Speaker/crossover parameters that influence the resulting sound

[speedle]

Yes, I know, most everything influences the end result, but I"m interested particularly in two. Without taking anything for granted, what aspects of a design, whether it's baffle width and diffraction or frequency response and phase response/group delay, get you an end result that images well, or has a good sense of depth, or presents the image in front of the speakers versus in a place behind the speakers? I've not been able to find much about this online anywhere at all, and I'd really like to know. It must be known, somewhat, to designers, right? And hence conscious choices are being made that get you different things in the end?

[JonMarsh]

Everything in the chain is important to this. also, the room and room reflections and comb filtering are very destructive of achieving this effect. There is a lot of discussion about this on this site, scattered among threads, and much discussion about what works and doesn't work in various hi-fi/audiofile sites. You probably won't see it discussed all that much in DIY sites. For a short discussion on some of those effects in digital components, take a look at the new Digital Audio section. Unfortunately, with regards to imaging, there can be significant differences between $1K DACs and $5K DACs. The mark of a really good DAC in my experience is how surprisingly good 44.1 kHz CD can sound... my personal best DAC with good 44.1 kHz CD material can beat out the $5K SACD player I have with the SACD layer of the same performance. This is the sort of thing rarely discussed...

Unfortunately, most people, whether DIY or not, have constraints from spouses and other issues which prevent them from setting up their system optimally (speaker and room) even for what ever level of performance is possible with their existing gear...

For a good starter in this area, check out this page at Cardas:

http://www.cardas.com/room_setup_main.php

Few designers or users unfortunately even take into basic room boundary issues impacting the immediate low frequency performance in a real room; see this short reference thread:

http://www.htguide.com/forum/showthread.php?10872-Guestimating-room-gain&p=138262#post138262

It's rare that I see pictures of folk's builds on HT Guide setup following these kinds of basic guidelines, or following the more advanced room setup practices that jim1961 uses, for example.


Also, its important to remember that while every component creates distortions, the subjective effect of the different colorations and distortions can be quite different. This is especially true when small signal details are masked by other large signal events- this also detracts from the reproduction of an acoustic space. This is an area where a lot of digital gear falls short, but even many popular ICs, such as those commonly used in volume control circuits in HT gear, have less than ideal characteristics.

Last, if the recording wasn't made in a natural way with the right microphone techniques, then there is no real acoustic space to reproduce, quite literally.

For an extreme comparison of what I mean, compare the Sheffield "King James Version" of Harry James's band, or "Jazz at the Pawnshop", with a typical modern pop, rap, or metal album.

For a discussion of one classic album often referred to for imaging, see this review.

http://www.tnt-audio.com/topics/james_e.html

I have both the original CD version and the 20Bit K2 dithered to CD version, which I prefer.



An album I often use evaluating the performance of digital gear comes from the same time period in the mid-late 70's- Jazz at the Pawnshop.

http://www.firstimpressionmusic.com/product_p/limuhd071.htm

It's actually a bit of a toss up as to whether the SACD release, or the new Ultra HD CD release; it depends on what gear you have to play it back (I have both), and I can also extract the SACD into 176.4K/24bit PCM, and play it back on the PCM DAC (haven't gotten around to that yet, though- ). Right now, the Ultra HD on my Total-DAC D1 beats the SACD version on Marantz SA-7. YMMV, of course.

Now, the thing is, you don't have to have super duper recordings but you can't have junk, either. In a good system, with decent recordings, the typical reaction should be like when my GF heard my system for the first time: "That's strange- I don't hear anything coming from the speakers, just sound between and behind them..."

[Carl V]

In a good system, with decent recordings, the typical reaction should be like when my GF heard my system for the first time: "That's strange- I don't hear anything coming from the speakers, just sound between and behind them..."

I've heard that a few times as well. Usually preceded with "is there something wrong with you speakers?"

Phase accuracy, power response (polar response plots both horizontally & vertically), Room symmetry.

I just visited a friend who has a AYRE & Utopia set up (used to be Avalon) nice very nice.
However, contrasting this with a colleagues Pass system Speaker as well...ARC CD.
it's entertaining to hear differences in darned good set-ups. both will tick off the boxes
of needed objectives....but both have a different 'sound'.

[speedle]

That's what I'm aiming for, yes! So, a good design addresses phase accuracy and polar response, and of course room symmetry is up to the end user, but is there any agreement as to what good phase is, and whether CD or broad response get us the best imagine?

[JonMarsh]

CD is fairly ideal, but difficult to achieve except with dipoles.

The flip side for a natural balance is both a smooth flat direct sound and a clean smooth roll off in the indirect sound. Many speakers go afoul of this because their off axis response has holes in it, due to driver directivity issues. The more reverberent room sound you have to deal with, the more this becomes an issue.

Finally, a good way to check on whether you're getting the subjective response you expect is a test CD like the Sheffield MyDisk which has both correlated and uncorrelated pink noise. The ear is fairly sensitive to a source like pink noise, and that actually makes it easier to identify changes in response due to on and off axis issues. And why correlated and uncorrelated? Uncorrelated gives you two separate channels with pink noise that is uncorrelated- it should sound like two separate independent sources. Correlated means just that- the same left and right - which in principle should give you a phantom center and little or no apparent sound from out side. Get that setup right and imaging with musical sources will be pretty solid- at least as far as the acoustics go. then you just have to deal with the limitations of source components and recordings, as mentioned above- if earlier components in the playback chain won't properly reduce the more subtle cues of a recorded acoustic space, the speakers can't make it happen by themselves.

[jim1961]

For me, the biggest improvement in soundstage came from treating my room. More specifically, controlling early reflections. When you consider that the brain psycho-acoustically blends everything it receives in the first 20-30ms or longer (depending on who you ask) into one coherent first impression, then its becomes easy to see how early high gain room reflections can subtract from or even collapse a soundstage. Once early reflections are minimized (to -20db to -30db) for the first 20ms or so, then what your hearing, but more importantly, what your brain interprets as the direct sound is mostly free from room interaction and you get a much more accurate representation of what the recording actually sounds like in all its 3-dimensional richness.

After the room, the speakers are the most relevant, IMO. Then the source.

When you consider that a good speaker might be +/- 3db from 30-20k and a typical untreated room can be +/- 15db (or worse), then it becomes obvious where the most damage to the intended sound encoded in your source material comes from.

[JonMarsh]

what he said, +1