Review of the Salk Veracity HT3 loudspeaker

[Dennis Murphy]

I know you asked Sean specifically but in your question "But that raises another question--how does the ear process more direct arrival times vs. more delayed room reflections with different frequency profiles?", everything I have read states that the sound that gets to the ears first determines the localization cues.

I need to dig out my Master Acoustics Handbook that's almost 30 years old. Would it be worth buying a new copy so I can be more up to date?

That's what I've heard and read too. But what does "dominate" mean? Totally obscure other information? I doubt that. The first arrival certainly helps us localize a sound. But the quality of the sound clearly depends on more than the first arrival. It's just a question of how much.

 

[gene]

Hi Sean Thanks for the many useful comments. One subject in particular has always bugged me as very minor speaker designer. I certainly agree that smooth on and off-axis response is a top priority, and I work as hard as I can to get rid of off-axis flares in the tweeter response. But if we confine ourselves to vertically arrayed driver configurations (and that's the vast bulk of the market), then we're really only talking about smooth off-axis horizontal response. For the most popular crossover designs, vertical response will be anything but smooth in the crossover region(s), and that should be reflected in the power response. But that raises another question--how does the ear process more direct arrival times vs. more delayed room reflections with different frequency profiles? A really good model of human hearing is critical to understanding just how important off-axis (vertical and horizontal) deviations are. All of which is to say that, although we have a pretty good understanding of what goes into accurate speaker design, we're still a good piece short of a complete understanding.

Dennis;

Have you read Dr. Tooles book on Acoustics which deals with human perception of early reflections among other things?

 

[Dennis Murphy]

Dennis;

Have you read Dr. Tooles book on Acoustics which deals with human perception of early reflections among other things?

Well, I'm making my way through it. But so far nothing that would answer my precise question, particularly in perception of tonal colorations in the 2k-4k region.

 

[tonmeister]

We are now off topic. However I should state right away that I do not listen to pop music.

For classical music minimalist mic techniques sound the best on good speakers. However minimalist mic techniques can sound very poor on poor speakers I have noticed.

I made hundreds of recordings for the Public radio station over many years.

The vast majority were made with my Neumann SM 69 FET stereo condenser microphone. This results in an intensity difference recording rather than the more usual phase difference recording.

I used to used either crossed figure of 8 or the MS technique, The the figure of eight sideways to the players, and the other capsule set to omni. I used the MS most often as mono radio listeners would have a perfect mono signal and the figure of 8 capsule is canceled in mono.

I think this type of technique gives the most accurate playback on good speakers in rooms of any size.

You can download a CD of a recording of mine using this technique here.

Or if it is easier I can send you one Sean.

Thanks. I will check it out. There has been some recent research on the perceptual attributes of different surround microphone techniques including spaced and coincident techniques like MS. For piano music at least, spaced omni's/cardioids seem to be preferred by listeners since they produce better spatial (envelopment and apparent source width) and timbral properties, which seem to withstand off-axis listening positions.

The least preferred microphone methods were coincident techniques using Ambisonics (no surprise there). I actually participated as a listening subject in the microphone comparison at McGill and the Ambisonics was clearly the worst one of the choices I heard.


See: the series of papers written by Sungyoung Kim et al.:
for example: http://www.aes.org/e-lib/browse.cfm?elib=13853
and also http://www.aes.org/e-lib/browse.cfm?elib=14732

 

[tonmeister]

That's what I've heard and read too. But what does "dominate" mean? Totally obscure other information? I doubt that. The first arrival certainly helps us localize a sound. But the quality of the sound clearly depends on more than the first arrival. It's just a question of how much.

The direct sound or first arrival does determine where you localize the sound as explained by the precedence effect.

However, the early reflections contribute towards the perceived loudness, intelligibility, timbre, and spatial attributes - listener envelopment (LEV) and apparent source width (ASW) and distance of the source(s), depending on their level, arrival time, angle of incidence, and spectrum.

 

[Dennis Murphy]

The direct sound or first arrival does determine where you localize the sound as explained by the precedence effect.

However, the early reflections contribute towards the perceived loudness, intelligibility, timbre, and spatial attributes - listener envelopment (LEV) and apparent source width (ASW) and distance of the source(s), depending on their level, arrival time, angle of incidence, and spectrum.

I think we can agree on that. But ideally we would like a computer model of a listener to help us with design trade-offs. If I achieve flat on-axis response at the crossover point I like, but there's a flare at 3 khz 40 degrees off axis, it doesn't necessarily follow that I should change my crossover frequency (or slopes) or build in a 3 kHz dip on axis. It depends on the relative importance of the on-axis and off-axis arrivals as perceived subjectively by the listener. Fortunately, modern design software allows you to hear crossovers before you build them, so a lot of experimentation is possible.

 

[highfigh]

That's what I've heard and read too. But what does "dominate" mean? Totally obscure other information? I doubt that. The first arrival certainly helps us localize a sound. But the quality of the sound clearly depends on more than the first arrival. It's just a question of how much.

Who mentioned 'dominate'? I see that used when many people mean 'dominant'. I don't know that the later arrival information is obscured but it is ignored to a small degree if it's at a low enough SPL. I'm still looking for my handbook- I know this is in there but it has been a long time since I had an acoustics class. I had some bizarre phantom sounds here because of the longer reflections. Sometimes, the source seemed to be about 10' outside of the right speaker but that stopped when I made my panels.

 

[highfigh]

Well, I'm making my way through it. But so far nothing that would answer my precise question, particularly in perception of tonal colorations in the 2k-4k region.

Some of what I do remember from the psycho-acoustics section is that certain frequencies cause the source to be far to the side (4KHz) or directly above (8KHz) but I don't remember a reference to 2K. I went in for a hearing test and since I had read this section not long before, I listened for specific sounds and as read, 8KHz did sound like it was directly overhead and it doesn't matter if it's coming in one ear or both. I asked if that was 8K and after looking at me like I had three heads, she asked how I knew. In listening to speakers and using equalizers over the years, I have noticed that when there's a rise in the 6K-8K range, the stereo image does seem taller.

 

[AJinFLA]

The majority of loudspeakers we test are what we produce (direct radiator with and w/o constant DI horns) and compete against. I have tested some electrostatic, magnetic planar, distributed mode types, and models with unusual directional characteristics (line sources, 180 degree radiation via acoustic lens, omni-directional, bi-pole, dipole).

Well, I was asking about the very specific case of piston source first order gradients. Commercial examples such as the Jamo R-909, ME-Geithain ME220k and Gradient Revolution (There are direct order types like the Linkwitz Orion and Musicanddesign NaO). Are the Harman products, listening panel judged against anything like these?

I'm not quite sure what you mean.

I was referring to this

It's only below the transition frequency (300 Hs or so depending on the size of the room) where the room largely determines how good the speaker sounds, and there judicious placement, equalization and multiple subwoofers can fix those problems.

How can multi (source) subs fix problems (via averaging) up to 300hz without localization problems? If the subs can only be used <100hz or so to avoid localization, then isn't the eq from 2-300hz going to be (largely) single point correction?
Also, as already mentioned, if the reverberant field is so important, why would you want polar smoothness only horizontally? Why not the entire polar field (point source, rather than vertically stacked decorrelating drivers)?

cheers,

AJ

 

[tonmeister]

Well, I was asking about the very specific case of piston source first order gradients. Commercial examples such as the Jamo R-909, ME-Geithain ME220k and Gradient Revolution (There are direct order types like the Linkwitz Orion and Musicanddesign NaO). Are the Harman products, listening panel judged against anything like these?
I was referring to this

No, I have not tested those specific models.

How can multi (source) subs fix problems (via averaging) up to 300hz without localization problems? If the subs can only be used <100hz or so to avoid localization, then isn't the eq from 2-300hz going to be (largely) single point correction?
Also, as already mentioned, if the reverberant field is so important, why would you want polar smoothness only horizontally? Why not the entire polar field (point source, rather than vertically stacked decorrelating drivers)?

cheers,

AJ

Now I understand your question. We don't recommend you use multiple subwoofers beyond about 120-200 Hz for reasons you suggest.

Equalization above that frequency will either be based on a single seat or a spatial average made across several seats. You can't make every seat perfect with multi-point equalization, contrary to what some people may suggest.

I can think of several reasons why wider uniform horizontal dispersion might be more important than uniform vertical dispersion: the most important reflections are off the side walls since they contribute spaciousness. The ceiling/floor reflections don't contribute to spaciousness, and the carpet (if you have one) will absorb much of the first floor bounce at higher frequencies. But I am speculating because I have not done an experiment where only the vertical radiation was manipulated while other factors were kept constant, and to my knowledge, no one has yet.

I have tested some loudspeakers with co-axial/coincident drivers that meet the criterion you suggest, yet there was nothing special about them that made them sound superior in any way. Some of them were downright awful.

It seems that loudspeaker compromised of well-designed vertical driver arrays seem to do pretty well. The CBT arrays developed by Keele et al, are a good example where good horizontal and vertical dispersion can be achieved. They also have desirable directional properties that make them good candidates as surround speakers, as discussed in Floyd Toole's book.

 

[skers_54]

We did a simple experiment in my physiology lab that showed that human hearing is much more attuned to direction in the horizontal plane. We basically had someone blindfolded and had another person snap at different positions relative to the test subject. Everyone tested could locate the sound to within a few degrees on the horizontal plane. On the vertical plane, however, the accuracy was greatly diminished and was often interpreted as originating from the horizontal plane. As the sound source got closer to directly above the head, it confused with coming from behind the subject nearly half the time.

 

[TLS Guy]

I know you asked Sean specifically but in your question but if I may, "...that raises another question--how does the ear process more direct arrival times vs. more delayed room reflections with different frequency profiles?", everything I have read states that the sound that gets to the ears first determines the localization cues.

I need to dig out my Master Acoustics Handbook that's almost 30 years old. Would it be worth buying a new copy so I can be more up to date?

I think you are correct. That is why I think intensity techniques produce a much better sound stage than phase difference techniques. The problem is that omni microphones as a group tend to have a smoother frequency response than microphones with directionality. For an intensity microphone technique you need mic capsules with directionality.

If you take two omni mics and space them (commonly done), what you really have is two mono sources. So if you put two spaced omni mics in front of a piano, you have two mono sources of pretty much the same amplitude. The only difference is phase, and I don't think that is an ideal fit for human hearing localization.

 

[TLS Guy]

Well, I'm making my way through it. But so far nothing that would answer my precise question, particularly in perception of tonal colorations in the 2k-4k region.

I think the need for crossovers is a huge pity. Those familiar with my posts know I'm an unrepentant full ranger at heart. I had a lot to do with the Jordan watts full range driver after Ted Jordan and Leslie Watts parted company. I was to a large extent responsible for the MK 3 unit.

The JW driver is power limited but in many ways the unit in a TL is a revelation.

I have a few cones still and a lot of MK IIs. The biggest problem with the unit is a suspension resonance at 375 Hz and its second harmonic at 750 Hz. If I had a partner, to help fabricate a new suspension and spin more 6 gm 4" aluminum tractrix cones, I'm convinced a high performance wide band mid range driver could be built, with a bandwidth four or five octaves wide and may be more, and no break up modes to cause problems.

I think we need speakers with drivers that can have a single driver handle the whole speech discrimination band where I believe the ear is highly sensitive to crossover ills and other aberrations. I believe the 1 to 4 kHz pass band is the target area where the ear is highly sensitive to problems in loudspeakers. Actually I think tolerance is pretty close to zero for errors here.

 

[TLS Guy]

Thanks. I will check it out. There has been some recent research on the perceptual attributes of different surround microphone techniques including spaced and coincident techniques like MS. For piano music at least, spaced omni's/cardioids seem to be preferred by listeners since they produce better spatial (envelopment and apparent source width) and timbral properties, which seem to withstand off-axis listening positions.

The least preferred microphone methods were coincident techniques using Ambisonics (no surprise there). I actually participated as a listening subject in the microphone comparison at McGill and the Ambisonics was clearly the worst one of the choices I heard.


See: the series of papers written by Sungyoung Kim et al.:
for example: http://www.aes.org/e-lib/browse.cfm?elib=13853
and also http://www.aes.org/e-lib/browse.cfm?elib=14732

I think the Calrec Ambisonic microphone with its UHJ code/encode system was ahead of its time, and they payed the penalty. However I have some two channel mix downs on Nimbus that sound very good. Unfortunately I have been unable to find, any of the Nimbus UHJ DTS recordings released on DVD A.

I do have some of the four channel Philips four channel Quadraphonic recordings made with a coincident four capsule microphone. These recordings on the Pentatone label are excellent.

 

[highfigh]

I think you are correct. That is why I think intensity techniques produce a much better sound stage than phase difference techniques. The problem is that omni microphones as a group tend to have a smoother frequency response than microphones with directionality. For an intensity microphone technique you need mic capsules with directionality.

If you take two omni mics and space them (commonly done), what you really have is two mono sources. So if you put two spaced omni mics in front of a piano, you have two mono sources of pretty much the same amplitude. The only difference is phase, and I don't think that is an ideal fit for human hearing localization.

I'm really sensitive to phase problems and it's not just low frequencies. Some speakers are just plain hard for me to listen to and it's not a comfortable feeling. I know someone with a small recording studio and he asked if I was able to borrow the RTA from work because he was having a problem with his mixing, but couldn't identify it. Before we set the mic up, he played something and I told him that he had something out of phase and I asked if he had recently replaced a horn driver. He had replaced one but he didn't do it himself and the other guy didn't think it would be a problem. The RTA showed a deep V at the crossover frequency and this was over 20 years ago, when a good RTA was a lot harder to come by at a reasonable price.

Highly directional mics are great for close-mic-ing but never for capturing the sound in a room. I recorded the band some friends were in and it sounded very real, but the room was terrible, acoustically. That's the only reason it sounded real- the side reflections weren't attenuated much, if at all. With headphones, the sense of space made it very much like the room that night, especially if I closed my eyes. As far as I know, it was one of the first digital recordings done here- the store where I worked sold the Sony PCM-F1 and I did this within a week of us getting it. I also had a customer who recorded Classical and Jazz performances, using Studer recorders and Schoeps mics.

 

[Matt34]

I have nothing to add expect I'm really enjoying this thread....even if it is way off topic.

 

[TLS Guy]

I'm really sensitive to phase problems and it's not just low frequencies. Some speakers are just plain hard for me to listen to and it's not a comfortable feeling. I know someone with a small recording studio and he asked if I was able to borrow the RTA from work because he was having a problem with his mixing, but couldn't identify it. Before we set the mic up, he played something and I told him that he had something out of phase and I asked if he had recently replaced a horn driver. He had replaced one but he didn't do it himself and the other guy didn't think it would be a problem. The RTA showed a deep V at the crossover frequency and this was over 20 years ago, when a good RTA was a lot harder to come by at a reasonable price.

Highly directional mics are great for close-mic-ing but never for capturing the sound in a room. I recorded the band some friends were in and it sounded very real, but the room was terrible, acoustically. That's the only reason it sounded real- the side reflections weren't attenuated much, if at all. With headphones, the sense of space made it very much like the room that night, especially if I closed my eyes. As far as I know, it was one of the first digital recordings done here- the store where I worked sold the Sony PCM-F1 and I did this within a week of us getting it. I also had a customer who recorded Classical and Jazz performances, using Studer recorders and Schoeps mics.

I don't like directional microphones unless you can open the back. The old STC microphone that is now the Coles 4083 is a fantastic microphone. That microphone has been virtually unchanged in over half a century!

I think a figure of 8 microphone works very well in a big space with good acoustics. If I use omnis, I usually adopt the baffled omni technique.

 

[TLS Guy]

I have nothing to add expect I'm really enjoying this thread....even if it is way off topic.

Glad you are enjoying it. Possibly not that off topic. The front and end of the musical chain is the most problematic. Microphones have to be chosen with as much though as speakers, and good ones cost as much as fine speakers. They have to be used with as much insight also.

 

[WmAx]

Glad you are enjoying it. Possibly not that off topic. The front and end of the musical chain is the most problematic. Microphones have to be chosen with as much though as speakers, and good ones cost as much as fine speakers. They have to be used with as much insight also.

Oh, how the times have changed.

I just don't see where you need to spend much for superb mics today. Pick carefully, and you can get top notch mics for $100. Take the Behringer B-5 for example. It has omni and cardoid capsules, and both have extremely low distortion, low noise, extreme linear response (+/- 1.5dB up to 16-20Khz) and very high dynamic range capability (140dB with no clipping;1%THD point). Now, I don't know of any mic truly comparable to the B5 until you get a far higher price point, so it is of a unique value.

If you want to replicate the non-linear, high dollar Nuemans, etc.; this is also no problem. There are a few choices for about 200 dollars that will replicate the sound characteristics to where you simply will not be able to tell the difference by the recording itself.

Of course, I'm sure it's no secret that these high value mics come from Asia. One company from the East, until recently, was making EXACT copies down to the original manufacturing specifications, of some super high priced vintage famous mics(well past their patent expiration dates), and selling them for just a couple of hundred dollars.

-Chris

 

[GO-NAD!]

I have nothing to add expect I'm really enjoying this thread....even if it is way off topic.

Me too! Most of it is over my head though....