small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

[charlielaub]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

Charlie, thanks for the time you took to do this.

Like others, I would encourage you to always measure the SPL. Not only so you can compare apples to apples but so we understand just how hard the driver is being pushed. Additionally, what would be nice (and I sometimes fail to do this myself) is to provide step measurements. IE; loud, louder. This gives is an idea of how the driver behaves at different levels to see if its nonlinear distortion has any linearity to it or if there's a "breaking point" where the driver loses its composure and maybe trace that back to a level and apply that to practical crossover locations. Because at the end if the day, HD is largely useful for determining crossover points. IMD data is more beneficial but isn't a must.

One person asked already so forgive me if I missed your answer: how ode was the mic placed for testing these? Which mic?

Thanks again for the efforts and resulting discussion.

- Erin

I positioned the mic at about 0.5m. That should be in the far field in this case, or close to it. The mic is an Earthworks M30BX.

I've always been hesitant to assign an SPL to my measurements. I take measurements through an outboard ADDA box, which has adjustable gain for both the outgoing and incoming signals, my amp has adjustable gain, and the software as well... too many places to screw up on the gain setting. I do have an SPL meter, but I don't really trust these things to give a very accurate SPL measurement since there are several factors that come into play. When I design speakers, I take all measurements of all drivers with the mic and amp at the settings that they will be used with, and then only the relative levels are important. As a result I've never really need to or bothered to pin down actual driver SPLs.

[ErinH]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

Charlie, thanks for the time you took to do this.

Like others, I would encourage you to always measure the SPL. Not only so you can compare apples to apples but so we understand just how hard the driver is being pushed. Additionally, what would be nice (and I sometimes fail to do this myself) is to provide step measurements. IE; loud, louder. This gives is an idea of how the driver behaves at different levels to see if its nonlinear distortion has any linearity to it or if there's a "breaking point" where the driver loses its composure and maybe trace that back to a level and apply that to practical crossover locations. Because at the end if the day, HD is largely useful for determining crossover points. IMD data is more beneficial but isn't a must.

One person asked already so forgive me if I missed your answer: how ode was the mic placed for testing these? Which mic?

Thanks again for the efforts and resulting discussion.

- Erin

[jonasz]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

The only thing unclear to me at this point is, if you put it in the right designers hands does it have the magic secret sauce? IE: does it just look good in the lab and on the test bench, or can it be magic to the ears too?

Well, in my NaO Notes they sure sound very very good! In this design they're used between 1000Hz and 6kHz. Can't hear any strain or misbehaving even at loud volumes, a very sweet yet detailed reproduction.

[Saturnus]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

I don't see any reason to retest it. It will not in any way change the outcome that the Scan Speak is quite obviously superior for this specific purpose. I hope it has come across at least that I consider the Scan Speak a superb mid range driver for a home audio application. Switch them around and make the Scan Speak fill the Faital's role and it would not come out as the best.

[r-carpenter]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

This is kind of a waste of wood in my opinion, but that is what will have to be done. The question is, should I bother in this case when the distortions are so different under very similar "nude" dipole operation?

There's of course no obligation. It is interesting and educational because the issues with testing will be dealt with by many more.
I would make 4'x4' panel with a provision for replaceable insert. Make an insert for each driver and re-test in near field and far field condition. The amount of diffraction you are getting from "naked" driver makes a direct comparison nearly impossible. In my limited "Faital" experience, the frequency response they are stating is very close to the actual response of the drivers (same with SS BTW). You could test in very "near field".
If you want to send me the drivers, I can test them on the 5'x10' panel and I probably have correct cutout size inserts already in my library.
On the other hand, if you are content with the results, then just call it a day. For your application, Scan is better.

[Face]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

So, the question is "what is a fair way to set the SPLs to be the same?". Above 2kHz the response of both drivers is trending differently, and the peaks and dips would made it difficult to establish a frequency at which to set the SPLs equal. Below 1.5kHz where the dipole 6dB/oct slope is observed the SPL is determined by the driver's physical geometry (given the SPLs are about the same). It seems to be that, in order to really make the SPLs equal in a fair way, both drivers have to be mounted on a large baffle or box. This will cause a relatively flat passband region to form in the frequency response, and within this region the SPLs can be compared on an equal basis. This is kind of a waste of wood in my opinion, but that is what will have to be done. The question is, should I bother in this case when the distortions are so different under very similar "nude" dipole operation?

It certainly would be an interesting comparison, but it's your call as it's your time and money.

[lorenmjones]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

Markk is not making irrelevant comparisons of a guitar speaker to a subwoofer. Both of these drivers could reasonably be considered for application as a midrange in a home hi-fi design. Could the Faital perform better in certain prosound applications...probably. Charlie was not saying the Faital is worthless, just that compared to a high quality "hi-fi" type midrange driver for this application it appears to have significantly higher distortion.

As to your contention that designing pro drivers is so much different that they are often purposely designed to produce higher levels of odd order harmonics, I must strongly disagree. Companies like Faital, 18Sound, B&C, Beyma, Eminence etc design drivers to have as low of distortion as possible just as do the engineers from ScanSpeak and Seas. The design goals are clearly different in terms of tradeoffs regarding TS parameters, power handling, efficiency and absolute output level but to say the designs intentionally introduce higher levels of distortion is just false. I am using all prosound drivers in my main speakers and they were all chosen because of documented LOW levels of distortion.

Sorry to only snip this bit and comment on but it illustrates the problem perfectly. It is utterly irrelevant what you, me or any other audiophile believes or think. The reality of professional audio design is completely different than home audio design. It's designed for different purposes and therefore require differently designed drivers that are good at different things. Why is this so hard to understand? You wouldn't compare a professional guitar driver to a home audio subwoofer and thereby conclude that the guitar driver is useless as a subwoofer. That is obvious. The subwoofer would be equally useless as guitar driver no matter how much electronic distortion you added. They are simply designed to fill different roles.

[charlielaub]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

I'd like to see the two compared again, but at the same SPL. Just to show that the FP doesn't excel at anything compared to the SS, no matter what label you want to put on them.

I've been thinking about this a little more, and it's a bit complicated in this case because I am testing both drivers "nude", e.g. without any baffle. This means that the frequency response of the driver is strongly influenced by the path length from the front (of the radiating surface, e.g. the cone) to the back. It's a dipole response. Now the drivers are very close in size, e.g. the frame OD is about the same (4FE23 = 100mm and 10F = 97.5mm) but you can see that the dipole peak of the 4FE32 is slightly higher in frequency (at about 2.3kHz) than the 10F (at about 2kHz) in the plot of their frequency responses, below. The manufacturer's spec sheets show SPL of about 90 dB/W each, and if you look at the MFG frequency response plots the 4FE23's on axis response is at about 92dB at 2kHz and the 10F is at around 92dB as well. This tells me that the driver SPL in the very nearfield will be the same but the front to back path for the 4FE32 must be slightly longer than that for the 10F, since everything else that would influence the dipole response is pretty much equal but the SPLs are offset below 2kHz. It's likely then that the different location of the windows in each frame are the source of the pathlength difference.



So, the question is "what is a fair way to set the SPLs to be the same?". Above 2kHz the response of both drivers is trending differently, and the peaks and dips would made it difficult to establish a frequency at which to set the SPLs equal. Below 1.5kHz where the dipole 6dB/oct slope is observed the SPL is determined by the driver's physical geometry (given the SPLs are about the same). It seems to be that, in order to really make the SPLs equal in a fair way, both drivers have to be mounted on a large baffle or box. This will cause a relatively flat passband region to form in the frequency response, and within this region the SPLs can be compared on an equal basis. This is kind of a waste of wood in my opinion, but that is what will have to be done. The question is, should I bother in this case when the distortions are so different under very similar "nude" dipole operation?

[Face]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

How can you say that? Apparently it excels at odd order hardmonic distortion! ;)

:D

In any case, where the Faitals are good are in small live music or instrument monitor where you want to accentuate the even overtones (odd harmonics) to make the music rich and engaging rather than pure and accurate. They will typically be used in a line of fours in such a speaker. Not to be confused with a line-array which is something completely different. The Faitals are specifically designed to have high 3rd and 5th harmonic distortion for this very reason as it would be very difficult to discern if the pitch of singing is correct or if a chord is strum correctly in a situation with a lot of background noise without such a harmonic distortion characteristics.

That's a pretty large shovel you're using...

[r-carpenter]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

well, testing is testing and really, the question is, what are we testing for? Pro drivers have been used in home audio applications by DIY for a long time, so there's nothing new about it. Is this particular model Faital mid workable for a good 3-way design?
Couple of other Faital drivers that I have tested, were very good to excellent in the non-linear performance, compare to other pro drivers or consumer audio counterparts (similar in intended size and passband) The TS were almost 100% on. Sort of not even fun to post the TS because .... it's the same as data provided by manufacturer.
But in order to define IF the mid in question is really better or can be used, proper testing needs to be done. So far we only ball-parked the performance.

[dlneubec]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

I'd like to see the two compared again, but at the same SPL. Just to show that the FP doesn't excel at anything compared to the SS, no matter what label you want to put on them.

How can you say that? Apparently it excels at odd order hardmonic distortion! ;)

[Face]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

I'd like to see the two compared again, but at the same SPL. Just to show that the FP doesn't excel at anything compared to the SS, no matter what label you want to put on them.

[lorenmjones]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

Well said on all counts sir. A very instructive and thorough post.

Well, it is true that certain studio monitor were, if not intentionally designed, at least they were marketed that way. Still, you won't find any sympathy on this board adhering to this philosophy. To crudely summarize, you are arguing that we should use the Faital in a DIY design and bring it to the next DIY as an example of a mediocre high distortion consumer design??? (Or, just bring an Auratone to the next big DIY!)



I'd have to disagree and I think I represent the majority of audiophiles. Even order is much more palatable, if you had to pick. Odd order provides a harshness, edge. It can certainly brighten up a dull piece, but this is just not the way to do it. I want my guitar to produce the harmonics, not my speaker. Is it necessarily a bad thing? Well, yes, yes it is. The reason is that, unlike an outboard effects processor, once I've designed the speaker, everything I play will have that same distortion. Maybe my metal sounds good with a little odd order distortion, but God forbid I listen to Allison Krauss. Worse yet, nonlinear distortion is level dependent in an unpredictable fashion (it is, after all, nonlinear...). Even if I find a little odd order helps "edge up" her voice, raising the volume now adds too much edge, and if I lower the volume, now there is not enough edge. No this is not the way to go.



It is true that, in general, pro drivers are designed to handle more power that drivers designed for home use. The problem is that lots of manufacturers know this and like to label drivers as "pro" implying higher power handling. It becomes a bit of a game. Having tested the 10F I can say it has quite a bit more power handling than the average driver. (The answer to this is a very long question. If I have time, I will explain it at some point.) However, since I've not tested the Faital, I can't comment on who would perform better.


A couple of general notes. Face posted a graph showing the 10F could take 250w with a filter. No way could the 10F dissipate that amount of power for any length of time. That graph, I believe is just max excursion at those frequencies. The driver will be thermally limited well before. Still, this it true for the Faital as well I'm sure. No need to test on that point. A transient at 250 w could be delivered by both. I guess it just depends on how you define transient...

Should the drivers have been tested at exactly the same spl? Yeah, probably, as nonlinear effects are, well, nonlinear. It is possible that those extra 2 dB are important from a distortion standpoint. But practically speaking we are correct in assuming regional linearity, so for small differences in drive levels Charlies assumption is probably correct.

[Saturnus]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

I'd have to disagree and I think I represent the majority of audiophiles.

Sorry to only snip this bit and comment on but it illustrates the problem perfectly. It is utterly irrelevant what you, me or any other audiophile believes or think. The reality of professional audio design is completely different than home audio design. It's designed for different purposes and therefore require differently designed drivers that are good at different things. Why is this so hard to understand? You wouldn't compare a professional guitar driver to a home audio subwoofer and thereby conclude that the guitar driver is useless as a subwoofer. That is obvious. The subwoofer would be equally useless as guitar driver no matter how much electronic distortion you added. They are simply designed to fill different roles.

[markk]

Re: small midrange shootout: Faital Pro 4FE32 and ScanSpeak 10F4424G00

You seem to be under the false impression that a studio monitor is supposed to always be speakers with very low distortion. That's one kind of studio monitor, other studio monitor are reference monitors used to represent the speakers that an ordinary consumer might be playing the music on, and then it's certainly not very low distortion that are the primary objective.

Well, it is true that certain studio monitor were, if not intentionally designed, at least they were marketed that way. Still, you won't find any sympathy on this board adhering to this philosophy. To crudely summarize, you are arguing that we should use the Faital in a DIY design and bring it to the next DIY as an example of a mediocre high distortion consumer design??? (Or, just bring an Auratone to the next big DIY!)

In any case, where the Faitals are good are in small live music or instrument monitor where you want to accentuate the even overtones (odd harmonics) to make the music rich and engaging rather than pure and accurate. They will typically be used in a line of fours in such a speaker. Not to be confused with a line-array which is something completely different. The Faitals are specifically designed to have high 3rd and 5th harmonic distortion for this very reason as it would be very difficult to discern if the pitch of singing is correct or if a chord is strum correctly in a situation with a lot of background noise without such a harmonic distortion characteristics.

I'd have to disagree and I think I represent the majority of audiophiles. Even order is much more palatable, if you had to pick. Odd order provides a harshness, edge. It can certainly brighten up a dull piece, but this is just not the way to do it. I want my guitar to produce the harmonics, not my speaker. Is it necessarily a bad thing? Well, yes, yes it is. The reason is that, unlike an outboard effects processor, once I've designed the speaker, everything I play will have that same distortion. Maybe my metal sounds good with a little odd order distortion, but God forbid I listen to Allison Krauss. Worse yet, nonlinear distortion is level dependent in an unpredictable fashion (it is, after all, nonlinear...). Even if I find a little odd order helps "edge up" her voice, raising the volume now adds too much edge, and if I lower the volume, now there is not enough edge. No this is not the way to go.

And on a final note about the Scan Speak having the same power handling as the Faital. I can assure you that the Scan Speaks are not designed to reproduce music at their maximum power handling for several hours like the Faitals, and you also forget that the difference in voltage sensitivity of 2dB dictates a power factor of 1.585, add in the required distortion you have to introduce electronically and it's around a power factor of 2 (and probably more), meaning the Scan Speak will have to take (at least) twice the power of the Faitals to reproduce the same sound at the same volume. Note that the 2 drivers are not measured with the same specification. The Faitals power handling is measured according to the AES standard, and the Scan Speaks are not measured for power handling at all. They are measured for maximum input voltage (IEC 60268-5 §17.3) which is then converted into a power handling. Typically this will result in a power handling that is around 50% higher than it would be under the AES standard.

It is true that, in general, pro drivers are designed to handle more power that drivers designed for home use. The problem is that lots of manufacturers know this and like to label drivers as "pro" implying higher power handling. It becomes a bit of a game. Having tested the 10F I can say it has quite a bit more power handling than the average driver. (The answer to this is a very long question. If I have time, I will explain it at some point.) However, since I've not tested the Faital, I can't comment on who would perform better.


A couple of general notes. Face posted a graph showing the 10F could take 250w with a filter. No way could the 10F dissipate that amount of power for any length of time. That graph, I believe is just max excursion at those frequencies. The driver will be thermally limited well before. Still, this it true for the Faital as well I'm sure. No need to test on that point. A transient at 250 w could be delivered by both. I guess it just depends on how you define transient...

Should the drivers have been tested at exactly the same spl? Yeah, probably, as nonlinear effects are, well, nonlinear. It is possible that those extra 2 dB are important from a distortion standpoint. But practically speaking we are correct in assuming regional linearity, so for small differences in drive levels Charlies assumption is probably correct.