April 2015 Toole video on sound reproduction
- Thread starter jkeny
- Start date
The way I understood what you posted was that accurate = natural sounding. I dis agreed with this because we are starting with a flawed capture of the audio event so that accurate just means that we retain an accurate transcription of this flawed capture to the speakers.
I was also trying to use the vinyl amplitude modulation as an example of how something that's not accurate may actual be more realistic because it incorporates some characteristics of real world sounds in it's flawed reproduction i.e the fact that there is almost no sound in nature that is a pure unwavering tone. The sounds of the natural world are what our auditory perception uses as it's learning canvas & incorporates into it's perceptual model. So given two tones, one pure & one fluctuating, I believe we will find the one that fluctuates as more natural simply because it ticks one of our perceptual learnings of the real world.
I'm talking about auditory perception - the conscious awareness of audio, not about the physiological functioning of the eardrum, etc. And this perception operates in the same way in all of us - hence we all come to a common understanding of the audio world. We all produce an internal auditory scene out of what we hear. In the same way as we all produce a visible scene out of what we see. Now what is different is what aspect of that scene we focus on at any point in time. But if I say listen to that cymbal, your attention will hear that cymbal even though you may have been listening to the saxaphone up to that point. So, it's not that we perceive things differently - all the auditory objects in an auditory scene are pretty much the same between us - it's rather that we can only focus on a small number of objects with awareness at any one time.
Sure - it's a hot area of research, I think?
Yep, it's a bigger, more general issue than analogue Vs digital - it's about understanding what model best maps our auditroy perception. Toole's/Olive/Harmon's work on speakers gives one small aspect of that - concerning the spectral difference between direct & reflected sound. This is uncovering a rule that is embedded in our auditory perception due to our subconscious learnings from how sound reflects in the natural environment.
So when Harmon talk about this preference, they are establishing a part of the general model of what makes our auditory perception believe that something is natural sounding. It's not a matter of "taste" - it's deeper than that - it's an underlying precept that applies to our auditory perception
You are right that I equated accurate to natural but I was including the entire recording/reproducing process in my consideration.
I have heard that argument before and, while logical, it leads us away from the goal of achieving greater success.
But that is simply not true. Our conscious view of events varies as the raw (transduced) information is interpreted and formed into a percept based on personal history, mood and context.
Ha! Exactly, that is how our individual history and current state of awareness dictate a unique perception of events completely without the intervention of an external influence.
To be sure.
And I can teach someone in 10 minutes to setup an arm and cartridge, there really is so little to it. Different skill sets for different needs and/or interests.
Sound pressure is very different, even large speakers don't have real bass below 40hz, the rest is room mode and really pressurized, isolation of lower frequencies becomes more problematic with subs. I can limit and tailor the upper end of the sub but can't do much in the lowest frequencies. True 50hz bass performance is more than acceptable in most of installations. Rarely do I have the need or the possibility to supplement lower frequencies for the main speakers and when I do, its very system specific and custom. That's when I order custom made electronic crossovers matching the subs' characteristics and basically equalize the low frequency to blend in correctly. Yes, that's straight forward and relatively simple when you know what you're doing. 95% of the time I have to take a top to bottom approach to room tuning, 16k-50hz where you face many additional challenges that any type of EQ will do little for.
Agreed and I do both including constructional acoustics. Everything starts with optimal positioning within allowable parameters of the domestic environment. Tuning starts from there. Speakers are selected according to that need, for example I can't use a rear firing speaker if I'm not allowed to pull them out far enough into the room. Everything has to follow the space and the budget. When the budget is there the analog source get's EQ'd with the best arm/cartridge/SUT combo and foundation, digital gets a Weiss EQ-1 and the speakers get bi or tri-amp'd eq'd with a custom made tube electronic crossover. Additional money buys them the services of a top acoustician, space & money gets them everything in a purpose built listening space! What Mike did, lucky guy!
On the other hand if they want HT or aren't willing to accept the bare minimum needed for good sound, and I don't mean just money, I direct them online for a Denon receiver and some decent speakers. Plunk it down where ever and let Audyssey do its magic, imo much better sounding than any half backed high system. By the way I'm a hardcore Denon fan! Beyond that, for high end HT I send them to people like you who have the speciality and tools to deliver everything on their menu including all that automation the wife MUST have. And for two channel I seriously try to convince them to stay with Denon until they're ready and able to accept the minimum, but that never works
.see above,
david
Nothing is accurate , it cannot be.. the reference you want for how it sounds is what the mastering engineer heard, you aren't there .. so you dont know.. the room mangles the speaker output and NO ROOM IS PERFECT.. so you are already in the realms of inaccuracy..
And even if you DO use DSP , its all taste based anyway , there is no universal truth in terms of a listening position curve.
The upshot is to get the system sounding the way the listener wants it to.. and you can do this many ways, passive treatments , distributed bass right up to full freq correction that does freq and time...
There is no need for arrogance , no position here is superior.
All I can tell you , what with a treated room , a few subs scattered here and there and DIRAC..I and others visiting me , have a more believable illusion of being there or the performers being in your room..it's all an illusion.. shape it whatever way you want.
And even if you DO use DSP , its all taste based anyway , there is no universal truth in terms of a listening position curve.
The upshot is to get the system sounding the way the listener wants it to.. and you can do this many ways, passive treatments , distributed bass right up to full freq correction that does freq and time...
There is no need for arrogance , no position here is superior.
All I can tell you , what with a treated room , a few subs scattered here and there and DIRAC..I and others visiting me , have a more believable illusion of being there or the performers being in your room..it's all an illusion.. shape it whatever way you want.
I think Toole has done as much good for audio reproduction as anyone in years, but his recognition of the "circle of confusion" could be his greatest contribution, if the recording industry would only adapt the obvious solution - universal, consistent monitoring standards. It wouldn't keep artists and producers from creating whatever they like, but it would give us a benchmark, so we could know when, within the limits of our systems, we were listening to the artists' intentions. I suspect it would do nothing to stifle these discussions, though. For many, the pursuit of something more natural, more representative of the original performance is their hobby, or at least a big part of it. They won't be letting go of the dream. In some cases, I look at pictures of their systems and their rooms and I know that even with the best live recordings it can't possibly sound "natural" in there. But they are heavily invested in believing that it does, so that's what they hear.
Tim
Can the pro DSP guys give some further feedback on DSPeak Anti Mode 2.0 Dual Core? I'm as 2ch analog non dsp as they come, w/a real computerphobic attitude to life. A curse I know, and for me e.g. downloading/streaming etc is not of interest. And here we have a complete system upgrade which is in effect the final frontier. I don't want to unecc miss out on the party.
Hence this unit appeals to me. Some qs
Is it as close to plug and play as you can get i.e. install it, set up the mic at listening position, let it do various test tone sweeps, and it comes up w/the analysis, and applies the solution w/nothing more than me than pushing the button that says in effect "agree/operate"?
Is it ok in the l/t to just use dsp solutions like this just for subs that i may add, and deal w/the frequencies higher up from the main spkrs by room treatments?
Is the tech/quality commensurate w/more advanced units like Trinnov, Dirac, Deqx, Anthem, Deqx, Illusonic, esp at this fairly restricted area of use i.e. just dap to sub frequencies/not wanting different curves for different genres of music?
If I can get a "DSP For Dummies" solution, then I'm game.
Hence this unit appeals to me. Some qs
Is it as close to plug and play as you can get i.e. install it, set up the mic at listening position, let it do various test tone sweeps, and it comes up w/the analysis, and applies the solution w/nothing more than me than pushing the button that says in effect "agree/operate"?
Is it ok in the l/t to just use dsp solutions like this just for subs that i may add, and deal w/the frequencies higher up from the main spkrs by room treatments?
Is the tech/quality commensurate w/more advanced units like Trinnov, Dirac, Deqx, Anthem, Deqx, Illusonic, esp at this fairly restricted area of use i.e. just dap to sub frequencies/not wanting different curves for different genres of music?
If I can get a "DSP For Dummies" solution, then I'm game.
I owned and used the antimode 2.0 so can help with it
It is self contained and doesnt require a degree in rocket science and will sort out your bass big time.
There are some things you have to know
optical input only if you dont want to go thru an AD
24/48 only
You HAVE to apply a bass house curve (hump) as flat bass sounds lousy and any parametric changes cannot be heard on the fly IE you still have to tune the bass and the rest to taste
It is also more expensive than a lot of other solutions.
If you feed it analog , it will do an analog to digital conversion , do the correction and then do a DA..so its best used as a pre and dac...
I would rather take the time to learn a little and use a miniDSP unit. A lot more flexible.
antimode make a cheaper unit just for subs antimode 8033
It is self contained and doesnt require a degree in rocket science and will sort out your bass big time.
There are some things you have to know
optical input only if you dont want to go thru an AD
24/48 only
You HAVE to apply a bass house curve (hump) as flat bass sounds lousy and any parametric changes cannot be heard on the fly IE you still have to tune the bass and the rest to taste
It is also more expensive than a lot of other solutions.
If you feed it analog , it will do an analog to digital conversion , do the correction and then do a DA..so its best used as a pre and dac...
I would rather take the time to learn a little and use a miniDSP unit. A lot more flexible.
antimode make a cheaper unit just for subs antimode 8033
http://www.stereophile.com/content/music-round-57
Yes.
Yes.
I am not sure what you mean by tech/quality but, for subs, I'd venture another yes.
Not sure why you want to get into DSP so hesitantly, i.e. for subs only. If you are going to try it, might as well get one that lets you use proper DSP and save different curves and play around. If you don't like it, you can use it only for subs.
DSP for subs is not far from entirely transparent and I am measuring my words here. It is not Plug & Play. Nothing is. Some are easier than others. Fully automatic results are rarely optimal but they can constitute a starting point. A minimum of measurements is always required. DSP does work though and well.
Once you get over say 500 Hz (arbitrary number I would have preferred 200 HZ), the transparency tend to become more problematic, especially if one has an aversion for things digital. Such strong emotions color our perception.One may refute it ,debate or try to fish arguments against it: Our bias color our perception.
One more thing , we audiophiles have become educated to listening carefully. We need at times to renew this learning process. I tend to believe that what some conceive as burn-in or whatever is an adaptation period during which we have to unliad our previous notions. This process is not conscious and does nott require that we constantly listen to the gear.. Simply that our vested interest in the gear or components allow us to push away previous standards and habits. Once the bass is cleared the rest of the spectrum become clearer even the treble... Small detail up to then unheard become startlingly prominent... This is what reducing 10 and up ( 10 dB is routine, 30 dB peaks and dips are not uncommon inmos dB peaks and valleys most rooms impose to the bass will do to listener perceptions. Give it a try . You may like it .. I am willing to bet many would be surprised by what it brings to the equation, especially those with problematic rooms that cannot afford the logistics of passive room treatments ...Madfloyd came to mind while writing this.
I see what you mean but I don't think we are even close to anything that will capture enough detail at the recording process for it to be possible to reproduce the actual auditory event in our listening rooms. So we are faced with something which is inherently < 100% accurate at the recording stage & the job thereafter being to make the most believeable illusion in our living rooms
I can see your point but disagree - my thinking is that knowing the rules by which our auditory perception operates will allow us to check all the necessary boxes for making the best illusion. It will also inform us what might be missing at the recording end & perhaps addressing this shortcoming for an even better, but not fully accurate, auditory illusion.
Indeed, I agree that it is difficult (maybe impossible?) to strip the higher cognitive aspect of our interpretation of the scene from the scene itself but I think these are two different processes - the scene itself is the same for all of us but what aspect of it we focus on & what our thoughts & feelings are will be different for each of us.
It's like looking at a painting - we can look at different aspects of it & vary our focus from looking at the whole & it's affect on us to looking at a particular aspect of the whole, the use of light & shadow, the treatment of the paint on the canvas itself, the brush strokes, etc. but all those aspects are available to everybody to see if they know what to look for. The emotional affect of the painting will likely be an individual thing, however & in our description of the painting this is what we normally talk about.
Yes but "perception of events" is misleading - I would say the "persona interpretation of events", maybe?
Anyway, we are both trying to get to the same place, I feel.
I agree with you and I prefer full range EQ myself. But, Toole does not, and EQ above the transition frequency is more controversial. Amir has tried to focus the EQ discussion in this thread just on bass below the transition frequency, where EQ can easily demonstrate tangible corrections to obvious measured room modes.
Then there was the analog vs. digital red herring. Since analog EQ is effectively dead in the marketplace, I suggested the hybrid approach using a DSP EQed sub channel with minimal disturbance to the analog main channels above the xover, avoiding a-d + d-a of the full range signal by using an analog xover. As Kal pointed out, though, the DSPeaker Anti Mode device I cited does use a digital xover, contrary to Robert E. Greene's review which I cited. I trust Kal on this rather than REG. So, it does not avoid a-d + d-a on the full range signal.
It now seems that this hybrid approach and trying to preserve unconverted analog signal above the xover would require going through many hoops. I doubt seriously that any device containing DSP EQ, whether stand alone boxes or subwoofers, would not also have a digital xover, since that is much more simply executed in the DSP rather than in a separate analog crossover circuit.
The only remaining hope for an analog xover + DSP EQ'ed sub channel I can see would be with one of the extremely rare remaining subwoofers containing no internal electronics or crossover. The only ones I have encountered are the Wilsons, which use a separate analog crossover box and which have no internal amps. A DSP EQ device could be inserted before the amps on the subwoofer output side after the xover. Hence, the main outputs would use the analog xover and be "unsullied" by any conversions to digital.
Agreed, that is a lot to go through and Wilson subs and crossovers are not cheap. But, it is a way for analog true believers to stay mostly in analog but also enjoy the benefits of DSP EQ in the deep bass. However, they should also weigh the simpler alternative of just using a typical DSP EQ box embodying a digital crossover, as cited by Kal, to see if the a-d + d-a process is sufficiently transparent for their needs. Again, the plain fact of life is that EQ technology these days is virtually all done digitally in DSP, sub xovers too. That is highly unlikely to change, ever.
Just a clarification here: Dr. Toole is not opposed to full range EQ. He is opposed to thinking that if you have a speaker with off-axis response that is different on-axis, can be remedied with EQ. The sum of those two signals will tend to show frequency response variations that should not be fixed with EQ. But rather, a good speaker. In his words, "if you get a good speaker, then you can EQ it."
Likewise, I am in favor of full band EQ but only if you can selectively turn EQ on and off in any specific position in the frequency response to see if the subjective result is positive or negative. I am not in favor of all or nothing for the entire range which unfortunately is the norm in mass market AVRs.
You are right that for the purposes of this thread, I have focused 100% on low frequencies for the reason you mention. There is no room for debate there as there are no psychoacoustics to worry about and what you see is what you get. If the response is way off, correction almost always makes it better. The only cost can be reduced bass response which can be fixed with a broad correction of the level (sloping down from low to high frequencies).
Amir - thanks for the clarification of Toole's position on full range EQ. That was not covered in the original video, so I just assumed. But, yes, even so, keeping focus here on just the bass issues makes a lot of sense. We can discuss full range EQ another day in another thread.
Great starter set on room measurement, Amir. I look forward to more on measurement and EQ.
In light of some earlier points in this thread, I think some of the key choices one must make in selecting a DSP EQ tool concern the user interface and degree of automation of the measurement/analysis/filter calculation/result presentation process. It is good to know the basic elements of the room correction process via EQ. But, especially for first time users, I like the class of tools that automate the process from beginning to end into kind of an "expert system", as we used to call them back in the 80's in computers/IT.
You and experienced users know this, of course, but other neophytes might not. These automated tools start with a target frequency response curve (usually a default, but often selectable or modifiable). They then step the user through multipoint mike measurements, providing the test tones and guiding mike position setup. After the test tone sweep, they automatically apply spatial averaging to the mike responses, and automatically calculate a complete set of filters, where the frequency range for the filters can sometimes be adjusted or limited by the user. They often perform other useful calculations automatically in the process such as channel-by-channel speaker delay and level trims for multichannel setups.
I find that the PC implementation of Dirac Live is very high on the ease of learning/ease of use scale, yet it also provides more advanced capabilities such as target curve modification and frequency range limits, if desired. It also allows for up to 4 different resulting filter sets to be selected quickly on the fly for comparative playback audition. The resulting graphs of before/after frequency and impulse response aren't too bad either. It takes me about 20 minutes start to finish to do a complete calibration starting with the mike sweep, that on a 7.1 channel system using 8 mike positions. A stereo calibration would take much less time than that.
Arguably, different tools might have even more sophisticated technical features and might even do a better job, likely at the expense of ease of learning/use. And, a step-by-step manual process might offer a greater sense of control, like driving a stick instead of an automatic car. But, here with EQ, doing it manually means requiring much more knowledge and time on the part of the user. Another analogy might be it is useful to know how to do long division by hand even if you always just punch a few buttons on a calculator.
One constant criticism of the automatic method is the inability to duplicate resulting response determined by the EQ tool via independent measurements, e.g., using REW. That is, however, an issue resulting from things like the inability to duplicate mike placements exactly in calibration vs. subsequent measurement. Also, many aspects of the algorithms used by the tool may be proprietary or unknown, such as multipoint averaging, time windowing, 1/x frequency smoothing, etc. So, independent followup measurement gets trickier with automated tools, unlike the more open and more deliberate completely manual calibration idea.
That is a good point. There are two approaches to this topic:
1. Manual equalization using a parametric equalizer. This is a bank of programmable filters. For each filter, you can set the frequency, how much it filters, and how wide its response is (so called "Q"). This is what makes it "parametric." For bass performance, the filter must have excellent design, being able to program it down to a single Hertz or even smaller frequency resolution. As I noted, this used to be expensive to implement but not so much these days.
In this context as I have mentioned, my focus has been in bass frequencies below 200 Hz or so. While the parametric eq (PEQ) can be used elsewhere, its use is not straightforward. And even below 200 Hz, its use across many channels can get quite tedious. So its best use is to use it for inserting 2 or 3 filters and in less than 100 Hz and call it done. This can make a remarkable difference which is good news as many subs come with 2 or 3 PEQs.
2. Automated EQ. I call these Auto EQ so when I don't include the word "auto," I always mean the manual process above. Audyssey is the mother of all such systems as far as distribution since it lives in many brands of Audio/Video Receivers (AVRs). Unless you get the Pro Kit which comes with a professional mic and software to control it, it is an all or nothing deal which I don't like.
Mistakenly all of these systems are called "Room EQ" but there is no such thing. The room is the room and it is not changed by electronic means. This is more of a terminology thing than anything else but best to use the correct term: auto eq.
The best of such systems give you the ability to tailor the target curve as you explained (and I dig way deep into here: http://www.madronadigital.com/Library/Room Equalization/Room Equalization.html), so that you can tailor what the overall response should be (not the filtering for correction). Flat response is not one that is preferred so you need a sloped one and that slope should be programmable and adjusted by ear.
The other critical item is to be able to defeat, each individual correction point. If the Auto EQ lifts a mid-range sag and that is there due to the speaker flaw, that correction may or may not sound better. You want to turn that on and off, and preferably blind. Close your eye and rapidly turn that part of the filter on and off and then test carefully and open your eyes when you think it sounds better. If it is when the filter is on, then leave it on. Otherwise, leave it off.
Alas, that is a feature that is hard to find in many Auto EQ systems and is the reason all high-end acoustic designers use parametric EQ. In my case, I use JBL Synthesis ARCOS which gives both abilities. It auto generates the filters but then lets you turn each filter on and off.
Many Auto EQ systems in contrast, create one FIR filter transform for the entire response and not a list of filter parameters that ARCOS generates. So there are no distinct filters to turn on and off. You can take their graphs and modify them but that transformation will always be there to some extent. Not ideal but heck of a lot better than not having the ability to change things.
Note that there is a half-way plan. REW program that I wrote a tutorial on, can auto generate filters for all the common low-cost DSPs. It can even program some of them. I find that good and bad. Good in that the filter settings it generates create a flatter response than I can do manually. The bad is that in my limited testing, while the curve is flatter than what I can do, the sound is not as good. I don't have a good explanation for this yet other than the large number of overlapping filters it generates. Less may be more there but this is a hand-waving explanation based on very limited testing.
1. Manual equalization using a parametric equalizer. This is a bank of programmable filters. For each filter, you can set the frequency, how much it filters, and how wide its response is (so called "Q"). This is what makes it "parametric." For bass performance, the filter must have excellent design, being able to program it down to a single Hertz or even smaller frequency resolution. As I noted, this used to be expensive to implement but not so much these days.
In this context as I have mentioned, my focus has been in bass frequencies below 200 Hz or so. While the parametric eq (PEQ) can be used elsewhere, its use is not straightforward. And even below 200 Hz, its use across many channels can get quite tedious. So its best use is to use it for inserting 2 or 3 filters and in less than 100 Hz and call it done. This can make a remarkable difference which is good news as many subs come with 2 or 3 PEQs.
2. Automated EQ. I call these Auto EQ so when I don't include the word "auto," I always mean the manual process above. Audyssey is the mother of all such systems as far as distribution since it lives in many brands of Audio/Video Receivers (AVRs). Unless you get the Pro Kit which comes with a professional mic and software to control it, it is an all or nothing deal which I don't like.
Mistakenly all of these systems are called "Room EQ" but there is no such thing. The room is the room and it is not changed by electronic means. This is more of a terminology thing than anything else but best to use the correct term: auto eq.
The best of such systems give you the ability to tailor the target curve as you explained (and I dig way deep into here: http://www.madronadigital.com/Library/Room Equalization/Room Equalization.html), so that you can tailor what the overall response should be (not the filtering for correction). Flat response is not one that is preferred so you need a sloped one and that slope should be programmable and adjusted by ear.
The other critical item is to be able to defeat, each individual correction point. If the Auto EQ lifts a mid-range sag and that is there due to the speaker flaw, that correction may or may not sound better. You want to turn that on and off, and preferably blind. Close your eye and rapidly turn that part of the filter on and off and then test carefully and open your eyes when you think it sounds better. If it is when the filter is on, then leave it on. Otherwise, leave it off.
Alas, that is a feature that is hard to find in many Auto EQ systems and is the reason all high-end acoustic designers use parametric EQ. In my case, I use JBL Synthesis ARCOS which gives both abilities. It auto generates the filters but then lets you turn each filter on and off.
Many Auto EQ systems in contrast, create one FIR filter transform for the entire response and not a list of filter parameters that ARCOS generates. So there are no distinct filters to turn on and off. You can take their graphs and modify them but that transformation will always be there to some extent. Not ideal but heck of a lot better than not having the ability to change things.
Note that there is a half-way plan. REW program that I wrote a tutorial on, can auto generate filters for all the common low-cost DSPs. It can even program some of them. I find that good and bad. Good in that the filter settings it generates create a flatter response than I can do manually. The bad is that in my limited testing, while the curve is flatter than what I can do, the sound is not as good. I don't have a good explanation for this yet other than the large number of overlapping filters it generates. Less may be more there but this is a hand-waving explanation based on very limited testing.
Similar threads
