specific speakers for optimum soundstage?

[bser]

For my next project I'd like to concentrate on getting as accurate of a soundstage as possible. However I've never heard a setup that has fantastic imaging so its going to be tough at first. The question I need answered to get started is what to look for in drivers to get an optimim soundstage. Off axis response, what else? Are there any other determining factors or speakers that are well known to be fantastic at imaging?

[Bill Fitzpatrick]

I, too, have a similar question.

Currently I have an excellent left to right soundstage which extends beyond the speakers. There isn't much depth, however.

It would be nice if someone, based on their experiences, could arrange in order, the elements which are most to least destructive to the depth of a soundstage.

I'm thinking that some of these elements might be:

1) Room excessively live with little or no absorbtion on the walls.

2) Electrolytics or other non-premium caps in the signal path. This to include the passive crossovers.

3) Drivers of inferior quality.

4) Driver phasing problems despite a good response curve.

5) Low resolution interconnects and speaker wire.

6) Enclosure diffraction.

7) Excessive group delay.

Also of benefit would be the recommendation of one or more recordings that have a great deal of depth potential.

[DSP_Geek]

My experience, such as it is, with soundstaging indicated:

1) Suppress early reflections from the front of the cabinet. In other words, a cabinet with a lip on the edge, such as almost anything from the 70s or early 80s, trashes imaging. I found that surrounding a problematic system's tweeter with a foam cutout[*] removed some of the vagueness it previously had.[*] imagine a square made of foam with a fat X cut away, and the tweeter living at the center thereof.

2) Oddly enough, depth of soundstage seems to be correlated with low frequency extension. I haven't the faintest idea why.

3) Try for even power response through the crossover region, which means set the crossover point before the mids get too beamy, but on the other hand you don't want to blow up the tweeter. It's an interesting set of compromises.

4) Avoid first reflections! This is what dipoles do really well - first reflections (side walls in most rooms) are almost completely suppressed, so you don't have a "phantom speaker" sending energy at you from where the sound from a monopole would typically bounce off the wall.


Best of luck,
Francois.

[Pan]

Low distortion drivers is essential. That means low harmonic/IM distortion as well as little stored energy iow. a fast decay and clean waterfall.

You likely will need to keep the speaker away from walls, something like 4 feet from backwall and 2-4 feet from sidewalls. You also need some treatments, absorbtion or diffusion of first reflection points.

Go with a two way since these can really excel in this area and are easier to design than a three way. Once you´re satisfied you can use an active sub to get more bandwith and dynamics.

My recommendations for drivers would be Seas L or W series of 5-7" midwoofers or Accuton midwoofers. Tweeters Scan Speak 9800 or Accuton C23.

You need to design this speaker with a baffle step correction unless baffle is very big. This means that you will end up with a full sound instead of the typical anemic performance of small speakers. Typicall midband sensitivity for a 7" unit is 88dB but on a small baffle you get a drop in energy in the 100-500Hz ragne and you simply use this baffle step correction to account for this. This will push down the sensitivity in the upper midband by 3-6dB or so, ending up with a 84dB or so sensitive speaker.

A typicall tweeter with 90dB sensitivity needs to be attenuated by 6dB to fit into this equation.

Take a look at Seas kits "Trym" and "Charis". Also lok up Ellis audio 1801.

There you have it!
I have listened to most types of speakers and systems and some of the best and this aproach gives the most outstanding view into the recording and the clearest 3d holography IMO.

/Peter

[Pan]

Bill,

1) Room excessively live with little or no absorbtion on the walls.

You definately need something on the first reflection points. Google "RFZ" (reflection free zone) and go to RPG for some advice.

2) Electrolytics or other non-premium caps in the signal path. This to include the passive crossovers.

I can´t say specific but electronics overall makes relatively much of difference. Low jitter is essential so a clock mod in the player or sound card may be in place. Prremium opamps or discrete analog I/V stage is worthwhile.

3) Drivers of inferior quality.

Definately one of the most importan issues. Even though "so so" drivers can imagine ok in a good acoustic set up ("RFZ") you will loose clarity, openess and focus.


4) Driver phasing problems despite a good response curve.

Makes some difference. However with a two way you have good phase response and even dispersion from 100Hz-2k or so and that is most important IMO. I´ve heard speakers with prototype x-over swith a dip at 2k imagine like crazy even though you hear that there is some phase problems and a whole in the curve.

5) Low resolution interconnects and speaker wire.

The last thing to worry about. However IMO a good cable can remove some grain and can slightly enhance focus. Small things though for sure.


6) Enclosure diffraction.

As long as the baffle is small it´s no big issue. A small dynaudio look allike monitor type speaker will make it.


7) Excessive group delay.

A sane two way design will be ok and the bass from reflex loading is needed to give bandwith and a full sound. Some of the "size" of the sound comes from low range.

/Peter

[conscious]

I would really do a great deal of room treatment to improve soundstaging. How can you get a good stereo image if first order reflections are messing up your soundstage?

Here's some links to check out:
http://www.soundandvisionmag.com/ide...?topicID=28194
http://www.soundandvisionmag.com/ide...?topicID=38945
http://www.soundandvisionmag.com/ide...?topicID=33004

Your other options are to build dipole or horn speakers, which don't have the severity of early reflections as typical box speakers

[Pan]

Remember that dipoles have much more output to the back which needs more attention regarding damping and/or distance to the backwall.

/Peter

[ScottG]

I've been working on a minor article to post on this subjet for a while now. (who knows when I'll finish it though..)

With respect to the original post though:

Soundstage accuracy and image accuracy are not neccesarily the same thing. Also people differ on the character of both elements. I can provide a general reference point for both (and why), but thats for article length - not as a response post.

with the above caveat:

Generally:

wide dispersion is a must BUT with DSP-Geek's proviso of limited near-time reflections AND with a limited (small) point source driver diameter. Also the type of crossover, how many elements are in the crossover, and the inter-relationship between driver mass and amplifier dampening factor.

Ideally you would want a small diameter driver capable of reproducing from 5kHz down to at least 800Hz. The easy driver choice here (though costly) is the Jordan JX53 on a spheric-esq baffle (about the shape of your head) used with a low dampening factor amplifier (like a tube amplifier). The problem though is that you'll sacrifice dynamics.

To Bill:

are you using a front loaded horn with good horizontal dispersion? That type of speaker usually provides the sound you have described. Without a drastic re-design of your speakers I'd suggest a non bass-enhancement SRS labs design processor. Most recordings have a good deal of depth, so providing one is pointless.

[Bill Fitzpatrick]

Originally posted by ScottG
Most recordings have a good deal of depth, so providing one is pointless.

[ScottG]

Ok, Ok..

outstanding recordings -

may not be your cup'o'tea:

Frankie Goes to Hollywood (pop), "Pleasuredome" is extraodinary (sound-wise, not content)..

Sting: Ten Summoner's Tale (pop with a smidg of cw), A fantastic album (content and sound)..

ANY of the RCA Living Voice 35mm series (classical)..

The one track I use for determining image localization and soundstage ability:

Rod Stewart (Best of): Your in My Heart. Its a good (sound-wise) analog recording that has an increasing number of instruments added to the mix (..very difficult for a speaker to track all the instruments near the end of the song as well as it did at the start of the song).

Live Recordings with lots of depth (excellent sonically and musically):

AC/DC: LIVE! (rock)

Pink Floyd: Delicate Sound of Thunder live (rock)

Monster Depth Track:

The Cult: Ceremony (title track beginning sequence) (rock)

So then, what is the configuration of this speaker that displays out speaker imaging but not depth?