Carmody S2000 DIY Speaker Kit Review

[bigLP]

While we are on space heater talk:

A gas(natural or propane) fired wall heater, available at any home store, would do a good job of heating your garage. Better than any electric I know of.

 

[Rick Sykora]

If anyone is interested in the pair I built, dm me.

Once I get the test one back from Amir, I plan to sell them.

 

[MZKM]

Something is seriously wrong with the scoring system or my measurements of F328Be.

To get a general feel, here are all the curves isolated (with aspect ratio following the 2034 standard) and restricted to the frequencies the formula cares about for each:

As stated, the steep slope of the S2000 boosts the SM_PIR score, if we set it to say 0.8, the score goes from 4.9 to 4.7.

 

[Mattcc22]

I have a pair of S2000's in my living room setup with an Ultimax 10" ported sub. They are very good sounding speakers for the money and especially for their size. I went with these because they just barely fit on top of an entertainment center to the left and right of the TV.

I measured mine the other day and was concerned about the drop off in the higher frequencies. I actually took them apart to check the crossover and made sure I didn't miss something and couldn't find anything wrong. Glad to see Amir's measurement matches up to mine!

 

[Angsty]

Looks like it would be a fun kit to build. I have an NOS pair of Cambridge Soundworks Model Seventeen's I had planned to use, but these are likely better performers.

 

[Newman]

The estimated in-room response is good and ultimately that's what matters more than on-axis anechoic measurements, unless of course you live in an anechoic chamber.

Incorrecto

 

[JohnBooty]

Have we forgotten our fundamental audio science lessons, and that most of what reaches our ears is reflected sound?

(Unless we are doing extreme nearfield listening, listening to speakers with exceptionally narrow beamwidth, are listening outdoors, or do in fact live in an anechoic chamber?)

 

[Newman]

Fundamentals breed fundamentalists, if ‘we’ don’t expand on them with further audio science lessons. Try it. ‘We’ might discover along with Toole that the direct arrival sound is distinguishable from the summed-over-time sound, and needs a flat, extended and smooth frequency response, if the speaker is to be highly rated in listening tests.

And, as an aside, you might try not to talk down to people who might actually turn out to be your teacher.

 

[amirm]

As stated, the steep slope of the S2000 boosts the SM_PIR score, if we set it to say 0.8, the score goes from 4.9 to 4.7.

The fact that the linear regression worked so well on bookshelves means then that listeners like much less treble than we think they do. That would explain my preference for this speaker then. I have often commented on speakers being bright and people point to flat on-axis response and wonder how that could be. Maybe we need higher slopes especially since my room is fairly reflective.

Just thinking out loud!

 

[amirm]

Fundamentals breed fundamentalists, if ‘we’ don’t expand on them with further audio science lessons. Try it. ‘We’ might discover along with Toole that the direct arrival sound is distinguishable from the summed-over-time sound, and needs a flat, extended and smooth frequency response, if the speaker is to be highly rated in listening tests.

I don't know that such research extends to 20 kHz. Most of the listeners used can't even hear that high. Here is Sean Olive's study with respect to different equalization:

As you see, the winning system had a measured response (red) that slopes 10 dB down from low to high. This needs to happen through combination of direct and indirect sounds.

Here is how the listeners rated the different setups:

As you see the highest range is only 12 kHz. And the band below is 4.5 khz.

BTW, please moderate your tone. It doesn't entice me to want to engage you. This is not the complaint department. We are here to figure things out.

 

[beagleman]

The Olive score turned out higher than I thought it would, which is the opposite of what happened with the F328Be.

Actually this tiny speaker seems to measure having better bass than the F328e, at least on "paper" a lower roll off??

 

[Newman]

@amirm My ‘tone’ was directed at someone who had just condescendingly written, “Have we forgotten our fundamental audio science lessons, and that most of what reaches our ears is reflected sound?” Please keep that in mind, and hence the possibility that you are blowing the whistle at the second man in.

 

[MZKM]

Actually this tiny speaker seems to measure having better bass than the F328e, at least on "paper" a lower roll off??

As Amir stated in the main post, his bass measurements are slightly off for the F328Be due to the complexity of the design (3 woofers with 2 ports), he is currently working on some optimizations.

 

[Newman]

I don't know that such research extends to 20 kHz. Most of the listeners used can't even hear that high. Here is Sean Olive's study with respect to different equalization:

As you see, the winning system had a measured response (red) that slopes 10 dB down from low to high. This needs to happen through combination of direct and indirect sounds.

Here is how the listeners rated the different setups:
View attachment 93453

As you see the highest range is only 12 kHz. And the band below is 4.5 khz.

This is a test only of summed-over-time response vs listener preference. It doesn’t say anything about direct sound, or its correlation to speaker preference. In fact, IIRC, the same speaker was used each time, so this test is actually quite silent on the issue of preference vs direct sound FR.

Furthermore, Toole has evolved his position since the date of those experiments, so IIRC he is now saying that, if the direct sound FR is smooth, flat and extended, and the dispersion characteristics are as per the preferred dispersion characteristics, then generic target-curve type EQ is best avoided above the transition frequency. Reason being that it will tend to change the direct sound FR away from its native smooth, flat and extended character, and to human perception this modification to the direct sound FR is distinguishable and less preferred.

I’m not sure why you raised the question of 20 kHz in response to my post, which made no claims in relation to that point.

cheers

 

[TimVG]

This is a test only of summed-over-time response vs listener preference. It doesn’t say anything about direct sound, or its correlation to speaker preference. In fact, IIRC, the same speaker was used each time, so this test is actually quite silent on the issue of preference vs direct sound FR.

Furthermore, Toole has evolved his position since the date of those experiments, so IIRC he is now saying that, if the direct sound FR is smooth, flat and extended, and the dispersion characteristics are as per the preferred dispersion characteristics, then generic target-curve type EQ is best avoided above the transition frequency. Reason being that it will tend to change the direct sound FR away from its native smooth, flat and extended character, and to human perception this modification to the direct sound FR is distinguishable and less preferred.

I’m not sure why you raised the question of 20 kHz in response to my post, which made no claims in relation to that point.

cheers

The speaker tested was the 802N

Loading its anechoic response in REW and adjusting it to the target curve, it turns out that the applied EQ actually made the direct sound flatter (with this particular speaker). That fact + it having the most extended bass after EQ, and no wonder it was the preferred curve. Doesn't mean it applies to every loudspeaker.

 

[Helicopter]

A gas(natural or propane) fired wall heater, available at any home store, would do a good job of heating your garage. Better than any electric I know of.
View attachment 93345 View attachment 93346

These are great. I had one that did not require electricity once, which was a bonus. Dirty generator power and modern computer control can be a bad combo.

 

[thewas]

Furthermore, Toole has evolved his position since the date of those experiments, so IIRC he is now saying that, if the direct sound FR is smooth, flat and extended, and the dispersion characteristics are as per the preferred dispersion characteristics, then generic target-curve type EQ is best avoided above the transition frequency. Reason being that it will tend to change the direct sound FR away from its native smooth, flat and extended character, and to human perception this modification to the direct sound FR is distinguishable and less preferred.

Toole didn't have to evolve his position, as those experiments were done by Olive and not Toole.

The speaker tested was the 802N
Loading its anechoic response in REW and adjusting it to the target curve, it turns out that the applied EQ actually made the direct sound flatter (with this particular speaker). That fact + it having the most extended bass after EQ, and no wonder it was the preferred curve. Doesn't mean it applies to every loudspeaker.

Not fully sure there as the EQ actually made a direct sounds peak at the presence, as it is correcting the steady state and so also the sound power dip due to the too large mid driver. Which would kind of show that at typical hifi listening distances which are usually in the "far field" maybe not only concentrating above the transition frequency on direct sound but also on the sound power can giver a higher preference which is also observed at that Harman loudspeaker score.

 

[BenB]

It's odd to me that the output level drops about 3 dB between 200 Hz and 60 Hz, even though the impedance curve doesn't show the asymmetric response at the woofer's resonance that is typical from an extended-bass-shelf alignment. Assuming that drop is accurate, then the diminished output below 200 Hz, and the lack of output below 60 Hz may partly explain the preference for the diminished high frequency output. Bass has a way of taking the edge off from the higher frequencies. With diminished bass output, a flat speaker can sound too bright, and grating. Reducing the treble output for such a speaker can improve our perception. If this speaker were supplemented with a subwoofer, then it might be able to support a neutral balance, but played stand-alone, it might be better with this tilted response.
Additionally, the drop in level of the tweeter appears to compensate for the directivity error (which manifests itself mostly in the vertical plots). The tweeter is putting the right amount of energy into the room to continue the trend established by the rest of the speaker, but it's doing it by sending less energy over a wider space. I suspect that people have a fairly high tolerance for the slope of the room power curve. What we can't tolerate is peaks in the curve that make some frequencies stand out, and mask the frequencies around them. I know personally I've messed around with EQ, making different slopes. I've sometimes left the EQ active, forgotten about it, sat down to listen to my speakers, and been none-the-wiser; happily enjoying a modified balance without even realizing it. Eventually a certain source material would clue me in; a dark source material mated to an exaggerated tilt, for example.

 

[andreasmaaan]

It's odd to me that the output level drops about 3 dB between 200 Hz and 60 Hz, even though the impedance curve doesn't show the asymmetric response at the woofer's resonance that is typical from an extended-bass-shelf alignment.

Looks asymmetrical enough to me:

 

[Newman]

Toole didn't have to evolve his position, as those experiments were done by Olive and not Toole.

I was talking about Toole's position, not Olive's. Amir raised Olive in response to my post about Toole, not the other way around.