how important is 'voice matching'? [Archive] - Audio & Video Forums

PDA

View Full Version : how important is 'voice matching'?



elbarono
02-14-2005, 07:45 PM
Is it really important, or just one of those tricks to try to get you to buy all of your speakers from the same company?

If buying two or more different brands of speakers, what are some specs to look for to make sure the system will sound good?

RGA
02-14-2005, 10:32 PM
Well for H/T the idea is to match the front 3 speakers because as recorded content pans from left to right you want the same sound. There is much to support that notion --- indeed, if you want it to truly match and truly pan across properly you would have three IDENTICAL speakers across the front at the exact same height --- the good new is that with Plasma and LCD and front projectors gaining popularity you can buy the same speakers and place them at the same size -- to me that is the way to watch movies at least the best set-ups I've heard --- luckily that also means that non home theater companies will have speakers that will work in this fashion.

Rear speakers is debatable --- however, it's not a bad idea to match them --- again for best matching you would want identical to the fronts which means the same speaker as the fronts. Some speakers lines though have a similar sonic signature so you could have the B&W 604 in front and the 602 in back and while the back would sacrifice bass the mid and treble would be in the ballpark such that it would be close enough for most people. If it's just movies it matters less because we don;t have the proper reference for rear sound content anyway for things that are not human voices and are not instruments -- everything else explosions breezes, winds, lightsabers, aircraft, bullets, explosions etc in the rear will make little difference whether it's matched

I heard a nice Linn surround set-up professionaly decked out the other day --- if you can match the rears do it --- if not don't sweat it too much.

elbarono
02-14-2005, 11:10 PM
Well, I ordered the front and center speakers from Polk.. The RM6901's and RM6902.. I have a set of Infinity Reference bookshelf speakers that I bought in about '97. I was wondering if I can just use those as the satellites until I get around to ordering new ones.. Not sure whether or not I want to buy the Polk satellites (6903) or something more exotic like a dipole/bipole arrangement.

RGA
02-14-2005, 11:22 PM
The easiest solution is to try it and see how it works --- if you like it keep it --- if not upgrade. There are too many varuiables for people to assume what the sound would be like in your room with your gear with the particular matching you've got. Assuming you set the home theater up correctly and have all the delay effects correct you will probably be fine with the sats you have.

I often tell people the ideal --- but the ideal is usually far far more expensive too. We all have to make a compromise that we can live with.

kexodusc
02-15-2005, 04:42 AM
RGA's gives good advice here. Much to the contrary of what many people believe, the same drivers, and same crossover in in a different box will not sound the same. At the very best the high end will, but the midrange will be subject to so many new variables that all you can hope for is a close match.
Logically, it stands to reason that using the same drivers will get you closer in matching than using other drivers. But throw in the impact room acoustics has on voicing or timbre and it can become a crapshoot.
Generally, it is a good idea to stick with speakers from the same product line, but because of all these variables, there's a great possibility that using other speakers will produce results just as pleasing or even more so. If building a system from scratch, I'd argue it's far easier to just buy matching speakers from the same line than mix and match an infinite number of possibilities. If you already have speakers however, things might become a bit more complicated.

Pretty much everyone agrees that subwoofers are the exception though, get whatever suits you and don't worry about timbre matching. I've had staff at Best Buy try to convince me to that not timbre matching my sub would annoy me to no end. :rolleyes:

markw
02-15-2005, 06:18 AM
... to be in the vocal range. This is where most anomolities will manifest themselves. Noises, gunshots and other such nicities are quite "non'liner" (for lack of a better word) but, we are very prone to notice changes in voices. If the tonal character of a character's voice changes as it traverses from speaker to speaker to speaker, we will most likely fail the ilssusion.

If you can't tell a difference in the voices from your speakers, I'd say you are good to go.

Of course, others may have different opinions

shokhead
02-15-2005, 07:10 AM
When they dont match,you now quick when the sound is going from left to center to right.

Sir Terrence the Terrible
02-15-2005, 10:59 AM
Well for H/T the idea is to match the front 3 speakers because as recorded content pans from left to right you want the same sound. There is much to support that notion

This is the old prologic line of thinking, but does not represent what is currently happening in film mixing today. These days we(re-recording engineers) treat all channels exactly the same. The days are gone that most of the panning and effect placement is only done in the front three speakers. Because of the discrete channels in the rear, you get flyovers, explosions, music, and voices eminating from the rear channels as well as the front, and in full range to boot.. If you have mismatched rear speakers, then every flyover, every effects pan to the rear would exibit a unnatural jump to the rears, would profoundly change the frequency response of the effect, and would not sound cohesive when all channels are played together.

Sounds naturally change frequency response as they move from the front of the head, to the sides. That is because the path to the inner ear is more direct from the sides of the head than it is from the front of the head. Speakers that are timbrally(and frequency response to a degree) matched will not emphasize this natural frequency response shift. Rears that do not match will however.


... to be in the vocal range. This is where most anomolities will manifest themselves. Noises, gunshots and other such nicities are quite "non'liner" (for lack of a better word) but, we are very prone to notice changes in voices. If the tonal character of a character's voice changes as it traverses from speaker to speaker to speaker, we will most likely fail the ilssusion.

Unfortunately voices do not often move from channel to channel for several reason. Mixers mix soundtracks so they play well in most conditions. With the awareness of the way theaters take care of the sound systems, and the proliferation of horizontally based center speakers in the home, we understand that panning voices will not always translate very well in the field. What does get panned are usually wide band effects such as explosions, flyovers and music. And what is becoming more common is the reverse flyover from rear to front.


Logically, it stands to reason that using the same drivers will get you closer in matching than using other drivers. But throw in the impact room acoustics has on voicing or timbre and it can become a crapshoot.

Room acoustic are really only dominate at 200-300hz and below. Above that, the speaker itself dominates what we hear. With timbrally matched speakers, you stand a much better chance of soundfield cohesiveness above room dominate frequencies, and in some cases below. It is not necessary to have the "same" speaker in every quadrant to acheive a timbre match. Speakers that match in frequency response(and voicing) from 80hz up is perfect, as it skips that bass problems you will encounter with large speakers in the rear.

My strong suggestion when putting together a speaker system is to make as few compromises as you can. If you start off with a timbre matched set of speakers, then you have a much stronger chance of having fewer audible conscequencies throughout the most audible and sensitive frequencies to the ear. You also have a much better chance of creating a solid, cohesive acoustical bubble, which IMO is the holy grail of film sound.

RGA
02-15-2005, 11:20 AM
Sir terrance I agree with you 100%.

The idea IMO for a proper home theater system and the only good ones I have heard that I would care to own are when all five speakers (in 5.1) are EXACTLY the same --- those center channels placed above or below a tv is a compromise (whether the drives are the same -- the sound is not) and puny rear speakers are a compromise as well. You either have 5 identical speakers all around or you are down the compromise road where then it's merely a matter of degrees of compromise. The first choice would be to have a so called "timbre matched" rear speaker but I don't think it's critical for movie effects.

We don't have references for the "sound effects" used --- we do have for voice and instruments - my problem with some of the SACD discs I've been hearing lately is that I'm actually seeing the band in front but I'm hearing vocal and instruments coming from behind me. I'm going to listen to a lof SACD to see if this is common -- if it is it's ruinous to music and totally NOT what a live experience sounds like but a pyrotechnic coolness trick.

I suspect one has to really know how to set up a room and cover all the room acoustics bases before putting SACD into the mix --- I heard it done reasonably well at one place and since then it's been terrible every time -- Don Henley's voice should NOT be coming from behind me -- that is not a reverberant field they're capturing - but one they're putting in -- and if the Eagles are a bad example stores should not be using it to demo the technology.

Still SACD is dirt cheap coming with many DVD players and if you have the speakers in place already why not.

kexodusc
02-15-2005, 11:30 AM
Room acoustic are really only dominate at 200-300hz and below. Above that, the speaker itself dominates what we hear. With timbrally matched speakers, you stand a much better chance of soundfield cohesiveness above room dominate frequencies, and in some cases below. It is not necessary to have the "same" speaker in every quadrant to acheive a timbre match.

Sorry Sir T, I rarely know enough to even begin to disagree with you but I'm jumping into the shark infested waters on this one. I have measured with my own spectrum analyzer (and ears) the effects that placing surround speakers close to a wall can and does have in the midrange between 1000 and 4000 Hz...
Generally, a speaker's crossover incorporates baffle step compensation to account for the transition from 4 pi to 2 pi, if placed a foot or two away from a wall (ie: with wall mounting brackets etc) you are causing a lot of reflection of the walls and creating a quasi-4 pi effect which often can lead to a +3 or +4 dB increase in frequencies. This usually a sharp transition too, not gradual. Alot of dipoles/bipoles actually remove the BSC circuitry from the crossovers because of this.
Not to mention the materials of your room and how that can affect the higher frequencies.
While I agree the largest variance will occur in the bass regions, it is definitely not limited there.

Woochifer
02-15-2005, 12:19 PM
It's no trick, it's the truth. You want the timbral characteristics of the speakers to match as closely as possible all the way around. With a timbre matched system, you'll hear a clearcut difference in how much better the encirclement effect gets conveyed, and the depth perception is beyond what can possibly pick up with any two-channel playback.

The center speaker match is always going to be the most compromised one due to the horizontal placement of the speaker. But, in general, if the center speaker match is reasonably close, then it is preferable to use a center speaker. If the match is significantly off, then you're better off without the center speaker. Keep in mind that the center speaker match can vary a lot by series and manufacturer, so definitely give it a listen during your speaker evaluations. If you're deciding between two equally good sets of main speakers, then the match with the center speaker could very well turn into the decding factor.

With the surround speakers, you can get away with using mismatched speakers with a lot of sources. If it's mostly ambient sounds or the soundtrack is mixed with not much activity mixed between the front and rear channels at roughly equal levels, then mismatched surrounds can work okay. But, with multichannel music and an increasing number of newer movie soundtracks, mismatches in the surround speakers will give themselves away very easily, and create an effect almost as bad as if you use different speakers for the L and R channels.

The results can be very impressive if you use timbre matched speakers and position them properly (Dolby's placement guidelines specify raising the surround speakers about 2' above ear level and pointing the surround speakers directly at one another slighly behind the listening position). With movie soundtracks, they convey not only a sense of direction, but can also create a sense of space, whether it's a wide open outdoor scene, or a cramped confining indoor space.

With multichannel music, the variety of 5.1 mixes out there already is very impressive. As with the early days of stereo, you get some good and some bad mixes. But, the best multichannel mixes can create a very broad depth perception with the instruments, and it can also solidify the side imaging in a way that I've never heard from a two-channel setup before. It also keeps the music from sounding overly congested by spacing out the instrumental sounds and using a discrete center channel rather than having to deliberately mix and EQ a phantom center effect. You can only hear the true benefit of 5.1 by going with timbre matched surrounds.

Sir Terrence the Terrible
02-15-2005, 08:26 PM
Sorry Sir T, I rarely know enough to even begin to disagree with you but I'm jumping into the shark infested waters on this one. I have measured with my own spectrum analyzer (and ears) the effects that placing surround speakers close to a wall can and does have in the midrange between 1000 and 4000 Hz...

That Kex would not be the norm amoung surround speakers. Most surround speakers are already designed with a boundary compensation in mind. Most speaker designers know that their surrounds will be placed in close proximity of the walls. I do not think any smart speaker designer would design their surrounds for free air mounting, or the woofer would have to be VERY substantial to deal with the inherent roll off of free air mounting at low frequencies. If you measure the boundary compensation in the THX ultra2 processing, you will find that it really only effects the frequencies between 200-500hz. The wavelengths at high frequencies are too short to support much audible enhancement at the ears. Also, speakers are very much more directional at 2-4khz, so the walls behind the tweeter have much less influence on the tweeter sound than on the woofers.



Generally, a speaker's crossover incorporates baffle step compensation to account for the transition from 4 pi to 2 pi, if placed a foot or two away from a wall (ie: with wall mounting brackets etc) you are causing a lot of reflection of the walls and creating a quasi-4 pi effect which often can lead to a +3 or +4 dB increase in frequencies.

Actually if you understand the dispersion characteristics of high frequencies, you will notice that any speaker who's tweeter interacts with the walls(dipoles and bipoles) usually have a reduction in high frequencies due to the longer path the sound must take to reach the ears. There is also the effect the air has on high frequencies that must be taken into consideration. You will find baffle step compensation on front speakers, not rear speakers.


This usually a sharp transition too, not gradual. Alot of dipoles/bipoles actually remove the BSC circuitry from the crossovers because of this.

This is so because of the reduced sensitivity to high frequencies in areas behind the pinna(the flaps of our ears)


Not to mention the materials of your room and how that can affect the higher frequencies.
While I agree the largest variance will occur in the bass regions, it is definitely not limited there.

Most rooms seriously lack asorption and diffusion. Most room reflections at high frequencies are so closely spaced and low in amplitude, that we cannot clearly make them out from the speakers direct output.

It is widely understood in acoustical science that boundary effects of walls and ceiling diminish with the increase in frequency. This is because speakers become more directional with the increase in frequencies, and why most boundary reflections are audible at low frequencies where speakers become less directional.

kexodusc
02-16-2005, 06:18 AM
I've only been at this acoustic physics stuff for a year or so now and only have a few books and articles under my belt, but I think I see the root of our disagreement here...


That Kex would not be the norm amoung surround speakers. Most surround speakers are already designed with a boundary compensation in mind. Most speaker designers know that their surrounds will be placed in close proximity of the walls. I do not think any smart speaker designer would design their surrounds for free air mounting, or the woofer would have to be VERY substantial to deal with the inherent roll off of free air mounting at low frequencies.

I agree with what you're saying here about surround speakers and boundary compensation, and understand your point...I guess my issue is that many, many home theaters (mine included) are composed of identical speakers, and these are quite often bookshelfs or mix of floorstanders and bookshelfs that are designed to be front main speakers, not surround channel speakers only. For example, say I had a Paradigm home theater...and I used Studio 40's for mains and surrounds (or even Studio 20's). These are built to be main speakers, not surround speakers like the Studio ADP dipoles are...there would be no boundary compensation, I wouldn't think. I've certainly never added any boundary compensation circuitry to any crossover I've built. (mind you I've only build 5 different xo designs).



If you measure the boundary compensation in the THX ultra2 processing, you will find that it really only effects the frequencies between 200-500hz. The wavelengths at high frequencies are too short to support much audible enhancement at the ears.

There you go using big words and stuff I don't know anything about again. :D
Okay, really high frequencies are directional...I'm with you so far...



Also, speakers are very much more directional at 2-4khz, so the walls behind the tweeter have much less influence on the tweeter sound than on the woofers. ..



Actually if you understand the dispersion characteristics of high frequencies, you will notice that any speaker who's tweeter interacts with the walls(dipoles and bipoles) usually have a reduction in high frequencies due to the longer path the sound must take to reach the ears. There is also the effect the air has on high frequencies that must be taken into consideration. You will find baffle step compensation on front speakers, not rear speakers.

Yes, I know as the frequency gets higher, the sound is emitted in a forward pattern.
I guess it comes back to my statement about people using main speakers as surrounds. Baffle step diffraction issues are usually most problematic from 200Hz to 2500 Hz (with diffraction issues going up higher)...typically speakers will see a - 6 dB drop by the time they hit the bass region, and are compensated for this...If you start placing a bookshelf speaker with baffle step compensation circuitry near a wall (especially wall mounted) you are basically undermining the compensation and your midrange response will vary significantly from 200Hz up to the 2 kHz region...again, assuming a main speakers are used all around. We might only be talking about a +/-2 dB difference for the frequencies in this region, but that's way more than it takes to change the "timbre" or voicing of the speaker from that of the front mains. Then again, for ambient cues and effects in movies it's probably not that big a deal. But that's why I don't place too much emphasis matching the surrounds with identical speakers.

I guess this would be a good argument in favour of using dipole/bipole speakers or on-walls for surrounds instead of main speakers. Damn I wouldn't want to begin to try to design a crossover, find a good cabinet volume etc, to come up with a direct radiating speaker that had a close timbre match to main speakers. My rear speakers are stand mounted a good distance away from walls but the side surrounds I have are currently bracket mounted, a necessary evil until I get up the nerve to build more stands...assuming the lady of the house lets me place 2 more speakers on our floorspace.



This is so because of the reduced sensitivity to high frequencies in areas behind the pinna(the flaps of our ears)

Good point...intuitively though, it would have to be a combination of both factors...the lower sensitivity of ears to high frequencies, and the reduced baffle step loss from the speaker being designed for wall placement, wouldn't it?
Here's a question then, what good is a manufacturers specs about a bipolar/dipolar speakers performance in an anechoic chamber then? Why bother? And if the response looked great in the anechoic room, wouldn't that be cause for concern?



Most rooms seriously lack asorption and diffusion. Most room reflections at high frequencies are so closely spaced and low in amplitude, that we cannot clearly make them out from the speakers direct output.
Here I think you can teach me a lot. My new room is basically wood flooring, mostly bare walls, and a few area rugs, pictures, and furniture I've placed at reflection points. Yet it is noticeably lively and brighter sounding than my old carpeted room...a bit too echoey too, but hopefully the new furniture and piano I've ordered will fill it up some. There must be a point where even the directional frequencies are bouncing off rear walls onto the side walls, floors etc...the significant other would like to carpet the floor or get the wood flooring refinished/polished to bring it up to snuff...I'm assuming I'd be further ahead acoustically to carpet...any thought?