• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Cosmik's system

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,027
Location
UK
#1
My philosophy is that the stereo recording should emerge from two speakers in the form of sound pressure variations that match it as closely as possible. To this end, I use:
  • digital source in the form of an asynchronous USB multichannel DAC
  • six channel AV amp (90W/channel I think - way more than is needed)
  • three-way active speakers with fairly low cost drivers (they have a very easy job to do)
  • re-used vintage speaker enclosures
  • large, sealed woofers
  • a Linux PC-based DSP convolution engine whose software I wrote. The PC does the following:
    • provides bit perfect playback: sample clock is provided by the DAC
    • plays CDs, audio files and streams off the internet
    • corrects the individual drivers to have linear phase
    • provides the crossover filtering
    • provides delays for time alignment
    • provides gentle pre-calculated linear phase EQ for best response at the listener's position
  • A fully-carpeted room something like 15x15 feet with a higher-than-average ceiling.
It is a minimal, yet maximal system. The Sony amp is pretty large; the DAC is a small box, as is the silent, fanless PC. The speakers are pretty big in the retro monkey coffin style - which I like. The software is doing much of the clever stuff.

It is the nicest system I have yet heard - with the standard bias caveats.

I may apply the same 'formula' to some other vintage speaker enclosures that I like the look of, but I don't expect any further substantial improvements in the sound. Omni versus directional may be an interesting avenue to explore. I could also be interested doing the Kii-style bass dispersion stuff, but I don't really feel that it's necessary at the moment.
 
Last edited:

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
9,910
Likes
8,302
Location
Riverview FL
#2
It's nice to see a description of your setup. Thank you.

Any measurements to share?
 
OP
Cosmik

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,027
Location
UK
Thread Starter #3
It's nice to see a description of your setup. Thank you.

Any measurements to share?
Not really. I have the impulse responses used to correct the drivers, but apart from that, my aim is not to, effectively, use feedback from the listening position to 'correct' the room. I apply pre-calculated EQ and time alignment delays rather than from measurements in the room - and maybe to people's horror, I set the depth of that overall curve by ear (pretty much like the 'target curve' you may sometimes mention). Beyond that, the room can do what it likes - as I have no doubt bored you with before(!):), I subscribe to the philosophy that says we "hear through" the room to the source, so I don't want to muck about with the source any more than necessary.
 
Last edited:

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
9,910
Likes
8,302
Location
Riverview FL
#4
Two philosophies then.

You correct the speaker drivers individually, because you can, and call it a day.

I correct the speakers from 10 feet away as a unit, because that's all I can do, and call that a day.

That's fine. I wouldn't dare to argue which is better.
 

fas42

Major Contributor
Joined
Mar 21, 2016
Messages
2,818
Likes
150
Location
Australia
#5
I'm curious how close they get to being invisible, in the sense that I talk of ...
 
OP
Cosmik

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,027
Location
UK
Thread Starter #6
Two philosophies then.

You correct the speaker drivers individually, because you can, and call it a day.

I correct the speakers from 10 feet away as a unit, because that's all I can do, and call that a day.
Sort of. The way I see it, if I just corrected the individual drivers to be 'flat' they would be correct in an anechoic chamber but not a real room.

If I EQ the speakers at the listening position to be flat (or whatever curve I choose), I am in danger of being influenced by reflections that my hearing compensates for, therefore rendering the direct sound as being perceived as odd. My strategy is to EQ the previously-flattened drivers based on baffle step by calculation, and set that constrained (to use a CAD term) curve's depth by ear - this will be affected by the room's s furnishings and so on.

I make it so this EQ doesn't affect the phase, so the direct sound still maintains the integrity of transients, but the in-room FR balance sounds neutral. I am adopting this strategy because a measurement of the FR at the listening position, to my mind, is not an indication of how it will sound ("hear through" the room etc.), but nor is a measurement of the speaker in an anechoic chamber.

If the speaker had constant dispersion at all frequencies, it wouldn't need this EQ curve.
 
Last edited:
OP
Cosmik

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,027
Location
UK
Thread Starter #8
I'm curious how close they get to being invisible, in the sense that I talk of ...
Around the seating position they produce the best imaging I have heard, and you can't tell that there are two speakers i.e. the 'soundstage' is continuous. If you move forwards and backwards that imaging is maintained. Is that the sense you mean?

If you stand up and walk walk around the room, they don't sound odd, but it is not the same as the apparent 'transparency' or 'holography' you get around the seating position - but I wouldn't expect it to be if they were working as intended.
 

fas42

Major Contributor
Joined
Mar 21, 2016
Messages
2,818
Likes
150
Location
Australia
#9
Around the seating position they produce the best imaging I have heard, and you can't tell that there are two speakers i.e. the 'soundstage' is continuous. If you move forwards and backwards that imaging is maintained. Is that the sense you mean?

If you stand up and walk walk around the room, they don't sound odd, but it is not the same as the apparent 'transparency' or 'holography' you get around the seating position - but I wouldn't expect it to be if they were working as intended.
That's the sense I mean. But where I look for an extension of this presentation is that the "holography" is sustained when you stand up, and walk around; the illusion you experience is invariant of the listener's location.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
28,136
Likes
72,255
Location
Seattle Area
#11
a Linux PC-based DSP convolution engine whose software I wrote. The PC does the following:
That's pretty cool. Didn't realize we had such skill set among our members. :)
 
OP
Cosmik

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,027
Location
UK
Thread Starter #12
That's pretty cool. Didn't realize we had such skill set among our members. :)
Many thanks. At heart, it is nothing more than buffers, FFTs and complex arithmetic. The most difficult aspect is the 'infrastructure' that interfaces the DSP with the app that is playing the music, and the DAC. Linux is pretty good for that sort of thing, but not very well documented.
 
Top Bottom