What are you experiences with imaging and sweet spots?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I recently heard a horn system which made me question what is it that makes for good imaging and a broad sweet spot?

The speaker in question has a CD in a large 200 Hz horn on top of a vented 15" pro driver. I'd call the sound "direct." I've heard it said that wide dispersion puts the musicians in the room and narrow dispersion gives a window to the musicians. I think that would describe what I heard.

So I'd like to hear what you guys have experienced - what speakers you have heard and what kind of imaging and sound stage they created as well as the sweet spot.

This is what I've experienced:

* Tannoy - fairly recent dual concentric with 6.5" midbass drivers (vented bookshelf)

Imaging appeared to be good, despite the room being much like a garage with carpet and a couch acoustically.

* My mains - MTM with dome tweeter and 6.5" midbass

They can image quite well as previously mentioned. The sweet spot is such that you can sit off centre without the image collapsing into one speaker - a centre speaker isn't needed for home theatre.

* My mains on an open baffle panel

Noticeably bigger sweet spot, which was welcomed. The sound stage was bigger and the overall sound more appealing. When I put the drivers back in their boxes again, the sound stage was suddenly smaller.

From memory I'm not sure if the imaging is better or not as I didn't pay specific attention to that aspect, but the sound was more transparent, creating a more real illusion of the musician being in the room.

* Adire HE10 (coax 10")

Imaging did appear to be good but the treble harshness made it hard for me to really evaluate.

* CD horn

See previous comment at start

I've heard a great deal many more speakers, but these are the ones where I paid some attention. I've heard some omnis and I can't say their imaging was actually any different to conventional hifi speakers from memory. It was a surprise at the time, as I felt an omni was the worst possible design from an imaging point of view.

A more theoretical question:
How does one achieve a large sweet spot?

I'd like to experiment with horns and dipoles quite soon as an upgrade, I just wonder if I will be happy with the sweet spot and imaging of horns and if I should design any horns to have dispersion as wide as possible. One of my current ideas is a 3 way with a large mid horn (either CD or cone mid driven), concentric supertweeter and 12" pro mid on open baffle.
 
what is it that makes for good imaging

Sy, good answer!

A short answer and leaving out driver qualities and filter requirements for the moment:
A speaker stereo (?) or multi-channel set-up that minimizes phantom localisation errors, vertical and horizontal together with a physical layout that maximizes phantom sharpness, minimises time/intensity trading smear/blur effects by either layout of the physical drivers if using minimal baffles and/or smoothing them out buy intelligent choice of baffle layouts.

b
 
SY said:


IME, a rather large room and omnidirectional radiators.


I rather agree with SY, ..... BUT:

In my opinion you will always have a "problem", if you don't have the same distance to your speakers - no matter what kind of speakers - because you will have a phase problem. Almost all sounds from your stereo comes from two points, and unless you have the same distance to these points (speakers), the sound are out of phase, right?
So stick to your sweet spot for serious listening
:smash:
 
bjorno said:


Medum, No contradiction at all but a sharpen up the general requirements of the thread owner’s wish: broad sweet spot and entering the lower level where imaging can be discussed?

b


of course - If you want a broader "sweet spot" - go for the reflections from the walls. My Marin Logan Clarity speakers have a rear mounted tweeter, that can be activated for "party-listening" - maybe that could be a solution. But I must say that it is not really a sweet spot. Unless you have the same distance to the speakers, you are not in the sweet spot.
 
If you want a broader "sweet spot" - go for the reflections from the walls.

No offence Jørn, but now the subject is expanding into including room effects and I find it hard to stay focused at the original question.

There is no indication that the room effects should be counted as I can see. I’m hesitating to further follow this thread because the risk that this matter is divergating too much for my taste.

But maybe this picture can be used as good starting point if a discussion of pros and cons of the sweet spot; sweet spot area is essential for the thread owner?

Look at:post#2

http://www.diyaudio.com/forums/showthread.php?s=&threadid=84731


b
 
FWIW, my experience has been similar. My first DIY's were an MTM. They sounded good, had good imaging, but the soundstage was not particulary big and was sort of flat front to back, nor were they forgiving of listener position.

My second major DIY was the NaO Mini, open baffle MTM. The soundstage was much bigger and deeper. The imaging was still very good and more forgiving of listener position. You get a real sense of 3d space around them.

My latest project is a hybrid omnidirectional design. Each speaker has a downfiring 12" sub for 100hz down, an up and down firing midwoofer covering from about 100hz up to 1450hz. Then there is a front firing tweeter that covers from 1450 up to 21khz or so. It has a much bigger and deeper soundstage than even the NaO and for the first time, I seem to be able to hear the size of the recording venue, for example one recording sounds like you are in a small nightclub and another sounds like the soundstage is well outside the walls of the room you are in. Instruments can sound like they are way behind the speakers. However, I think becuse of the direct firing tweeter, they still have good imaging and are very forgiving of listener position, IMO. YMMV.

When I was building and testing the omni prototype, I was able to compare a single speakers sound from each of these types. It was eye opening to me, but sort of hard to describe. Here is the best way I came up with to describe it. You know how a single box MTM sounds so much smaller than a pair of stereo boxes? I could describe the single box MTM as having about 1/4 of the soundstage as two of the box MTM's in stereo. Compared to the box MTM, a single open baffle NaO mini could be described as having 1/3 of the sound of the pair of NaO's. Compared to the others, the hybrid omni had more like 1/2 the sound of the pair of omni's. A single omni almost sounded like it could be as big 3/4 of the soundstage as the two boxes. Here is a photo of the hybird omni:

Mentor1Omni2s.jpg
 
Digital crossovers do a good job of enlarging the breadth, height, and width of the sweetspot..but at the expense of uhm..analog micro transient integrety..which is where the heart of the muisc lies. So..they make it work notably better..but the bathtub strangely has no baby in it...it has been tossed out.

What you get is perfect phase integrerty in the crossover range..but nothing has it's original 'fine detail' which is where the music itself lies.

This, I find, with a near equivalent to a $10-12k digital crossover. A total of 28 op-amps ain't doing anyone any favors either (about 32 in balanced I/O mode!!). Makes for a nice pretty pro audio diagram, but it don't do fecal for music. My next trick after this audio show in April is over: Transformers in..transformers out. As few op-amps as possible. I'll leave the in and out (ADC/DAC in/out op-amps), but the rest are taking a walk.I'm actually looking at the transformers..right now...they are waiting.

What I think I need to do, is to leave the drivers in a near 'orginal' condition, concerning their outputs..and use the digital crossover for a SIMPLE 6db per octave corssover ONLY..and leave ALL other corrective algorthims and mathematical manipulation...out of the equation. Only then, I suspect, will the music even attempt to re-emerge from the system.

I was using the behringer dcx-2496, with a corrected clock, in this simple function..and I gotta say..it sounded a hair better than this expensive thing I've got now. Why? the simple mathemematical transforms enected in a limited way (frequency range covered), which also has an effect of not delaying the digital signal's arrival to the given dac's input, thus cutting down jitter. I have also yet to correct the clock on this new device. How can I say such with any sort of 'numbers' to back it up? well, let me say I can hear the phase integrety on simple signals..and it goes away on complex passages. Thus, I suspect the complex mathematical work done on the more complex signal, delays the signal's entry to the given DAC for that driver. There is a 'window' for arrival of that given signal..and I suspect the complex math enacted on simple vs complex passages causes that to jump around in the micro-timing-arrival moment sense. Thus, complex signals being responsible for more jitter. Second point: The complex mathematical transforms create a situation where the ORIGINAL transient POINTS are SKEWED....thus the actual music signal itself is fundamentally.....SCREWED. Enecting a simple crossover function alleviates the vast majority of this - as issues go.

Follow the logic, ferret it out...you'll eventually get what I'm saying.
 
here's the critical point of understanding:

The human ear works with transient leading edges, their time period, level, and distance between. Ie, a diode like function, but transient peak in nature.

This means that what is seen on a scope, nearly 90% of it can be tossed away, it is irrelevant to the ear.

The newest thinking and enacting of MP3 algortihms and low data rate similars seems to show this point in a very clear manner. LP's keep timing between and relative differences in levels of transients....they get this right. Same as analog tape. These new algortihms do this trick, thus preserving what we NEED to hear, to call it 'musical'.

Thus, any mathematical transform systems that even slightly damage the placement of transient peaks, even in a 0.01% manner (remember this is whole signal weighting!!! not the critical transient that must be analysed in isloation!), are going to wreak total havoc with what the human ear does with a given signal, when it comes to 'hearing' the music. This point also holds true in terms of anlog manipulation, in all ways. (passive crossover, full analog signal manipulaton, etc)

The fact that a horm speaker has about 40% distortion..but most if not many folks find them to be startlingly clear and correct in their presentation..well..this is tied to the fact that they preserve leading edge transients..but not much else. This point illustrates....my point - quite effectively.

Then there is the fact that we can hear down into the picoseconds of differences of jitter in differing clocks....

This effect I am speaking of, where simple vs complex change, in terms of what sound like phase (thus transient peaks) issues..this also rears it's head in the 'free floating' tripath chip's 300khz oscillator, which from my feeble understanding, it's behaviour is tied to the complexity of the incoming signal. On simple signals...the chip sounds great. With complex signal..it turns to mud. I'm hearing a hint of similar effects here, and I'm searching around for the cause.
 
The human ear works with transient leading edges, their time period, level, and distance between. Ie, a diode like function, but transient peak in nature.


The fact that a horm speaker has about 40% distortion..but most if not many folks find them to be startlingly clear and correct in their presentation..well..this is tied to the fact that they preserve leading edge transients..but not much else.

Dear KBK,

could you please provide some reference(s) substantiating the assertions quoted?

Thank you,

M
 
It varies with the given horn. Horms destroy the back wave or negative character of the waveform. Some horns are at about 25% distortion. Some are far higher. Depends on the given frequency, and design. See Martin Colloms book (Loudspeaker Design), for example. One of many places where the information can be found. Measurements of horn speakers vs 'dynamic drivers-baffle loaded' reproducing square waves, for example, which would take the entire UHF magazine catalog. UHF shows the recreation of a square wave in the acoustic sense, for every speaker they've tested, over the years. In a horn, the very VAST majority of the distortion is in the backwave, or negative part of the wave.

This, as a piece of data (one of many, noted, not cataloged or documented), and the rest of the data together, show that transient peaks (the ear as a diode, single ended, like our voices-documented in many places) their placement and level are the critical components of the idea of humans and sound. The data is scattered..but the understanding is clear.

You will find little of this documented, but the horn distortions are well known and documented.

File it under 'lead, follow, or get the hell out of the way', which is that strange grey area where those who -do- march forward, and those who need paper and documents and prior art....are left in the dust. :)
 
I think I now see where your misunderstanding is coming from. Crossover filter algorithms be they FIR or IIR do definitely differ from compression algorithms in that tey are time-invariant (apart from MANY other things). They do not analyse what is going into them and then behave differently depending on program apart from some driver protection algorithms but these are rarely used for home audio.

They do however introduce errors like any other crossover does. IIR filters introduce phase errors like ordinary analog filters (increasing with filter order). FIR filters introduce pre-ringing. Both of these are errors in the time-domain.

The switching frequency of the Tripath is not depending on the input signal complexity but on the drive level.

I do however agree with you that timing is critical and I terefore use a transient-perfect active crossover. This is using some of those "feared op-amps" though. But additional A/D- followed by D/A- conversion is still worse than that IMO.

Regards

Charles
 
got it - thanks. I knew this, but one also has to be told a few times to actually have it sink in. A slight misundestanding of questions vs answers here, but I think we both got something useful out of it. As far as the nanes and types of the digital filters and their exact technical complexity, I am relatively ignorant of them, beyond casual analysis. No time to specialze that deeply, too many other things to do. This does not mean lack of intelligence, merely ignorance -- which I wish to, and strive to clear up. But I know what I hear!

As far as digital goes, I think I'll have to go to, as stated, simple 6db per octave crossovers..and leave all other processing that the unit is capable of....out of the sonic equation. I'll keep the more complex 'settings' for 'wowing' people with (the kind of people who enjoy such sound), but for myself and others..this is the only way we are going to be able to enjoy digital crossovers. Keep them simple, and maybe, eventually...they will be OK.

As for that crossover you use..what is that? Custom or available?
 
My crossover is a custom one. It uses a 2nd order highpass slope and 1st order lowpass. The step response od the speaker can be seen here:

http://www.diyaudio.com/forums/showthread.php?postid=1139711#post1139711

Kepp in mind that the MSW has some response irregularities that arew not yet taken care of and that's where the "nervous part" after the rising transient might come from. As soon as I can make outside measurements again I will take care of these as well.

Regards

Charles
 
What about using transformers for a 'bleed in' of the (-) phase of the HF, to create the LP? That's not the whole story, mind you, but ..just a thought ( A dual wound transformer sandwiched between resistive-active sections for isolation, gain and control of the effectiveness). I dunno. Just throwing it out there.
 
b: What happened with: What are you experiences with imaging and sweet spots?

Thank you Dan for sheering your omni speaker findings with us and believe this is in line with what the tread owner are asking for. But what has happened next? Im confused. Maybe a tread mix-up (?) or read this quote:


Digital crossovers do a good job of enlarging the breadth, height, and width of the sweetspot..but at the expense of uhm..analog micro transient integrety..which is where the heart of the muisc lies. So..they make it work notably better..but the bathtub strangely has no baby in it...it has been tossed out.What you get is perfect phase integrerty in the crossover range..but nothing has it's original 'fine detail' which is where the music itself lies. This, I find, with a near equivalent to a $10-12k digital crossover. A total of 28 op-amps ain't doing anyone any favors either (about 32 in balanced I/O mode!!). Makes for a nice pretty pro audio diagram, but it don't do fecal for music. My next trick after this audio show in April is over………….


b:Is this really posted to the right thread (?)Reading this tread further, I also find some answers, politely and accurate as usual by DIY member Charles!

But still Im confused and may write to the wrong thread, cannot keep up with the phase that this tread has reached and don’t want to enter any thread trying to explain matters of sweet spot concerning or other psycho-acoustical findings I find outrageously as quoted above.

Im not of this native language and must say Im 99 % handicapped to express my thinking of this matters in English even though I been following this research area closely for about 30 years, psycho/music –acoustics research and audio design almost a half century.
So I could be lost and even misunderstand the relevance of these suddenly appearing postings or not capable to reach the level at which this post is aiming at.

I tried also to read behind the lines but fond, believed I entered the audiophile High End Voodoo landscape twilight zone when I begun to read this:

Digital crossovers do a good job of enlarging the breadth, height, and width of the sweetspot..but at the expense of uhm..analog micro transient integrity

b:I never have heard of statements like this clarifying any aspect of audio from the professional or scientific research worlds I visited and if I for an occasionally circumstance looked in a nonsense Hi-Fi magazine using terms like this I simply stop my reading further.

b:Quoting myself:...outrageously as quoted above. It continues:

Then there is the fact that we can hear down into the picoseconds of differences of jitter in differing clocks…


b:I can only agree within the realm of a couple microseconds from the research area when talking about differential timing between the ears and the brain in charge but where can I read about the picoseconds that destroy imaging or the sweet spot? I recall a paper that cover this but not state the quoted above:

http://peufeu.free.fr/audio/extremist_dac/files/jitter92.pdf

Or the Quote from: http://www.hydrogenaudio.org/forums/lofiversion/index.php/t43841.html


Then there is the fact that we can hear down into the picoseconds of differences of jitter in differing clocks…
Quote :WmAx Apr 21 2006, I can only speculate that they saw no reason to do further research due to the very high values found to be required in the preliminary. Since the objective of this test was to determine if jitter was an issue of sound quality in average equipment, no reason existed to perform a test establishing precise values. Those graphs you posted were extremely interesting. The threshold for even the best listeners was in the order of nanoseconds, and tens and hundreds of nanoseconds in some cases. This is interesting because, as you say, other studies have claimed audibility of jitter in the order of a few picoseconds. The method you described indicates that the test was less than perfect, but the results are still interesting.More interesting reading (which places the threshold of jitter audibility much lower than the dolby article):Jitter: Specification and Assesment in Digital Audio Equipment by Julian Dunn Stereophile: Bits is Bits It really seems as though nobody has done a double blind test with decent methodology. Every test I have seen is flawed in some major way. It's a pity that these tests are so hard to do or we could organise one here and get a good answer to the problem. It seems the best answer you can give at the moment is that the audibility threshold of jitter reduces with frequency and lies somewhere between 10ps and 100ns.

b:There could be other contradicted findings in this matter, I would be grateful to be acquainted with, if existing.

QUOTE]…You will find little of this documented…[/QUOTE]


b:Can this be found at the urban legend library?

…File it under 'lead, follow, or get the hell out of the way', which is that strange grey area where those who -do- march forward, and those who need paper and documents and prior art....are left in the dust…

b: This is a DIY-forum not a grey area residing in the heads of besserwisser’s and we all need proofs or guiding we can accept before we allow us to swallow the dust.

b:I think this is beyond audio engineering connected to psychoacoustics and even beyond music as an art entering the psycho-music area and think this is a layman’s approach to express his conviction in a way he think is common and passable and common here at DIY-Audio, but what do I know.


Charles:I do however agree with you that timing is critical and I terefore use a transient-perfect active crossover. This is using some of those "feared op-amps" though.

b: I agree

KBK:As for that crossover you use..what is that? Custom or available


b: Last comment: Not knowing transient perfect crossovers says it all.

I’ll stop here commenting matters I don’t find relevant but maybe for others to focus on but hope that the tread owner and the other who posted earlier, returns to this tread I found interesting starting only from the beginning but took a leap into matters I find is extreme for most of the forum members.
Someone must agree?
b[
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.