
Highs: Cylindrical radiation pattern (Beveridge 2SW-1 on the side walls)
Lows: Eight identical powered subs scattered all over the room
Correction loop: Dirac Live
So my soundfield has two really different 'shapes' or cubic geometries, one above & one below crossover.
I have a miniDSP SHD and up until 2 weeks ago I was running 2.1; all 8 subs shared a mono signal. But I've added a Flex Eight to the bass path and this will hold MSO cals for each sub individually. Should be fun but suddenly it's control loops inside of other control loops and so on.
Beveridges don't go deep, I've mapped power bandwidth vs. drive level (using white noise) and a conservative crossover frequency for headroom is 80Hz. At lower volume they make it to 60Hz but it pulls in when you drive them harder. An 80Hz stonewall is indicated here, I'm using LR48 slope.
The natural high frequency rolloff in my subs is 12dB/oct starting at about 140Hz. Listening tests have strongly told me this natural rolloff sounds great and restricting sub bandwidth any further reduces the sense of a solid central image - the sound moves away from the center of the room and retreats to the sides, if that makes sense.
Conclusion: Fine tuning woofer highpass has a profound, profound effect of the 'curtain of sound' drawn between the two speakers.
Thanks for playing along everyone.
Highs: Cylindrical radiation pattern (Beveridge 2SW-1 on the side walls)
Lows: Eight identical powered subs scattered all over the room
Correction loop: Dirac Live
So my soundfield has two really different 'shapes' or cubic geometries, one above & one below crossover.
I have a miniDSP SHD and up until 2 weeks ago I was running 2.1; all 8 subs shared a mono signal. But I've added a Flex Eight to the bass path and this will hold MSO cals for each sub individually. Should be fun but suddenly it's control loops inside of other control loops and so on.
Beveridges don't go deep, I've mapped power bandwidth vs. drive level (using white noise) and a conservative crossover frequency for headroom is 80Hz. At lower volume they make it to 60Hz but it pulls in when you drive them harder. An 80Hz stonewall is indicated here, I'm using LR48 slope.
The natural high frequency rolloff in my subs is 12dB/oct starting at about 140Hz. Listening tests have strongly told me this natural rolloff sounds great and restricting sub bandwidth any further reduces the sense of a solid central image - the sound moves away from the center of the room and retreats to the sides, if that makes sense.
Conclusion: Fine tuning woofer highpass has a profound, profound effect of the 'curtain of sound' drawn between the two speakers.
Thanks for playing along everyone.
I've found that this piece of firmware doesn't do a very good job in the Schroeder transition band (80-210 Hz) in my room with my loudspeakers. It always seems to produce a thin sounding result, and when measuring the loudspeakers again at 1 m (with temporary floor absorption to handle the floor bounce), always seems to depress this band's SPL by 1-3 dB.
I find instead that manual measurements made at 1 m with floor absorption (and checked again at the listening positions) consistently produces a superior result. YMMV.
Have you seen the effects of using IIR filters of such high order on your phase response (i.e., the higher frequency drivers/phase plug extensions lead the lower frequency subs by two full wavelengths)? This will significantly and negatively affect the perception of bass. I recommend using first order filters--or no crossover filters at all crossing from the mains to the subwoofers. (There is precedent for this.)
Chris
The fellow who taught me more about the perception of sound than any other (yes, it took me years to fully comprehend what I was told, sadly), Clark Johnsen, thought that Dunlavy designed the best loudspeakers he had ever heard. They were what he used and he sold them when he had his shop.
John Atkinson continues to explore the links between measurement and perception. We are lucky to have him.
As always, a post from Chris A. is required reading for the serious DIYer. And even superfluous DIYers like myself ...
John Atkinson continues to explore the links between measurement and perception. We are lucky to have him.
As always, a post from Chris A. is required reading for the serious DIYer. And even superfluous DIYers like myself ...
Nothing superfluous about you, Mr. McInnis. How's the sound quality nowadays?
Chris
Here's what the low end of a Beveridge ESL panel looks like. This is a 2SW cabinet, an original Model 2 should tune different.
This measurement was taken loud, a few dB above my usual listening level. The drive signal is white noise known flat down to 10Hz. A 2KHz stone wall has been applied to this drive signal to be merciful to my electrostatics. The microphone is a 1 inch calibrated by CSL, it also is known flat. Mic was nearfield up against the lens foam. When I drop drive level 10dB the peak shifts left to about 60Hz but the shape hardly changes.
The problem with a gradual crossover slope is it forces me to throw away some of this energy. It can not turn my steep rolloff into a gradual rolloff. It can only mask my steep rolloff by overlaying a gradual rolloff that is tuned much higher, and the region between those two curves is "ray generator magic" right a critical region of the spectrum and I freakin' want it.
So I 100% take your point but I think some of these exotics are so shy of power bandwidth & SPL that the real world gets in the way. Does that make sense?
I read the article, thanks for posting it.
This measurement was taken loud, a few dB above my usual listening level. The drive signal is white noise known flat down to 10Hz. A 2KHz stone wall has been applied to this drive signal to be merciful to my electrostatics. The microphone is a 1 inch calibrated by CSL, it also is known flat. Mic was nearfield up against the lens foam. When I drop drive level 10dB the peak shifts left to about 60Hz but the shape hardly changes.
The problem with a gradual crossover slope is it forces me to throw away some of this energy. It can not turn my steep rolloff into a gradual rolloff. It can only mask my steep rolloff by overlaying a gradual rolloff that is tuned much higher, and the region between those two curves is "ray generator magic" right a critical region of the spectrum and I freakin' want it.
So I 100% take your point but I think some of these exotics are so shy of power bandwidth & SPL that the real world gets in the way. Does that make sense?
I read the article, thanks for posting it.