The tweaking imperative

CopperTop

Member
2009-02-28 9:40 am
UK
I'm not just referring to New Age tweaks, but whether there is an absolute need for audiophiles to make major changes to their system on a frequent basis.

My experience is that I assemble a dream system and it sounds so good that I simply want to spend as much time as possible listening to it. I've done it! Maybe now I can simply leave it be and listen to the music. This happy state lasts for maybe a month or six weeks but then it gradually palls, eventually reaching the point where I no longer can remember what I liked about it in the first place. It becomes flat, two dimensional and more of a chore to listen to it than a pleasure. Even leaving a gap of a week or two doesn't help by then - so it's not listening fatigue as such, and there's nothing obviously grating about the sound.

The natural reaction is to want to make it better - obviously I have grown used to that level of quality, and I must therefore improve it. It will probably involve the expending of some money, time and energy.

I wonder if this cycle is inevitable. Some DIYAudio-ers have suggested that if you tire of a system, it must be because it has some flaws and that once you discover the perfect synergy between source, amp, speakers and room you will never look back. (The strange thing is that they never seem to achieve this state of contentedness themselves, and seem as keen on ceaseless tweaking as the rest of us.) I'm not convinced: I am beginning to wonder whether it is simply impossible to listen to a stereo system - even a perfect one - in the same room ad infinitum. In fact, perfection will forever remain a mirage you can never reach. I suggest that it is like going to a concert every night in the same hall and sitting in exactly the same seat, so that no matter what the performance, you are endlessly hearing the same acoustics superimposed over it. After a while, your ears are incapable of being surprised by what they're hearing, and surprise is one of the main appeals of music (to me anyway).

Moving to DSP-based active, I have found that it, too, isn't immune from the problem: there isn't a single perfect setting that doesn't pall after a few weeks. However, I find that I can then obtain the necessary aural refreshment by relatively small tweaks to the crossovers. There's no overtly obvious change (as indeed there shouldn't be), but the element of surprise is restored; the system suddenly regains its colour. I presume that changing the crossover of any multi-way speaker in fact performs quite a complex modification of its interaction with the room that is different from simply applying EQ to the signal, say.

I now intend to include a 'randomize' option to the software, so that every time the program is started, it applies limited random variations to crossover frequencies, slopes, phase correction and so on. I see this, potentially, as another on the list of advantages of active over passive: get a new pair of speakers every time you listen to them.
 
No, I don't believe it's inevitable. Decades ago I became aware that audio can achieve a very high state of quality, where the speakers in every sense become "invisible" acoustically. When this is arrived at the pleasure in the sound is immense, I've never tired of, or lost interest in what I was hearing when the system was working at this level.

This is difficult to achieve though, requiring extreme focus and fastidiousness in tweaking to make happen. If you don't know what I'm talking about, then you haven't got there yet; and when you do, all that restlessness will be put behind you ... ;)
 

CopperTop

Member
2009-02-28 9:40 am
UK
If you don't know what I'm talking about, then you haven't got there yet; and when you do, all that restlessness will be put behind you ... ;)

Hi fas42

Yes, I have been told something like that before, but as I say, I am doubtful that a real system in a real room can ever be 'perfect'. Each driver is putting out a wavefront that strikes the walls and is partially absorbed or reflected in complex, but endlessly repetitive ways that you, from your listening position, must ultimately become bored with over a period of time (I suggest). Maybe an occasional tweak of the speakers' orientation or elevation might alleviate this. Maybe a move around of the furniture may help.

But how many people in these forums ever leave their equipment for weeks at a time without tweaking in some form or another? Very few I would guess! If anyone had reached the audio nirvana you suggest, why would they be hanging around here anyway..?!
 
But how many people in these forums ever leave their equipment for weeks at a time without tweaking in some form or another? Very few I would guess! If anyone had reached the audio nirvana you suggest, why would they be hanging around here anyway..?!
Well, in my case because I'm working with fairly cheap gear which is rather fragile -- so part of the exercise is refurbishing to keep it together. Also, no-one knows all the answers as to what needs to be done to get optimum performance, so there is a decent component of straight out investigating, to try and understand what counts, and what doesn't.

'Perfect' is not a word to be used, because the recordings aren't "perfect". But they can be made to reproduce in an immensely satisfying way, and that's good enough for me! Reflections in your room don't matter at this level of playback, they become irrelevant. Of course, if one happens to be a person who would get bored with top notch musicians playing live in their listening room, then I don't think there's much hope for them - what are they doing listening to music at all? :)

Cheers ...
 

SY

diyAudio Moderator Emeritus
2002-10-24 10:19 pm
Chicagoland
www.SYclotron.com
I am doubtful that a real system in a real room can ever be 'perfect'.

If you wave your hands and keep repeating the same meaningless mantras, sure it can.

I'll confess that I rarely "tweak" or change anything once a system is set up and running properly. My hifi is for playing music, not for exercising some imagined creativity- I prefer the creativity of the musicians whose recordings I own to the storytelling of audiophiles. :D

The exceptions are when I have to change out a component to accommodate a new source (e.g, my soon-to-be-installed MM preamp). And room acoustics, which is a constant challenge when you move as often as I've been forced to do. Every once in a while, I'll also throw in some sort of test box if I get curious about someone's claims about an electronic device (for example, my Bastard Box), but I much prefer listening to music than listening to hifi.
 
My equipment stays unchanged for years, as I think about what to build next - I am a slow designer and builder!

Exception is repairs: two new EL34 (one developed intermittent red-plating so I changed the pair), and two new anode resistors (I had underspecified them originally and they were only lasting a few months).

I could not conceive of anything worse than tweaking every few weeks or months. This would ruin music for me.
 

Cassiel

Disabled Account
2004-09-30 3:53 pm
Madrid
I thought once that I could build a great sounding amp anytime I wanted to, so, for the sake of experimenting, I have killed many great amps for parts. Recently, I ended up with bad sound and couldn't get it right. I was trying to like SS stuff. Unbelievable - a lesson to learn. I have lived with good sound so many years, it has become a natural thing and then one day it was gone. DISTRESS. I just built a 6AS7 PP amp in a hurry because I felt like a JUNKIE. This one is not my best work but sounds GOOD and my nerves are OK now. It will do for the moment. No more SS for me - lesson learned. Oh and always keep at least one tube amp alive - another lesson learned.
 

Jay

Banned
2003-02-11 9:02 am
Jakarta
This is difficult to achieve though, requiring extreme focus and fastidiousness in tweaking to make happen. If you don't know what I'm talking about, then you haven't got there yet; and when you do, all that restlessness will be put behind you ... ;)

What tweaking??:D

Too much tweaking might be a sign of cluelessness :D

A good system starts with a good speaker. A good speaker starts with a good driver. First, find drivers that can be filtered with second order electrical without notch. Second, know that driver sound is mostly the sound of the cone material (surprise!). So pick your cone material properly. Everything else is just simple science.
 

CopperTop

Member
2009-02-28 9:40 am
UK
The thing I have in mind is that if you think about real speakers in a real room, conceptually it is no different from feeding your music through a reverberation effects unit on one setting, forever. Simplifying it, when that drumstick hits the drum, a delayed echo then hits your ear 16.34 ms later, and it happens every single time in the same way, over and over again. You may play a wide variety of music, but it all gets passed through that effects unit on the same setting. You may move around the room a bit, but on average the effect remains the same. It's OK at first, as the combination of reflections, and the complexity of the music keeps your ears interested, but eventually surely your ears are going to become jaded. It doesn't mean that the speakers or room acoustics are bad, merely that you are listening in a real room and not an anechoic chamber.
 
What tweaking??:D

Too much tweaking might be a sign of cluelessness :D
I have a strange abberation, I want the music that I like to sound good; and it was very apparent in the earlier years of hifi that it works in reverse on many system: only "correct" recordings work properly, and that's no use to me. I want to put on a rough 'n' ready rock and roller, a Faces album is a good example of that, and it should blow me away. Not because it's loud or distorted or grungy, but because it sounds as "good" as a pristine, audiophile recording. Standard systems are a million miles away of being able to do this, you can tell it every time because they make a mess of the drumkit sound, cymbals only very roughly resemble anything that a real drummer ever hears.

So, the point of tweaking is to get, say, the drumming on a "crude" 70's rock recording to have the impact and finesse of an uber recorded audiphile effort ...
 
Simplifying it, when that drumstick hits the drum, a delayed echo then hits your ear 16.34 ms later, and it happens every single time in the same way, over and over again. You may play a wide variety of music, but it all gets passed through that effects unit on the same setting. You may move around the room a bit, but on average the effect remains the same. It's OK at first, as the combination of reflections, and the complexity of the music keeps your ears interested, but eventually surely your ears are going to become jaded. It doesn't mean that the speakers or room acoustics are bad, merely that you are listening in a real room and not an anechoic chamber.
Obviously you haven't been in a small space with a drummer letting rip on a real drumkit: if you think this would sound ho-hum after a bit then there's no hope for you ... :D
 
CopperTop,

If I had a continuum of musicians playing in my living room, I would not being paying much attention to the living room.


I see you have spent some time now with a convolution engine, and have explored manual phase linearization. Some posts suggest this has brought you further glimmers into hi-fi. What have you used for speakers with this?

What is your measurement microphone?

Does your listening environment allow extending listening at realistic levels for different types of music?
 
CopperTop,
If I had a continuum of musicians playing in my living room, I would not being paying much attention to the living room.
But if you had them playing in there daily for six weeks? I suspect that you would crave a different acoustic after a while.

What have you used for speakers with this?
Initially Mission 702e converted to 'active', then Tannoy R2, and now my own three way active speakers based on 12" driver in 90 litre sealed cabinet, plus 4" mid and 1" tweeter in separate, smaller enclosure. I don't want to brag, but they do sound magnificent.

What is your measurement microphone?
WM61A capsule housed at the end of a (mechanically damped) aluminium tube, plus my own op amp-based pre-amp.

Does your listening environment allow extending listening at realistic levels for different types of music?

Yes. I'm lucky that I can turn up the volume as high as I like.

I think my main point is that a common thread in many posts is an assumption that a form of perfection can be found if a few variables can be tweaked correctly. Following each tweak people may think they've hit perfection only to be disappointed after a while. They may falsely attribute that disappointment to needing even better quality than before, or maybe they think the equipment's gone wrong, or that they burned the cable in with too much rock music and now it's ruined for classical. I, too, feel that disappointment at times, and I'm looking for the reason why. Maybe it's a placebo effect, but my experience with active speakers is that a bit of random variation in the crossover characteristics can keep the sound fresh, and so I'm homing in on the never-changing nature of the speaker/room acoustics as the possible culprit. What do you think to the analogy with feeding your music through a reverberation effects box on a single setting, forever?
 
Copper, I feel you.

we're dealing with a few issues here.
first, IMO, most audiophiles are not actually music lovers. one needs to take a look at the music collection to check that. any Chesky albums? ok, true audiophile, knows jack **** about music. ok, I'm exaggerating, they have released some decent albums. but you're getting the idea.
guy comes to my place to listen to my new speakers, says "play some soft jazz with a female voice". audiophile alert! ok, girl with a guitar is not a bad thing in itself but it's a "red flag". LOL.
I actually know very few music loving audiophiles.

second. the continual system upgrade. it gives them something to do. audiophiles are people who need to feel special. if audiophile friend just tested a new vibration absorbing product and seems extatic about it, unsecure audiophile needs to follow suit.

IMO, there's a definite threshold for "listenability". once that threshold is reached, people should just stop, listen to music and enjoy life. at this point, it should be noted that life doesn't equal audiophilia. when funds allow and the improvements gained for the money are OBVIOUS, go for it.

disclaimer: the above is referring to the "non music loving" subcategory only. don't tell me there aren't such animals, I know better.
 
If room is too small/too live it masks low level reverberant details in recording. This is most apparent with acoustic music and trying to capture live space. Pop music is mostly packaged for casual listening; cars, headphones, crappy mini stereos. In these situations artificial reverberation is often engineered in; typically at levels much higher than encountered in natural spaces. Likewise dynamic compression when pushed too far leads to effect of gnawing sameness. You turn overly compressed recording up to get peaks at live level, and none of the reverberant details work out right in the mind.

With high definition system this is price of admission.

If listening space is too live, then indeed it asserts a sameness on all sources. Details lost, or given contrast are specific to room/speaker behavior.

Great recording/engineering stands out in capacity to deliver music experience across many listening situations. Great recordings intended for controlled listening are stellar. These are recording where new details emerge over multiple listening sessions, and where new details emerge when played through system capable of revealing these details.

Changing crossover characteristics for freshness effect is poor choice if room reverberation is real issue. On other hand, if small changes here cause perceptual shift, then speaker/crossover is still culprit.

Standard for me is getting reverse null of driver pairs as deep and symmetrical as possible. I aim for null of at least -35dB. Center distance between driver pairs less than quarter wavelength of crossover frequency is required for coherent convergence as single virtual source. A lack of coherent convergence leads to sound radiation through crossover region in multiple directions at differing intensities. For reverberant source signals within crossover region this leads to doubled up sound field that is prone to detection through small head movements, leading to unmasking of system as source. This is further sullied up by need for two sources to get phantom images.

Manual EQ as independent phase and amplitude control for smoothing responses amounts to partial inversion of system impulse response. Matching Q of filters becomes very important. This increases number of filters required to get smooth target response. Balancing levels of filters becomes too time consuming. Automated methods as with REW software is still only partial solution.

When considering temporal aberrations do to energy storage and release, radiation of reflection from inside enclosures back through driver, on to external reflections/diffraction responsible for speaker acting as multiple source, to longer reflections effecting both bass performance and producing standing waves, manual methods are dubious at best. Manual methods are highly dependent on understanding of engineer.

Fully automated inversions are problematic as well, but lead the way. Big pitfall is picking reference point(s). Another potentially is measurement of the system IR in terms of noise/distortion. And also issue of microphone response being acceptably flat or correctable.

I've gone acceptably flat route, using Earthworks OM-1 for measurements. With them I can recover typical response behavior of Behringer ECM8000 microphone as measured by others. The WM61 is perfectly applicable for measurement microphone. Simple pre-amp with reference calibration works well. Typical is rising response above about 4kHz that peaks between 7kHz-12kHz. Some builders include compensation filter in pre-amp to get flatter response.

I don't find noise an issue with capturing IR, inverting it, and using this as primary correction with convolution engine.

I follow quarter wave criteria and measure between tweeter and mid/woofer on listening axis 20cm-23cm from tweeter face. For cross to woofer <100Hz, quarter wave criteria is met too.

One, two, or three measurements: With multiple channel convolution set up I've tried this a number of ways. For two-way Pluto Cone, I've done both single measurement with both drivers running with own crossover (DCX2496). In this scenario I've gone with/without EQ from DCX2496. With EQ I've done it with multiple iterations, one tweak at a time, and with modeling from REW (both manual and automated). With manual methods I've tweaked DCX2496 until reverse null deeper than 35dB is seen. Under these conditions the speakers perform exceptionally well, likely on par with Linkwitz's implementation. From this point I've taken IR, inverted and used for correction. And also from point where DCX2496 is solely crossover. Both cases produce corrected measurements that are virtually indistiguishable by ear.

With multiple channel convolution I've measured drivers independently, generated separate corrections, and crossover filters. This result is also virtually indistiguishable from convolution corrected DCX2496 results.

All convolution corrections produce superior results (In my not so humble opinion) to manual results.

Speaker is driving source of all standing wave (modal) behavior. Mode peaking is greatest in lower frequencies <3kHz, where modes are also sparsest. Through most of this range, microphone at 20cm measurement distance is less than one wavelength from source, yet correlates to all modal peaks within room. Benefits extend well beyond sweet spot. Timbre and amplitude of instruments across there ranges is balanced and natural.

Phase relationships between harmonics are maintained.

Reverberation excited in room decays smoothly across spectrum.

Various "room correction" software accomplish this. Common variant is multiple measurements proximal to listening position. This is hunting for modal peaks, when these all happen nicely at source.

I do inversion with Kirkeby inverse as implemented by Farina, and available as plugin for Audacity and for Audition/Cool Edit.

Are you doing multiple channel convolution for active system?

With REW or ARTA it is relatively easy to get multiple measurements from single microphone position for tweeter/mid/woofer that maintain time base. Other methods with periodic sweep and stable DAW/sound card are possible too.

You've taken the first steps in becoming familiar with power of convolution engine, and limits of manually derived filters.

Your microphone may not be flat, but it is likely flat enough to begin exploring mathematically derived inverses.

To this end I'll post inverses via Kirkeby if you post some 20-23cm measurements.

What are dimensions/type of mid/tweeter setup? With multiple channel convolution tweeter crossover may be made very steep and low enough to meet quarter wave spacing to mid without damage or intolerable IMD generation. In this regard I've found that tweeters may be crossed lower, and driven harder with lower IMD than possible with typical 24-48dB/oct crossovers. Steep crossover also helps maintain speaker as single source in crossover regions.
 

Jay

Banned
2003-02-11 9:02 am
Jakarta
I have a strange abberation, I want the music that I like to sound good; and it was very apparent in the earlier years of hifi that it works in reverse on many system: only "correct" recordings work properly,

I cannot understand this common statement that a good speaker can be worse than a bad speaker because it reveals flaws instead of masks them.

For me, good system/speaker is a good system/speaker. It reveals flaws in the recordings, but it still produce better sound than bad speakers.

Standard systems are a million miles away of being able to do this, you can tell it every time because they make a mess of the drumkit sound, cymbals only very roughly resemble anything that a real drummer ever hears.

Most of the time (if not all), we have to live with far from perfect loudspeakers. We have to choose between compromises (including higher cost of perfect drivers). So it is about designer's knowledge on what is achievable and what is important.

I think many speakers (including high end ones) are terrible, simply because the designer relies highly on measurement tools (and nothing else).

So, the point of tweaking is to get, say, the drumming on a "crude" 70's rock recording to have the impact and finesse of an uber recorded audiphile effort ...

There is a situation where tweaking is just equalizing the system. The system may sound okay for certain music/instrument but sound terrible for everything else.

For me, the basic requirement is that any speaker must be first designed correctly. Then from those correct designs there are bad speakers and there are good speakers.
 
There is a situation where tweaking is just equalizing the system. The system may sound okay for certain music/instrument but sound terrible for everything else.

For me, the basic requirement is that any speaker must be first designed correctly. Then from those correct designs there are bad speakers and there are good speakers.
Well, for me the basic requirement is that the system works correctly, :). Having gone the exercise of sorting out audio playback a number of times, I have no trouble seeing the problems lying beyond the speaker, reaching right back through the whole chain, up to and including the quality of the mains power.

There is a level of tweaking where all recordings, I repeat, all recordings, can come good; not easy to do, but certainly attainable with sufficient focus and effort.
 

Jay

Banned
2003-02-11 9:02 am
Jakarta
There is a level of tweaking where all recordings, I repeat, all recordings, can come good; not easy to do, but certainly attainable with sufficient focus and effort.

Everything, I repeat, everything, is scientifically explainable, unless it is a guesswork. And I want to know the formula and quantitative/qualitative goals of this not-easy-to-do tweak.

Unfortunately, like I mentioned previously, I have never experienced those situations where a good system is supposed to sound bad just because the recording is bad. And I'm still confused about this.

My main speaker is more revealing than Grados and Sennhesisers that I have compared to. Mp3 sounds like sheet through the speaker (compared to CD/FM radio), but I still enjoy it more than through the Grados/Sennheisers.

I mean, I don't understand why it has to sound bad just because it is revealing.

The harder problem imo is to relate good sound and measurements with enjoyment. Like I said before, choose drivers that can be crossed with 2nd order electrical filter without a notch filter, and you are close.

But the most revealing or detailed drivers are those from light but rigid cones. Cheap ones need a lot of interventions such as notches and higher order filters. The higher the order, the more critical the precision of the components values, and it is not measureable. This is the trap of many ill sounding high end speakers imo.

These details, is not a must for an enjoyable sound. But we want an enjoyable sound that has details. So what we do is pick the challenge, not compromise the enjoyment for the sake of details (etc).
 
I cannot understand this common statement that a good speaker can be worse than a bad speaker because it reveals flaws instead of masks them.

For me, good system/speaker is a good system/speaker. It reveals flaws in the recordings, but it still produce better sound than bad speakers.

Whether or not it's a flaw in the recording or real, live music, sound can sometimes be quite hard on the ears, and my ears do get tired there's no doubt about it. I could certainly envisage a live performance where someone's guitar pedal has a constant high frequency whistle (there's a couple of Jimi Hendrix tracks like that I seem to recall) where a speaker with a rolled off top end would make it more bearable, for example. Constant pounding of the ears with the full dynamic range of an orchestra or rock drums will tire your ears quickly, whereas listening to it on a system with limited dynamic range (i.e. a compression characteristic that may also give distortion) may allow listening for longer - which may be misinterpreted as less listener fatigue = better speakers.

Or, I was reading a blog only the other day where a musician was recommending that if a clean digital recording is sounding a little sparse, dub it to cassette and back to make it more cohesive. Not completely implausible I would say.

I have some old Kef Concorde speakers that are quite big, but have a very relaxed, old fashioned sound that is certainly not hi fi in my opinion, but nevertheless is very pleasant and easy on the ears.
 
IMO, there's a definite threshold for "listenability". once that threshold is reached, people should just stop, listen to music and enjoy life. at this point, it should be noted that life doesn't equal audiophilia. when funds allow and the improvements gained for the money are OBVIOUS, go for it.

disclaimer: the above is referring to the "non music loving" subcategory only. don't tell me there aren't such animals, I know better.

@mr push_pull

I absolutely agree that from my experience at a recent audio show, there's something not quite right about a bunch of middle aged men listening in reverence to a recording of a girl singing MOR songs to a guitar. It does make me wonder exactly what it is that my fellow audio enthusiasts are listening for.

I hold my hands up and admit that my ears do get tired after listening long and loud. I am no longer able to judge what is good and what is bad sound. As for identifying cable A from cable B, forget it! It would appear that I may be almost unique in this regard around here. However, I most definitely am a music enthusiast who finds it hard not to latch onto, and follow, music whenever and wherever it occurs. Which is an irritation, because my way of listening to music is to sit down and listen to it very carefully indeed - it's the little unexpected details that make the hairs on the back of the neck stand on end. But careful listening in the face of unchanging reverberation sounds like something that should grind the listener down, eventually. I think I must look into improving the room acoustically - it's not great at the moment.