crossover design- depth of field

drewan

Member
2015-03-03 6:10 pm
UK
Hi All,
I embarked on this projet 3 yrs ago having spent nearly a year reading this forum. There is so much knowledge here I cant begin to explain how humble I feel.
Anyway I built a three way with the Peerless 830452, SB satori 16mw8, and sb29rdc, crossed at approx 120 and 1800Hz. I designed the crossovers (many thanks to AllenB and Bwaslo) .
I have now reached a point where the speakers are totally and I mean totally invisible, BUT the downside is that the soundstage is two dimensional, and has no (illusion) of depth. is this something that can be designed in or does the speakers disappearing act kinda preclude that... one or the other?
I deliberately haven't posted any circuits etc yet cos I want to learn:)
many thanks
Andrew
 
Greets!

Been there, done that! :(

Regardless, recommend using a sufficiently adjustable EQ to ~ match this response [less minor dips] as a starting point since it was a very successful ~1936 - 1984? two way 15"/horn studio monitor designed specifically to highlight/maximize speech intelligibility, singing vocals clarity. Because of the expo horn, it 'projects' a bit too far for some HIFI/HT apps since with some recordings it can put the singer in your lap at 12-14 ft on axis corner loaded in a typical USA size room, ditto with some 'bright' cone/dome tweeters.

Valencia FR | GM210 | Flickr

Obviously, it's tuned to 50 Hz and the XO notch at ~7-800 Hz can be ignored and of course your room is going to muck with it, but as a starting point you may get lucky finding a room curve you like.

For sure it's helped me 'voicing' many various rooms/systems way back when before DSP, etc., either didn't exist or priced higher than new Corvette with all the extras.

Compare the response to these two charts to get a 'feel' for how it relates and if successful, maybe you or someone else can work out some XOs, though assume that some frequency shaping filters will be required to make them work.

http://www.audio-issues.com/wp-content/uploads/2011/07/main_chart-610x677.jpg

https://cdn.surfacedstudio.com/2010/10/fletcher-munson-curves.gif

GM
 

Sangram

Moderator
Paid Member
2002-09-25 11:01 am
India
A description like that is near impossible to translate into a technical problem/solution.

I doubt it's an issue of FR or crossover design issue, you might want to consult dave's (planet10) notes about 'Downward Dynamic Range'. Much as a I dislike acronyms and marketing sass in a technical field, there is some smearing of low level detail due to poor cabinet construction and insufficient control over diffraction - mainly secondary effects that have nothing to do with actual crossover or driver selection.
 
I didn't know about the perception of depth arising from speakers.

But as Toole explained, there is a major matter of taste between systems that offer good ambiance and those that possibly are more ping-pong in their stereo. As much as many of us cherish ambiance and disparage systems lacking it, that's not the same as z-axis localization.

Cues to perception of all aspects are established by the recording abetted by the engineer, more and more today with over-cooked recordings and synthetic instruments*. There are several cues to depth and we always pray that we are buying recordings made by engineers who understand perception psychology.

B.
*Anybody believe a keyboard has "a depth" somewhere (except in your mind)?
 
Last edited:
If the speakers really "disappear," they're probably not the heart of the problem. Seating location relative to speakers and other room aspects may be more important to realizing soundstage depth than drivers or crossovers or enclosures. Electronics can also play a part. A lot of SS gear will not provide enough ambient information, from source to pre to amp. I might get clobbered for this, but even cabling can contribute or detract.

You don't give us any information about the rest of the system, and it's all important.

Peace,
Tom E
 
If you have a decent front end and the problem is with the speakers, the usual culprit for lack of depth is poor phase control. The speaker may well have an even response, low colouration, low distortion etc., but without good phase it will sound 'flat'.

With correct phase there comes that magical moment when you can see into the music and place musicians in a clearly defined spot separated from other musicians by space.

As always, posting your design and measurements and information about the rest of your system will help in diagnosing your problem.
 
If you have a decent front end and the problem is with the speakers, the usual culprit for lack of depth is poor phase control. The speaker may well have an even response, low colouration, low distortion etc., but without good phase it will sound 'flat'.

With correct phase there comes that magical moment when you can see into the music and place musicians in a clearly defined spot separated from other musicians by space.
.

+1 ..... i’m a newcomer to dsp integration (hybrid passive/active setup) I think it’s the easiest way to figure it out, still quite elusive but if the stars align it’s really something to behold.
There’s some on here that might have a better handle on it than me, wesayso,pano,ro808,camplo,gm.......
 
Last edited:

AllenB

Moderator
Paid Member
2008-10-18 11:31 am
A comb filter will show this. Using the blue trace, +6dB with 0 degrees separation, -∞dB with 180 degree separation. When nulling, the slightest change causes the greatest variation in level but when summing, the variation is least.

When I wrote my first crossover simulator the first thing I did was chase phase. I had the two way matched to within 1 degree across the crossover region, before I decided to call it. Then I backed off to 5, 10, 20. My notes would be long gone now, but I changed my ways after that.

With some designs and situations, phase seems to come together well, but this is in combination with other things, so there is a bigger picture. One common concern is that drivers are only near each other, so phase may only match where you measured, then when it does it doesn't consider whether it best compensates the other angles with regards to room power.
 

Attachments

  • comb.jpeg
    comb.jpeg
    72.7 KB · Views: 373
Last edited:

AllenB

Moderator
Paid Member
2008-10-18 11:31 am
Yes I have, and no I don't believe so. For a time I'd use a listening position measurement and a clean speaker measurement together, equalising between the two, one peak at a time looking for those in room artefacts that could be 'fixed' without making things worse in other ways. Not very productive unfortunately, this led me to begin considering the room when I designed a speaker.
 
I suppose persistence and blind luck may have paid off in my case......I’ll be able to tell soon how much ‘skill’ was involved as a new room and new speakers are in the works!

Drewan you might consider some form of dsp in the future......a couple years ago I was totally against it but now that I’ve experienced what it can do I’ve been researching my next steps.
 
Last edited:

AllenB

Moderator
Paid Member
2008-10-18 11:31 am
Best with that, I'll have a beer to your success.

My bass waveguide prototype badly needs a revision that I can't settle on a solution to. So I'm sizing up my tweeter for a cross at 500Hz using multiple entry port/s.

Here, measuring the cancellation notch/location.
 

Attachments

  • 20191102_212819.jpg
    20191102_212819.jpg
    787.8 KB · Views: 357

drewan

Member
2015-03-03 6:10 pm
UK
Many thanks to all.
Most seemed to think it was more a room issue, and GM's 'house' curves gave me ideas. Despite the system measuring almost ruler flat from 30-3k and then a slight rise of about 5dB to 20k at the listening position I decided to try bypassing the BSC part on the crossover and moved the speakers about 8in closer to the rear wall to compensate. Great result many thanks.
For those interested the peerless is in a 40 litre box (Qtc about.303)originally built as BR but now sealed, the satori in 18 litres the same, sealed but lined not stuffed giving a Qtc a little over 0.5.
I'm now more than ever convinced speaker crossovers need to be specifically designed for the room they are going to be in:)
 
The OP question is why is depth missing from his otherwise satisfactory system?

Stereo kind of like oil paintings.

For an oil painting, the canvas is flat. The objects on the x- and y-axes are localizable by various familiar cues to perception. But the depth is "faked" using perceptual cues like interposition, perspective, familiar size, change of colour with distance, etc.

What are the cues that "fake" depth in stereo*?

B.
*actually, the situation with stereo is degenerate of oil painting because it is obviously just a single dimension of the x-axis while a canvas is x and y. So what cues are available to put y- and z-axes into your cognition?
 
Last edited: