Narrow vs. wide baffles... which one is the best?

[fjhuerta]

Hi,

As I was designing a new speaker, I decided to use the narrowest baffle I could use. In fact, the edges of the 7" woofers are flush with the baffle walls, without a round-over. I used the FRD Baffle Diffraction Simulator, and everything looked fine.

Then, as I was investigating more on narrow baffles, I found this:

https://www.htguide.com/forum/showpost.php?p=292359&postcount=5

Well...

Narrow baffle and deep cabinets are prone to "boxy" sound if they're not done carefully. Depending what you mean by "narrow"

So, more specific parameters would be helpful in pointing you one way or another.

C

http://www.ejjordan.co.uk/faq.html#faq16

Q: In your VTL box design, why is the JX92S fitted in the wide face when it is common knowledge that the box should be as narrow as possible?

A: 'Common knowledge' and scientific fact are often very different. The narrow front face is a fashion concept supported by some very questionable marketing rational. The indisputable scientific fact is that the ideal mounting for a loudspeaker is an infinitely large flat baffle and this is the concept used for all loudspeaker analyses. A wide baffle always sounds better.(emphasis mine)

And...

http://www.diyaudio.com/forums/showthread.php?postid=448157#post448157

Only very briefly. Narrow baffles make the sound comparably thin EVEN IF you apply baffle loss correction due to poor wavelaunch and altering directivity throughout the format range. In addition, any diffraction problems by narrow baffles fall outside the HAAS window and are comparably much higher in level, meaning imaging is very much damaged and made un-natural.

A wide baffle moves the diffraction problems more to within the HAAS window (after mor ethan 1mS) and due to a larger distance to the edge the level of diffracted sound is lower. A wide baffle will support "wavelaunch" (remember, waves are spherical) to a much lower frequency and give better support to lower formants.

Acoustically speaking the modern narrow frotn tall tower speakers are about as bad as it gets, unless you build a linesource. (again, emphasis mine).

Sayonara

So far, I haven't found anything on the web that indicates narrow baffles are better, except for some comments regarding better imaging.

Is this correct? Are narrow baffles not recommended?

[dlr]

Wide vs. narrow for the face is primarily aesthetics

I have seen no references to any scientific studies indicating that narrow or wide is perceptibly better. I've heard excellent systems using both. Systems such as dipoles and cardioids use a wider baffle specifically to achieve a desired response. The narrow M/T boxes I've used in my 3-way systems are, I believe, as good as any wide baffle, all else being the same (that's a BIG caveat, of course).

There are reasons to use a wider or more narrow baffle as alluded to earlier. It's related to internal dimensions and resonance modes. The depth needs to be enough to allow for sufficient damping in a closed box system, maybe most important for midrange drivers in a closed box. For ported designs some optimum relationship between the three dimensions is required and since there is insufficient damping material for full damping, the dimensions are even more critical.

As for the baffle front dimensions, narrow is nothing more than current fad due to usage with HT systems with the exception of the choice being integrated with the design for improved off-axis response or requirements for a dipole/cardioid. If the drivers and their directionality are taken into account, the diffraction effects can be ameliorated with baffle dimensions (height and width, not just width), driver placement and crossover Fc selection. It's a complicated set of choices when all aspects are considered.

Regarding the idea that the step location has a huge impact in the perception, I don't believe any of it. I've seen no evidence of any kind other than anecdotes. The step in a 3-way in particular can have the most impact, but this would primarily be related to the selection to reduce excursion requirements for one driver or another. Tweeters and midranges on wider baffles benefit by maintaining the low end in 2-pi space to lower frequencies, tweeters benefiting most. I'm toying with the idea of making a much wider baffle for the midrange myself. I like small midrange drivers. They are limited in the low end and would benefit from a wider baffle.

It always possible to use the crossover to correct for the baffle step in a properly designed 3-way. The small difference is the point at which the system goes from 2-pi to 4-pi. Wider baffles place this lower, so theoretically the power response will change over that of a wider baffle. Again, I've seen nothing to indicate that this has a real perceptible impact. It's most likely the change in other diffraction effects coupled with altered requirements of the drivers, i.e. more displacement requirements for the drivers on narrow baffles (assuming the same driver is used, of course).

In fact, offset drivers on a wide baffle placed to ameliorate diffraction effects are in something of both areas. The near edge acts as a narrow baffle, the far edge acts as a wide baffle. Taller baffles vs. short ones in height have the same effect, so any discussion of wide vs. narrow has way to many variants to be definitive. The only possible exception would be a pseudo-infinite baffle, such as mounting the drivers on a wall.

Dave

[fjhuerta]

Thanks, Dave. I, myself, have never found anything regarding narrow baffles being a bad idea for design - only that the frequency at which the 4Pi - 2Pi transition begins. And I can imagine why this could be a problem if diffraction is also properly accounted for during crossover design.

[jkrutke]

Javier,

Your quote from DIYaudio is completely wrong. (common problem over at DiyAudio, home of EnABL and other bad ideas) Allow me to describe why.

In my eyes, cabinet width affects 2 main issues for the woofer: 2pi to 4pi space corner frequency and diffraction ripple magnitude. The 4pi space corner frequency is not an issue to me, it's easily resolved in the crossover wherever it happens to fall. It's the diffraction ripple magnitude that's an issue, and I don't really see a lot of people being as concerned about it as I am. The magnitude of the ripple is essentially based on the ratio of piston diameter to enclosure width. The closer the piston diameter is to the enclosure width, the smoother the response.

For example, take a look at the raw woofer response of the Scan Speak 15W8530K00 in my ZD5 design. It's a 7" wide enclosure, relatively narrow for this woofer size.



There's hardly any diffraction ripple at all in the woofer response. And of course, on an infinite baffle, the woofer's response has no ripple other than it's own linear distortion issues. Now, put that driver in a 12-14" wide cabinet and see what the response looks like. It's pretty bad.

Of course, this is just in reference to a woofer / lower mid. Mids or tweets will always have ripple problems, some of which can be resolved on axis by offsetting. This comes at the cost of non symmetrical horizontal polar response. But mostly, I like to keep a lot of ripple out of the 500-1500 band because that's where I think it's the most audible. In an average 2-way, that's always going to be in the woofer's area. In an average 3-way, you're screwed with the mid unless you do what Dave does, separate tweeter mid enclosures. Or a tapered/mitered Avalon style box, or some other solution that isn't a big monkey coffin.

I prefer narrow baffles myself, though if I did a wide baffle, it would be really wide and probably curved back too. So in summary, I generally recommend as narrow as possible or as wide as possible, but nowhere in between. Considering most people think wide enclosures are ugly, that leaves the slim option.

I have a 13" wide 3 way tower enclosure sitting in my basement for 5 years now, unused. It was for a largish woofer (the cabinet was set up for 8 to 10" woofers) with a 4 inch midrange. That presented the worst case enclosure - midrange ripple was horrible, and offsetting the mid only helped on axis.

I've really got no new ideas to present here. Everything I mention can be seen for yourself by playing with a diffraction simulator.

[norcad]

Except for a pair of CV AT100 i owned about 20 years ago, all my speakers had narrow baffle. But when I started my first DIY project it was an wide baffle. The reason for this was several, first of all I didnt want a speaker that was to deep and had to be placed in the middle of the room, and second I wanted a high sensitivity and therefor without BSC.
My Seas 3-way is about 21 inches in total widh and that pushed the baffle step down to about 250-300hz, where I put the xo point between the mid and woofer. It is biamped with active xo for the bottom and passive xo for the top. The baffle has a rather huge radius of about 5 inches
The benefit is a tweeter without any diffraction, a higher sensitivity and like Troels Gravesen says; "it comes with a room"!

The drawback is like jkrutke says; a ripple in the midband. But this riple is steep and rather easy to deal with by using a notch filter.

You can look at some measurements and pictures here: http://www.htguide.com/forum/showthread.php4?t=30421

[Curt C]

In my humble opinion, they can both be best depending on your application

If the baffle has no edge treatments, the wider baffle will have more, diffraction artifacts, but they model to have less response deviation from ideal per artifact. The narrow baffle will have fewer artifacts, but they will be of higher deviation.

The ubiquitous ‘bump and dip’ response of the driver just above the 2pi transition frequency varies with frequency, with the largest response variation occurring at lower frequencies on a wide baffle, but is nearly absent on a narrow baffle. The wider baffle better ‘supports’ lower frequencies and may in some cases have merit as it can increase the usable bandwidth of a driver. It can also be utilized in to ‘fill in’ a dip in the driver response, assuming said dip and said bump correlate to a reasonable baffle size.

The larger concern here IMO, is not baffle size, but edge treatments. All baffles respond well to larger chamfer or roundovers, with again the wide baffle having more ripple, but smaller response variations than a narrow baffle. In either case, the larger the edge treatment the better. Ripple typically models to be less than 1dB with a 1.5” roundover, but closer to +/- 2 dB with only a .75” roundover.

As to whether baffle size affects imaging, I have no documented evidence that it does or does not. Of course the wave propagated off of the baffle edge would be delayed in time, and could detrimentally affect the perception of transient detail. The larger the baffle, the longer the delay would be. This supposition would tend to favor the narrow baffle, but only if diffraction effects were prevalent. That is, baffle edge treatments were not applied.

I suspect a large curved back baffle with good edge treatments would image just as well as a narrow baffle with good edge treatments. I can say that I’ve found a large sphere to have excellent imaging.

C

[Saurav]

Of course the wave propagated off of the baffle edge would be delayed in time, and could detrimentally affect the perception of transient detail. The larger the baffle, the longer the delay would be. This supposition would tend to favor the narrow baffle

Thorsten's quoted post seems to suggest the opposite. If I understood it right, he's saying that there's a lower limit to the Haas effect, such that if the diffracted sound arrives too soon after the original sound, the brain won't be able to separate them. But delay the diffracted sound by about 1ms, and now the precedence effect comes into play, and the brain ignores the second arrival.

Quoting from Javier's quoted post:

A wide baffle moves the diffraction problems more to within the HAAS window (after more than 1mS) and due to a larger distance to the edge the level of diffracted sound is lower.

[augerpro]

Everything I've read suggest much longer times before the secondary sound is perceived to be separate from the direct sound, like >10ms.

[Saurav]

Here's my understanding.

> 30-40ms (according to Wikipedia), the brain perceives the 2nd sound as a separate sound, i.e. an 'echo'.

< 30-40ms, the brain filters out the second sound, so you only hear the first.

What I think Thorsten is saying is that there's a lower limit to that window, so if you get < 1ms, the brain hears both sounds as one, which would (I think) result in fuzzier imaging/localization.

Of course, psychoacoustics is totally not my field, so take this with a big grain of salt. All I'm trying to explain is how I interpreted the post that Javier quoted.

[Saurav]

Hmm, this pretty much matches what Zaph is saying too, I think. Check my math here. Speed of sound = 340m/s, = 340mm/ms. So for the sound to be delayed by >1ms, the reflection needs to be > 340mm away. I'm guessing the delay time would be equal to 1/2 the baffle width, so you need a baffle that's > 680mm, which is 26"?

Which is what Zaph said - go really wide, or go narrow.

[Curt C]

Everything I've read suggest much longer times before the secondary sound is perceived to be separate from the direct sound, like >10ms.

I agree. Experience has proven that placing a speaker too close to a reflective side wall will artificially widen the soundstage outside the speaker. In effect, the ear perceives the reflected sound from the wall and the direct sound from the speaker to be the same, and not interpreted as separate direct/ambient sound. This will cause an audible smearing of the perceived position of the sound. The delay caused by the distance between the speaker and the wall would be much longer than a 1 ms, and would represent a baffle width much larger than most large baffle speakers.

C

[augerpro]

Hmm, this pretty much matches what Zaph is saying too, I think. Check my math here. Speed of sound = 340m/s, = 340mm/ms. So for the sound to be delayed by >1ms, the reflection needs to be > 340mm away. I'm guessing the delay time would be equal to 1/2 the baffle width, so you need a baffle that's > 680mm, which is 26"?

Yeah that's how I'd look at it. Basically you need much more time than even a wide baffle width would give you to keep the direct and secondary sounds "separate" according to the brain. Given that, a smaller baffle would at least keep the direct and secondary sound closer in time than a wide baffle. On paper that would be better. It would be an interesting experiment to see how it really sounds.

[Saurav]

I guess that's one 'good' thing about dipoles - the baffle width has other jobs to perform, so this is one less thing to worry about

[Dennis H]

I agree with Curt. Anti-diffraction treatments (roundovers, chamfers, facets, felt, etc.) are more important than baffle width within the range of widths you'd be likely to build. If you can reduce the amount the edge radiates, that's more important than a fraction of a msec difference in when the radiation occurs. Of course, wider cabinets give you more room for edge treatments, like what Norcad did.

As an aside, I've been reading John Kreskovsky's dipole stuff today. He's added a lot since the last time I looked. For a dipole mid and tweeter, there are definite power response advantages to keeping the baffle as narrow as possible, like no wider than the shortest wavelength the driver plays. The tradeoff (there are always tradeoffs in engineering ) is you sacrifice low-end SPL with narrow baffles.

[Sakura]

I'm thinking of trying some ~ 2" radius roundover (on the baffle edges)cabinets and I think I've got a neat idea how to make it easy, if it works out, I'll post the results...

[Alaric]

I'm thinking of (ok , working up to) building a pair of speakers , and this debate intrigues me. My current speakers are narrow baffle , and I really like them. I also owned a pair of wide baffle , shallow box , speakers I really liked (A/D/S L1530). Are there any clear advantages of one over the other?

[David G]

A wide baffle has another disadvantage that has not been mentioned yet - resonance.
The sound output of a panel is directly proportional to its area, and it's far easier keeping a small panel inert and rigid, than a large one.
The front panel of the speaker is especially critical because the drivers are mounted on it, and it points directly at the listener.

[Sakura]

When making a narrow baffle, I really don't think it's worth it to sacrifice the roundover in order to make the baffle narrower. I would say that 3/4" roundover is the "sweet spot" where you get the most benefit, but diminishing returns if you go larger.

But, try to do at least a 1/2" roundover. 1/2" router bits are inexpensive, and not too hard to use in a hand held router, as long as you are careful to keep the router held flat on what you're cutting, don't let it tip or you'll get a gouge.

[Curt C]

A wide baffle has another disadvantage that has not been mentioned yet - resonance.
The sound output of a panel is directly proportional to its area, and it's far easier keeping a small panel inert and rigid, than a large one.
The front panel of the speaker is especially critical because the drivers are mounted on it, and it points directly at the listener.

Good point David. Extra bracing is a must for larger baffles.

One small point: The resonant frequency is inversely proportional to the baffle size. Unsupported larger baffles will resonate at lower frequencies, likely in the passband of the mid/woofer, while a narrow baffle's resonant frequency will likely be outside the mid/woofer's passband, and therefore will not be excited at all. Narrow front baffles usually mean deep enclosures, so proper side bracing will obviously be important, and perhaps the only bracing required.

C

[norcad]

A big enclosure needs to have big panels. Big panels resonate and make a sound you dont want to hear. It doesnt matter if this is front, back or sides. You have to use bracing on big panels anyway.

IMO both narrow and wide baffles has their benefits.
In my Seas 3way the biggest surprise was that they are much easier to place.
The wide baffle makes the first reflections down to 200hz, and therefor the distance to the backwall isnt that critical.

Thats why Troels Gravesen says that "they comes with a room"

On the other hand, the short depth in the enclosure is often a problem, because the reflection inside the box, from the backpanel to the drivercone. And there isnt much space for dampening either.

[looneybomber]

The wide baffle makes the first reflections down to 200hz, and therefor the distance to the backwall isnt that critical.

What is this "distance to the backwall" reflection thing you're talking about? And how does the baffle width play a part in it?

[norcad]

What is this "distance to the backwall" reflection thing you're talking about? And how does the baffle width play a part in it?

Its hard to explain this in english
When the wavelength is half of the baffle width, it starts radiate behind the speaker. It hits the backwall of your listening room, and will be reflected back to your ears with more or less time delay.
As your baffle gets wider, the frequency where this happens is lower. Because lower freq has longer wavelength.

If your bafflewidth is 50-60cm or more, this will happen mostly in the upper bass and lower, and not in the midband. Then the first reflection from the tweeter and mid will be from the enclosure itself, and not from the walls in your room. That makes the speakers sound more like the same, even if you put them in an another room.

Again; it comes with a room

Please understand that this is very simplyfied and theoretical, but it explains the IMO biggest benefits of my wide baffle projects.

Of course, no baffel at all, in a room without (bad) reflections would be mutch better!

EDIT: Just add a picture so you can see that I'm not only build wide baffles! And no, I dont say that wide baffle is "the best", or the only right way of building speakers!