Research Paper Help Please! 2-way vs. 3-way

[Drewdwn93]

This questionnaire is for a school research paper, and I would appreciate as many responses as I can get. I understand that building loudspeakers involves a lot to take into consideration when designing one. But, for the sake of this paper please answer the following questions Yes or No, knowing that I understand there is no one size fits all approach. Designing, building, or buying our own loudspeakers adds a certain element to what we perceive or hear in a loudspeaker, making the audiophile topic highly subjective. I would like to get an idea of the averages concerning the questions involved.
Questionnaire:
Considering a room size of ~400sq. ft., with typical obstructions. A subwoofer is used, and crossed at 80Hz. The build is for front LR mains and consists of drivers no larger than 6.5 in. Cost and build complexity is not considered. And, multiple listening positions is the goal.
1. Is the phase audible at frequencies other than 180 degrees out of phase?

YES, or NO?

2. Considering phase shift is non-linear throughout the room at different frequencies, does it matter if it’s introduced by the speaker, if it can’t be controlled anyway?

YES, or NO?

3. Is the time alignment problem that can be introduced when using a 3-way 4th order crossover, important in considering whether or not to use a single driver or simpler (less components) two-way design?

YES, or NO?

4. Is a time aligned crossover important considering that the other components in a stereo introduce their own misalignment, i.e. Microphones/distance in recording, filters and enhancements added to the recording before release, amplifier used and any equalizers, DAC and ADC, and speaker voice coils not aligned?

YES, or NO?

5. Does the number of components and complexity of a 3-way 4th order crossover add more distortion and phase/time shift than any intermodulation introduced by using only two drivers covering the frequency range of 80Hz-20KHz? Room size considered.

YES, or NO?

6. Do you believe intermodulation distortion is audible in a two-way speaker given the parameters listed above?
YES, or NO?

7. Considering the cost, complexity, and time involved, would you build a two-way or three-way loudspeaker for the conditions stated above?

TWO-WAY, or THREE-WAY?

[BobEllis]

Is this to be a poll of designer's preferences or are you looking for more answers to your questions to form your own prefences? If the latter, I suggest reading the numerous design threads of the more popular projects here to get a feel for what our resident experts feel is important.

1. Absolute phase or phase discontinuities? In my experience, discontinuities are audible down to a fairly low number. Absolute phase, not so much.
2. Discontinuities in phase are going to be noticeable.
3. Time alignment is an issue no matter what design, it may be more difficult to get right over multiple crossovers, but system design targets are a bigger driver.
4. As above. It's discontinuities that draw attention to themselves. Remember that we are trying to reproduce what the artist, engineer and producer want us to hear. The phase and timing changes in the production process are a different issue.
5. Doubtful. IM is less of a concern for me than pushing a driver to its upper an lower limits.
6. Depends on the particular drivers and crossover points.
7. Again, it depends on the performance target. How loud, what type of music, acceptable distortion, etc. My preference in that room would still be a three way, given the budget. Crossover complexity isn't an issue unless design/build time is. Although you propose to cross to a sub at 80 Hz, I think the main issue us power response rather than IM distortion. In a 2 way in that room, I'd probably go with a pair of 6.5" midwoofers to give you some headroom on the bottom end SPL. You'll tend to push the crossover lower to avoid beaming, making the tweeter work harder. There are a few tweeters that can handle low crossovers, but you still end up limited at high SPLs. A smaller dedicated midrange allows you to carry the tweeter crossover higher and maintain smooth power response. In leaner times, the

Good luck with your project.

[Evil Twin]

Many of your questions are highly flawed as they start off with incorrect assertions regarding the basis for the question.... In many cases outrightly incorrect, such as asserting the differences between arrival times for different full range microphones (which are roughly equivalent to changing the seating distance of the performer from the minimum phase equivalent for the recording to differences in a loudspeaker driver time alignment, which is necessary (whether achieved physically or by modification of the phase response in the crossover) merely to achieve a smooth measured on axis response (ignoring for the moment the impact on power response for non-co-incident with different phase slopes in the crossover region has...)

You are drawing large generalities about characteristics (such as two way versus three way) without any fundemantal discussion or underlying understanding of the specific driver or transducer characteristics...

With regards to several of your poorly put questions, I would ask you, do you know what the Duelund transfer function equation is for a three way crossover and how it relates to the total system transfer function, including phase?

If not, and considering how you have proposed these other questions in a manner most naive, my suggestion is that you are not ready to do a research paper, and you need to do a reset and do your homework on these topics in a more thorough manner.

[neuro]

1. Yes.
2. Yes. 10dB frequency response peaks may be introduced by room modes, so does frequency response of the speaker matter, if it can't be controlled anyway?
3. No. Assuming you value time alignment (which would be wise); a poorly designed two-way could fail time alignment just as could a three way.
4. Yes.
5. No. Without further information it can not be said that more crossover components means more distortion and/or phase shift.
6. No. The parameters above are not sufficient to determine that IMD would be audible. In the absence of data I'll answer no.
7. Three-way

[Drewdwn93]

Thanks for your reply. This isn't for designing a system. It's simply to see what the average DIYer does with their own system, so yes a pole. I currently have MTMs for my mains and center, MT for rears and 2 tapped horn subs. Speaker building is highly subjective and regardless of how many books we all read, or how many measurements we take, it still comes down to personal preference. I have two sets of speakers that I've tested numerous times and they are almost identical in frequency response. Both are nearly flat but one has lower extension, but for some odd reason they sound completely different. I prefer the one with LESS extension. The vocals sound so much better. The only reason I can figure this, is because I bought the one I like better and the other woofer came from a center channel that had blown tweeters.

[cjd]

Speaker building is highly subjective and regardless of how many books we all read, or how many measurements we take, it still comes down to personal preference.

This is false. Almost all of speaker building is highly objective, if you understand the research. I'd wager that, if you get data on an individual's hearing you could then determine what they'd want in a system. Given the wide range of recording quality and source material fidelity, that may also influence what will sound good, but the data is there - it's not subjective.

For your questions, I'll wade through the odd assumptions and risk some answers.

1. Yes - assuming you mean the relative phase of components in the system.

2. Yes - assuming you mean relative phase of the components...

3. What problem? Too many assumptions in the question, but there is no time alignment problem - time alignment has to be managed regardless.

4. Time alignment is critical in the crossover if it's not dealt with in the physical domain (sloped baffle, etc.)

5. Crossover component contribution to distortion is insignificant relative to the power of the dar... sorry, the distortion introduced by the mechanics of a driver doing its thing, which ultimately drives the primary distortion signature of a system. Higher order crossovers certainly change absolute phase differently than lower-order crossovers, but unless you're only looking at electric order, the acoustic order has nothing to do with parts count; It is possible to have a 4th order acoustic slope with fewer components than a 2nd order acoustic slope.

6. If there is IM distortion in a system, it may be audible, or not... but none of your presumptive parameters provide any basis for assuming anything at all.

7. I'll always choose a 3-way over a 2-way when it is reasonable to do so. I'll even go 4-way if I'm looking for proper full range response (16Hz-20kHz)

[agrippa]

This is false. Almost all of speaker building is highly objective, if you understand the research. I'd wager that, if you get data on an individual's hearing you could then determine what they'd want in a system. Given the wide range of recording quality and source material fidelity, that may also influence what will sound good, but the data is there - it's not subjective.

I would also agree with that. Most peoples ears are not "educated" enough to understand what is the "best" sound. I go with the measurements/data over the golden ears.

As far as 2 sets of speakers having similar frequency response but sounding different you still have to consider a number of other factors. Harmonic distortion being one.

Here is someone who says it better than I, on subjectivity vs objectivity, and is not afraid to voice his opinion http://zaphaudio.com/evaluation.html Check out his article on non-linear distortion it might help.

If I were you I would comb the internet for much of what you seek.

[Alaric]

This is false. Almost all of speaker building is highly objective, if you understand the research

This was my very first thought, and I've never even built, let alone designed, a speaker. Just reading some of the design and build threads of Jon Marsh, cjd, ET, Jim Holtz, and others shows very specific goals leading to driver/crossover/cabinet design and selection within well defined parameters. Evil Twin's current Kurosawa thread is a great example of the design process. The initial poll makes it seem like a crap shoot with a box of random parts and how to best work around the terrible mismatch of components. Without a LOT more research and understanding of the design and implementation process I don't expect much of a grade on the OP's paper. Just my $.02.