Sorry for highjacking the topic a bit but this might be as good a moment as any.
I've been thinking at this old post ever since I first read it and have a couple of questions:
Thank you!
D7608, normalized
Audax TW025A28
By comparison, the Bliesma M74A looks much more well behaved.
That makes sense, but after having heard both the MDM55 and D7608 in various applications, the SS has the clear edge in terms of SQ and SPL capabilities. Don't get me wrong, the Morel is a very good mid, but it doesn't sound as detailed and open as the SS on vocals, strings or piano, despite being the smaller driver with a higher theoretical cutoff. The D7608 just sounds more agile and open overall. You can't cross the MDM55 as low either despite the claimed 2mm xmax. About 800 hz is the limit with a 12dB filter. Its better suited for small 3 ways with 8" LF drivers. They work very well in compact 3 way nearfield applications which need tight spacing to integrate drivers over a short distance. I still prefer to use the Morels in those situations.
A LP of 3.5k is optimal for mating the D7608 with smaller domes like the Seas 22TAF/G, SB 29RDNC and SB 26STCN. I've tried the good old Morel CAT378 with a 1st order filter along with the natural rolloff of the D7608 with good success, but the 10k peak of the SS needs to be notched out for that to work.
The most detailed metal dome which blends even better with the D7608 than the Seas is the pricier Audax TW025A28, but it only has a very slight edge. You wouldn't expect a metal dome to match the D7608 in harmonic character, but it does so more than most soft domes, likely due to the 2nd order HD profile around its upper rolloff.
The classic XT25 ring radiator also does well with the big SS dome, but it can't keep up in the dynamics department crossed with a 3.5k LR12 HP. You can hear the tweeter drawing attention to itself at higher SPLs and it begins to sound fuzzy.
The D7608 and Seas combo sounds effortless with acoustic, classical and string instrument music. Ribbons and planars work well too. The only major drawback is their narrow vertical dispersion can create an unnatural sounding off axis hole.
Theres a story behind my discovery of the D7608 which was purely by accident. The best combination I've tried came from a quick design using parts I already had thanks to a friend of a friend who didn't pay for parts he ordered and left me stuck with them. I came up with the idea to use a D7608 in a 3 way with the Audax TW025A28 and dual HM210C0 running in staggered LP (technically a 3.5 way). I used a bunch of left over crossover parts to throw a set of speakers together over a weekend, including the cabinets. I didn't even sim anything and just used a ton of clip leads, some white noise and my old Audio Control RTA to whip up a crossover in a few hours. I ended up living with these speakers for a few years in my main system.
I used to play Cello and believe to have a good ear for the instrument in a live setting - the above described 3 way played solo Cello so well, you could swear it sounded like the real instrument a couple of feet in front of you when closing your eyes. It was frightening how accurately proportioned the stereo image was.
Many years ago, a conductor friend of mine with a very picky set of ears auditioned the Audax/ScanSpeak 3 ways and gave me his critical feedback. He claimed it was the first time he could hear proper separation of string instruments in a full orchestra coming from non-elextrostat or planar speakers. He claimed the whole orchestra was correctly proportioned and balanced in the stereo image and couldn't believe the speakers were home made in a few days. This was coming from a man who owned the best Soundlab electrostats available at the time and was extremely critical when commenting on audio gear. He repeatedly said how clean and natural the mids sounded, even compared to his setup and he was even more surprised when I told him the type and cost of the drivers. I removed one of the D7608s and showed him what it looked like inside. His jaw dropped and he commented "how is that flimsy plastic frame driver making all that beautiful music?"...
The D7608 is that good compared to many other mids costing more, regardless of design. There's obviously nothing really special about it other than being a dome mid - no shorting rings or a fancy motors either.
If you insist on the Morel MDM55, the best tweeter match is the CAT378 for several reasons. The primary one is the perfect acoustic center alignment, but also the softer smoother character and the ability of the CAT378 to use 1st order filters without much drama.
For most people, the SPL capability of a single D7608 is more than enough for domestic use and even in a monitor application. I'm a bit more demanding when it comes to dynamic range, so running two SS mid domes with staggered filters is my preference as long as the CTC spacing isn't conflicting with the chosen crossover.
There are of course alot more choices in mid domes. The more expensive and exotic Bliesma mids are very nice, but also very expensive. A pair of silk or Alu M74s is over $800 here in the US. I can buy a pair of D7608s for just over $200, which are capable of 90 percent of the performance at that price. The lack of a rear chamber can be an advantage to the more experienced designer, so there are some benefits.
The people who claim a cone mid can do just as well or better likely haven't heard how good a well implemented large dome mid can sound. You'll be spending at least $200 on a cone mid that can come close to the harmonic capability of the D7608. No, its not perfect and it definitely is a difficult driver to work with, but its capable of reproducing very complex details an average cone mid usually can't.
Even the smaller MDM55 is a step above most cone mids. The only thing which can do better is a very well designed horn system, but those are even more expensive and tricky to work with. The SS dome has the right balance of detail and smoothness. When correctly implemented, it just disappears into the room. The FR curve looks ugly on paper, but it performs alot better in the flesh.
I've been thinking at this old post ever since I first read it and have a couple of questions:
- Would you still recommend the Audax as the best match for D7608?
- I think you mentioned in other posts you like crossing this midrange to 3-3.5Khz, is it the case for the Audax too?
- Looking at the directivity of the D7608 and the Audax at 3.5Khz, it seems there's a severe mismatch at about any angle off-axis. This might be addressed to some degree by careful baffle design but it still looks worrisome?
Thank you!
D7608, normalized
Audax TW025A28
By comparison, the Bliesma M74A looks much more well behaved.
@SunRa If you can afford the M74A then I wouldn't hesitate to recommend it. Its definitely in another performance league than the D7608.
Looking at your sim, there are more factors to consider ie. filter slopes, baffle geometry/size/layout to name the main ones. What data are you using for the sims? The manufacturers info on the D7608 and TW025A28 should be taken with a grain of salt.
I'd still recommend the TW025A28 over the other metal domes. Its an incredibly transparent tweeter with excellent dynamic capabilities. The T25B is of course a bit better, but not by much. A WG fitted to the TW025A28 would be a smart idea, but isn't really needed. I'd still suggest the 3k xover on the Audax. It depends on the baffle step and compensation.
Looking at your sim, there are more factors to consider ie. filter slopes, baffle geometry/size/layout to name the main ones. What data are you using for the sims? The manufacturers info on the D7608 and TW025A28 should be taken with a grain of salt.
I'd still recommend the TW025A28 over the other metal domes. Its an incredibly transparent tweeter with excellent dynamic capabilities. The T25B is of course a bit better, but not by much. A WG fitted to the TW025A28 would be a smart idea, but isn't really needed. I'd still suggest the 3k xover on the Audax. It depends on the baffle step and compensation.
Thank you for the reply!
The plots are not sims, but measurements performed by hificompass. Sorry, I should have mentioned, links are here: D7608, M74A.
The Audax plot is from the Madisound datasheet.
I couldn't find other measurements of the Audax on a flat baffle. However it seems it behaves really well with an waveguide.
Although, if I were to go the waveguide route, I would probably use a known combination, such is the D2604 with a WG300 clone, than to hack at a relatively expensive tweeter like the Audax.
The plots are not sims, but measurements performed by hificompass. Sorry, I should have mentioned, links are here: D7608, M74A.
The Audax plot is from the Madisound datasheet.
I couldn't find other measurements of the Audax on a flat baffle. However it seems it behaves really well with an waveguide.
Although, if I were to go the waveguide route, I would probably use a known combination, such is the D2604 with a WG300 clone, than to hack at a relatively expensive tweeter like the Audax.
I’m targeting a budget build. M74A is probably easier to work with allows a lower crossover and it appears it would blend well with a non-waveguide tweeter.If you can afford the M74A then I wouldn't hesitate to recommend it. Its definitely in another performance league than the D7608.
But I’ll try first to figure out if a budget build is possible around D7608.
@SunRa No hacking needed to WG the Audax. The front plate is flat and the dome opening is almost square profile. The WG148 fits right on it. Only need to drill holes for the different hole locations.
Are you referring to the R2604 and WG300 combo mentioned on Heismann's Acoustics site? The D2604 is a regular dome and R2604 is a ring radiating dome. Also take note the Peerless XT25 isn't as good as the SS R2604 despite appearing the same.
Are you referring to the R2604 and WG300 combo mentioned on Heismann's Acoustics site? The D2604 is a regular dome and R2604 is a ring radiating dome. Also take note the Peerless XT25 isn't as good as the SS R2604 despite appearing the same.
Last edited:
I was thinking at the dome version, specifically at D2604 / 833000, the one with higher SPL. Measurements in a WG300 are here: https://heissmann-acoustics.de/en/test-scan-speak-discovery-d2604-833000-p-audio-pct-300/.
Good to know it's easy to fit an off-the-shelf waveguide on the Audax, I still kind of favor it because of your strong recommendation. And because I really want a retro looking gold dome in my build 🙂)
At any rate, a tweeter waveguide appears to be mandatory with the D7608 if one aims for a smooth directivity transition between the two. As you said, I'll know more when I do some measurements and baffle simulations, I guess.
Good to know it's easy to fit an off-the-shelf waveguide on the Audax, I still kind of favor it because of your strong recommendation. And because I really want a retro looking gold dome in my build 🙂)
At any rate, a tweeter waveguide appears to be mandatory with the D7608 if one aims for a smooth directivity transition between the two. As you said, I'll know more when I do some measurements and baffle simulations, I guess.
A pair of very old (SN 29 & 94 -Early 70's?) and beaten up ATC midrange domes for sale on Ebay: https://www.ebay.com/itm/296043432484
Not recommended, but interesting from an historical point of view. Labelled as 'Soft Dome 2' each with a different power rating label, one from a PA-75 bass driver and the other from a PA-100, I assume that means there was a Soft Dome 1?
Cheers, Ralph
Not recommended, but interesting from an historical point of view. Labelled as 'Soft Dome 2' each with a different power rating label, one from a PA-75 bass driver and the other from a PA-100, I assume that means there was a Soft Dome 1?
Cheers, Ralph
Attachments
The following statement may ruffle some feathers, but it has to be mentioned. One significant thing to consider when crossing a mid to a tweeter is choosing a turnover frequency that suits the overall acoustic character of the mid driver. It basically means paying attention to the overall strengths of the mid driver. Getting right down to it, the most important driver in the entire speaker system is the midrange. Yes, I said it. Its the most critical thing to focus on when designing a speaker, especially on a 3 way system. That's really why you would even consider building a 3 way speaker to begin with.
The magic frequency range being discussed here is focused around 500 hz - 5 k. If a speaker is to sound accurate, engaging and simply down to earth correct, it needs to excell in this range - without exception. It's the bandwidth the ear will immediately notice at first listen and continue to pick apart. It contains the essential "meat" of the entire audible bandwidth. To put this into clear perspective, you could lop off the bands above and below 500 - 5k and still comprehend the jist of what's being representing by itself. I'm not saying bass and treble aren't important, but if you need to compromise, it should never ever be in the critical midrange. That's why you'd always select the mid driver first and the woofer / tweeter based on the mid, never ever the other way around.
Getting back to the point I'm trying to make here, you have to first and foremost do what the midrange driver wants and needs when considering a crossover implementation. However, you shouldn't ask the mid to do something it can't do, depending on the tradeoff of SPL vs HP frequency you're willing to make. The second critical decision is the LP frequency, more specifically choosing a point where the cutoff will blend coherently with the tweeter in a way both mid and tweeter will sound like a single driver when both are playing together. Sometimes this means you have to select a turnover point that will sound "natural" when listening off axis. This will also vary depending on playback SPL (I'll explain later). The notion you'll have perfect power response and natural sounding, coherent summing on axis is almost always idyllic. The exception to my previously stated rule is don't ask the tweeter to do something that it doesn't want to or can't do. By that, I mean forcing the tweeter to play too loud in relation to the turnover point and slope chosen. This also means staying out of the roughly 1500 to 2500 hz band with the crossover when designing/building a speaker intended for the purpose of music consumption and enjoyment.
Why, you should ask? Its because the human ear is VERY sensitive to abrupt timing, distortion and transients in this narrow range. It has to do with the primary resonance onset peak of a person's pina (ear lobe shape and geometry). Oversimplified, it means a natural exaggeration of distortion, phase and a sharp amplitude shift perceived by the human ear in this narrow range, specifically when listening at high SPLs. This Isn’t so much an issue for most people when listening at lower levels, the threshold level being roughly 95 - 100 dB A weighted, depending on how sensitive you are. You wouldn't consider this so much if the speaker was primarily intended for critical monitoring purposes, as ideally, the engineer would compensate for this during mixing / mastering, pulling some EQ from the 1500 - 2500 hz band, so the music sounds better at higher SPLs on speakers that aren't compensated for this small but important hearing anomaly. Sadly, most of the time, with most rock / pop material, this isn't considered (especially on most 80s - late 90s music). This is a separate issue from the fletcher munson loudness curve, which concentrates on reducing the 2.5 to 5 k range and inversely pushing the low end.
Why all the concern with this specific midrange to tweeter turnover? Considering the 1500 - 2500 hz hearing anomaly when designing a 3 way speaker system, it makes for a more enjoyable speaker when listening louder with the 99% of music released in the end of the last millennium. Chosing a crossover point higher than 2500 hz avoids the issue by slightly reducing the off axis energy around the 1.5 - 2.5 k range exagerated by a smaller diaphragm tweeter's wide directivity above this point. This is even the case with flat EQ. A WG on the tweeter can help mitigate some of this issue, but avoiding the lower end rise in the tweeter's odd order HD at higher SPLs (as minor and benign as it may appear and be considered by number) will greatly reduce listener fatigue - especially the case crossing a smaller dome above 2.5k. It almost always sounds better just doing this, even with mids that tend to beam a bit higher up.
Most typical average sized diaphragm mids will have much lower distortion below 2.5k and reasonable off axis performance compared to a typical 1" HF dome. The tradeoff between midrange power response and lack of audible fatigue in the same range is worth it. Such is the case when crossing the D7608 to most 1" domes. I know there are many people who will point out the lower end HD rise in a tweeter would be insignificant, but you'd be surprised if you heard the difference for yourself between a very low HD mid doing the sub 2.5k vs a small dome at higher SPLs. If you don't believe me, play a pure low HD sine sweep through the 1.5 to 2.5 k range on your average speaker and pay close attention to the sudden level and intensity jump in this area. I actually tested this on a few other people, some experienced musicians, audio techs and everyday musically educated people. Most of them noted a jump in this range along with more aggressive, fatiguing character in sound. It was pleasantly analytical at lower SPLs, but shifted to a very aggressive character at louder volume.
The magic frequency range being discussed here is focused around 500 hz - 5 k. If a speaker is to sound accurate, engaging and simply down to earth correct, it needs to excell in this range - without exception. It's the bandwidth the ear will immediately notice at first listen and continue to pick apart. It contains the essential "meat" of the entire audible bandwidth. To put this into clear perspective, you could lop off the bands above and below 500 - 5k and still comprehend the jist of what's being representing by itself. I'm not saying bass and treble aren't important, but if you need to compromise, it should never ever be in the critical midrange. That's why you'd always select the mid driver first and the woofer / tweeter based on the mid, never ever the other way around.
Getting back to the point I'm trying to make here, you have to first and foremost do what the midrange driver wants and needs when considering a crossover implementation. However, you shouldn't ask the mid to do something it can't do, depending on the tradeoff of SPL vs HP frequency you're willing to make. The second critical decision is the LP frequency, more specifically choosing a point where the cutoff will blend coherently with the tweeter in a way both mid and tweeter will sound like a single driver when both are playing together. Sometimes this means you have to select a turnover point that will sound "natural" when listening off axis. This will also vary depending on playback SPL (I'll explain later). The notion you'll have perfect power response and natural sounding, coherent summing on axis is almost always idyllic. The exception to my previously stated rule is don't ask the tweeter to do something that it doesn't want to or can't do. By that, I mean forcing the tweeter to play too loud in relation to the turnover point and slope chosen. This also means staying out of the roughly 1500 to 2500 hz band with the crossover when designing/building a speaker intended for the purpose of music consumption and enjoyment.
Why, you should ask? Its because the human ear is VERY sensitive to abrupt timing, distortion and transients in this narrow range. It has to do with the primary resonance onset peak of a person's pina (ear lobe shape and geometry). Oversimplified, it means a natural exaggeration of distortion, phase and a sharp amplitude shift perceived by the human ear in this narrow range, specifically when listening at high SPLs. This Isn’t so much an issue for most people when listening at lower levels, the threshold level being roughly 95 - 100 dB A weighted, depending on how sensitive you are. You wouldn't consider this so much if the speaker was primarily intended for critical monitoring purposes, as ideally, the engineer would compensate for this during mixing / mastering, pulling some EQ from the 1500 - 2500 hz band, so the music sounds better at higher SPLs on speakers that aren't compensated for this small but important hearing anomaly. Sadly, most of the time, with most rock / pop material, this isn't considered (especially on most 80s - late 90s music). This is a separate issue from the fletcher munson loudness curve, which concentrates on reducing the 2.5 to 5 k range and inversely pushing the low end.
Why all the concern with this specific midrange to tweeter turnover? Considering the 1500 - 2500 hz hearing anomaly when designing a 3 way speaker system, it makes for a more enjoyable speaker when listening louder with the 99% of music released in the end of the last millennium. Chosing a crossover point higher than 2500 hz avoids the issue by slightly reducing the off axis energy around the 1.5 - 2.5 k range exagerated by a smaller diaphragm tweeter's wide directivity above this point. This is even the case with flat EQ. A WG on the tweeter can help mitigate some of this issue, but avoiding the lower end rise in the tweeter's odd order HD at higher SPLs (as minor and benign as it may appear and be considered by number) will greatly reduce listener fatigue - especially the case crossing a smaller dome above 2.5k. It almost always sounds better just doing this, even with mids that tend to beam a bit higher up.
Most typical average sized diaphragm mids will have much lower distortion below 2.5k and reasonable off axis performance compared to a typical 1" HF dome. The tradeoff between midrange power response and lack of audible fatigue in the same range is worth it. Such is the case when crossing the D7608 to most 1" domes. I know there are many people who will point out the lower end HD rise in a tweeter would be insignificant, but you'd be surprised if you heard the difference for yourself between a very low HD mid doing the sub 2.5k vs a small dome at higher SPLs. If you don't believe me, play a pure low HD sine sweep through the 1.5 to 2.5 k range on your average speaker and pay close attention to the sudden level and intensity jump in this area. I actually tested this on a few other people, some experienced musicians, audio techs and everyday musically educated people. Most of them noted a jump in this range along with more aggressive, fatiguing character in sound. It was pleasantly analytical at lower SPLs, but shifted to a very aggressive character at louder volume.
There was also a list of dome speakers. I can't find the post.
Can it be added again including the ones mentioned after the list was created?
Thank you!
Can it be added again including the ones mentioned after the list was created?
Thank you!
I forgot to mention something important in my last mile long post. Again, this may not sit well with the design philosophy which some of the more.experienced. speaker designers cater to. It is however important if you want to end up with speakers that overall sound good playing real world music in typical listening settings and scenarios.
Sometimes, if you have to make a compromise decision regarding the mid - tweeter turnover point, it will be better to design the crossover to provide a slight off axis dip around 3 - 4 k. This actually can work greatly in your favor when dealing with a highly reflective room and wide radiating baffle design, especially if you prefer wide dispersion in difficult acoustical settings. This strategic dip I'm advocating can permit tuning the speaker placement to ease the aggressive mids that make your ears bleed with louder, poorly mixed music, especially older 80s albums with hot EQ in the upper mids. The off axis dip designed into the speaker can be adjusted by horizontallly toeing them in or out in regards to the listening location. It sounds simple, but it actually isn't, especially if you want accurate representation of the soundstage, acoustical setting and imaging of individual instruments conveyed in the recording itself.
Please keep in mind I'm not advocating building monitor type speakers here this way. In that type of application you ideally need absolute flat response across a wide listening window without any fluff or embellishment. You need to get a proper idea what's going on in a studio setting mixing and mastering music properly. This isn't what I'm advocating or after here.
The speakers I'm mainly referring to here, to design and build are mainly for the purpose of higher end music enjoyment and consumption, enabling you to enjoy a wider range of material, some of which may have been recorded or mixed poorly, but still have great artistic and musical value to enjoy.
A good balance between all out accuracy, detail and smooth FR are important qualities a well designed set of speakers need to possess. Most speakers which are designed for professional monitoring applications actually don't sound that good even with decent recordings and almost certainly will sound worse with bad ones. They are just meant to prvide the engineer with a cold hearted perspective of what's going on in a recording without prejudice. They're not made to embellish anything, despite the often very high price tag. No one would ever buy them for the purpose of just listening to music for enjoyment. Sure, you could do a decent mix on hifi speakers, but you would need to be very familiar with their strengths and weaknesses to do so. On the other hand, monitor speakers ie. the ubiquitous Yamaha NS10Ms, don't really sound good at all with any kind of music. They're just a familiar reference for engineers who work in other studios. That is their strength and purpose as a tool. Sure, there are some monitors which actually sound enjoyable when playing decent recordings, but usually will be too revealing and in your face for most people playing typical modern production type music. A decent speaker for music enjoyment is one that (in theory) can play everything well, not just specific material. The in room FR of that type of speaker won't be flat, manly because flat simply doesn't sound that good in the real world when listening to music for the purpose of enjoyment. We don't hear music ruler flat in reference to the actual live event, especially considering most recordings are done in a studio setting with added (often multiple) artifical spacial cues and lots or signal processing. Its the difference between taking a picture with a camera and panting it on a canvas with a brush. Even live recordings aren't done in an accurate reference to the real world live performance. That simply doesn't sound that good to most people, even highly experienced engineers.
The above is the main reason why the Harman curve was developed and suggested as a guideline for the average consumer listening to typical music in a typical environment. For me personally, its not a good enough SQ goal for my listening tastes and experienced ears. I prefer my own specific in room response characteristics.
The in room FR balance will highly depend on how loud you play your music. This balance actually changes with playback level in reference to the actual SPL the mastered recording was mixed and mastered at. Your final EQ setting will therfore change based on the SPL the recording was being monitored at when the final mix and mastering was done. Thats a factor which most people don't pay attention to and it makes a huge difference in how the recording is perceived. A significant shift in playback volume will completely change the way your music sounds in your environment. This is another reason why absolute flat FR doesn't sound good or even accurate, based on the playback level alone.
When I design a speaker that would suit my needs, I'd logically start with very transparent, highly resolving drivers but combined in a way which still allows me to enjoy the bulk of music that currently exists despite its shortcomings when it was recorded ie. some of the older 60s, 70s and 80s highly compressed, over EQed and over-colored analog recordings. Those need to be enjoyable to listen to but not irritate the ears. There's a delicate balance of resolution vs smoothness which is actually hard to combine.
In short, I like to listen to most styles of music in a real world setting, sometimes at higher volume levels. A wide dynamic range, accurate sound stage and very low distortion with excellent midrange transparency are very important to me. The midrange needs to be accurate (especially vocals and piano) but somewhat forgiving to keep listening fatigue low. The importance of a super wide listening window is secondary to me, as I don't entertain people in my home. I want smooth power response but not at the price of a tweeter crossed too low screaming at me. I don't listen to test tones and pink noise for enjoyment, especially on flat sounding, unforgiving speakers, so I have to design and build my own to get the happy medium I'm searching for.
Thats why I advocate the compromises pointed out here, so that sufficient balance exists.between high resolution and smooth, non fatiguing sound.
Sometimes, if you have to make a compromise decision regarding the mid - tweeter turnover point, it will be better to design the crossover to provide a slight off axis dip around 3 - 4 k. This actually can work greatly in your favor when dealing with a highly reflective room and wide radiating baffle design, especially if you prefer wide dispersion in difficult acoustical settings. This strategic dip I'm advocating can permit tuning the speaker placement to ease the aggressive mids that make your ears bleed with louder, poorly mixed music, especially older 80s albums with hot EQ in the upper mids. The off axis dip designed into the speaker can be adjusted by horizontallly toeing them in or out in regards to the listening location. It sounds simple, but it actually isn't, especially if you want accurate representation of the soundstage, acoustical setting and imaging of individual instruments conveyed in the recording itself.
Please keep in mind I'm not advocating building monitor type speakers here this way. In that type of application you ideally need absolute flat response across a wide listening window without any fluff or embellishment. You need to get a proper idea what's going on in a studio setting mixing and mastering music properly. This isn't what I'm advocating or after here.
The speakers I'm mainly referring to here, to design and build are mainly for the purpose of higher end music enjoyment and consumption, enabling you to enjoy a wider range of material, some of which may have been recorded or mixed poorly, but still have great artistic and musical value to enjoy.
A good balance between all out accuracy, detail and smooth FR are important qualities a well designed set of speakers need to possess. Most speakers which are designed for professional monitoring applications actually don't sound that good even with decent recordings and almost certainly will sound worse with bad ones. They are just meant to prvide the engineer with a cold hearted perspective of what's going on in a recording without prejudice. They're not made to embellish anything, despite the often very high price tag. No one would ever buy them for the purpose of just listening to music for enjoyment. Sure, you could do a decent mix on hifi speakers, but you would need to be very familiar with their strengths and weaknesses to do so. On the other hand, monitor speakers ie. the ubiquitous Yamaha NS10Ms, don't really sound good at all with any kind of music. They're just a familiar reference for engineers who work in other studios. That is their strength and purpose as a tool. Sure, there are some monitors which actually sound enjoyable when playing decent recordings, but usually will be too revealing and in your face for most people playing typical modern production type music. A decent speaker for music enjoyment is one that (in theory) can play everything well, not just specific material. The in room FR of that type of speaker won't be flat, manly because flat simply doesn't sound that good in the real world when listening to music for the purpose of enjoyment. We don't hear music ruler flat in reference to the actual live event, especially considering most recordings are done in a studio setting with added (often multiple) artifical spacial cues and lots or signal processing. Its the difference between taking a picture with a camera and panting it on a canvas with a brush. Even live recordings aren't done in an accurate reference to the real world live performance. That simply doesn't sound that good to most people, even highly experienced engineers.
The above is the main reason why the Harman curve was developed and suggested as a guideline for the average consumer listening to typical music in a typical environment. For me personally, its not a good enough SQ goal for my listening tastes and experienced ears. I prefer my own specific in room response characteristics.
The in room FR balance will highly depend on how loud you play your music. This balance actually changes with playback level in reference to the actual SPL the mastered recording was mixed and mastered at. Your final EQ setting will therfore change based on the SPL the recording was being monitored at when the final mix and mastering was done. Thats a factor which most people don't pay attention to and it makes a huge difference in how the recording is perceived. A significant shift in playback volume will completely change the way your music sounds in your environment. This is another reason why absolute flat FR doesn't sound good or even accurate, based on the playback level alone.
When I design a speaker that would suit my needs, I'd logically start with very transparent, highly resolving drivers but combined in a way which still allows me to enjoy the bulk of music that currently exists despite its shortcomings when it was recorded ie. some of the older 60s, 70s and 80s highly compressed, over EQed and over-colored analog recordings. Those need to be enjoyable to listen to but not irritate the ears. There's a delicate balance of resolution vs smoothness which is actually hard to combine.
In short, I like to listen to most styles of music in a real world setting, sometimes at higher volume levels. A wide dynamic range, accurate sound stage and very low distortion with excellent midrange transparency are very important to me. The midrange needs to be accurate (especially vocals and piano) but somewhat forgiving to keep listening fatigue low. The importance of a super wide listening window is secondary to me, as I don't entertain people in my home. I want smooth power response but not at the price of a tweeter crossed too low screaming at me. I don't listen to test tones and pink noise for enjoyment, especially on flat sounding, unforgiving speakers, so I have to design and build my own to get the happy medium I'm searching for.
Thats why I advocate the compromises pointed out here, so that sufficient balance exists.between high resolution and smooth, non fatiguing sound.
Some great reading! @profiguy, do you feel there are equivalent decision points regarding bass-dome integration (point, slopes)? Curious also, if you don't jive with the Harman curve, do you have one of your own or illustration of the off-axis contours you described, that you think are good?
@motokok Yes, I have a specific curve I attempt to build into my designs which factor in all the above mentioned criteria. Obviously keep in mind my hearing is stll pretty good for being in my mid 50s, but I don't claim to hear past 15k.
If I were to describe the target curve I prefer, just take the Harman curve and dip the 2.5 - 5 k roughly 1.5 - 2 dB centered at 3.2 - 3.5 k. Also keep the top end flat right past this dip. This is at an average 85 dB reference listening level. Going below this add more 2.5 - 5 k until the dip is removed at about 65 dB. Going louder than 85 dB, gradually pull more mid dip to end up at about 3.5 - 4 dB when you reach 105 dB.
Generally speaking, with most 3 ways I build, the LF - mid turnover is preferred to be no higher than 300 hz 2nd order on your average typical 10" woofer. That puts it at roughly 200 - 250 hz 2nd order for 12" and around 150 - 200 2nd order for 15" LF drivers, but those will almost always need a separate midbass to run a mid dome (4 way).
Essentially 300 hz is the point where the omni LF radiation turns more directive. The larger LF drivers 12"+ usually require an extra midbass to sound best at higher SPLs unless you've got a very good woofer with excellent motor and cone that goes cleanly into the mids ie. SB34NRXL - not very common and getting into SS Revalator territory. You can get away with a larger dedicated cone mid for a 3 way if you're using something very high end, such as least a 6" mid with decent linearity ie. Purify paper cone models, ScanSpeak slit cone 18M8631T or 18M4631T. Those are excellent and can go past 2.5k easily to pair with a decent HF dome without a mid dome.
If you want to run a mid dome, you'll need a very good 10" LF or an even better (rare) 12". That will mean using a 3" mid dome, crossing in the sub 600 hz area.
In the case of an 8" LF or dedicated midbass (4 way), I'd cross a 3" mid dome around 600 hz. Some few 8" cones will do 800 - 900 hz cleanly (watch the 1st mode surround reflection FR kink with most 8"s around the 800 - 1000 hz area. It will show in the impedance curve and is sometimes rather audible ie. lower end ScanSpeak discovery stuff, SB Acoustics, etc. That will be noticeable at higher volume levels, otherwise stick with a 6" - 7" and mate with your favorite 2" - 3" mid dome. There are always a few unicorn drivers out there which can bend the rules a tiny bit, but again, very rarely do so.
I don't advocate steeper than 2nd order filters below 1k. Those are audible in transient phase shift playing back piano, percussion and other rhythm instruments. 3rd order is ok between 3k + turnover. Once you go above 5k, you're not going to notice the upper harmonics of those frequencies as a complex connected harmonic structure. Anything higher order HD won't be noticeable anymore. Only the fundamental notes will be noticeable, mainly because the harmonics land above 10k, which depending on youe upper hearing cutoff won't sound like much more than a sine wave to the ear. If you want to hear what this sounds like, try playing a 5k+ sine wave on your speakers and switch back and forth between a square wave. You won't be able to discern the difference between sine and square wave past 5k or so. Thats why tweeter HD isn't that critical past 8k.
The nitpick range of most typical 1" domes will be the 1.5 - 3.5k range, where your ears are VERY sensitive. This is the main area where you should pay very close attention to THD and IMD, as it defines the overall acoustic signature of a dome tweeter. Always watch for a sharp rise in lower order HD when shopping for tweeters. You'd preferably want to cross it at a frequency above the point where THD drops below the 1% point at 90 dB/1M levels. Tweeters like the SS R2604 and Tymphany XT25 will have a sharp rise in HD below 3k. Yes, its predominantly even order HD, but still very noticeable crossing from a good dome mid, not so much from an average 4 - 5" cone mid.
Hopefully that gets the point across.
If I were to describe the target curve I prefer, just take the Harman curve and dip the 2.5 - 5 k roughly 1.5 - 2 dB centered at 3.2 - 3.5 k. Also keep the top end flat right past this dip. This is at an average 85 dB reference listening level. Going below this add more 2.5 - 5 k until the dip is removed at about 65 dB. Going louder than 85 dB, gradually pull more mid dip to end up at about 3.5 - 4 dB when you reach 105 dB.
Generally speaking, with most 3 ways I build, the LF - mid turnover is preferred to be no higher than 300 hz 2nd order on your average typical 10" woofer. That puts it at roughly 200 - 250 hz 2nd order for 12" and around 150 - 200 2nd order for 15" LF drivers, but those will almost always need a separate midbass to run a mid dome (4 way).
Essentially 300 hz is the point where the omni LF radiation turns more directive. The larger LF drivers 12"+ usually require an extra midbass to sound best at higher SPLs unless you've got a very good woofer with excellent motor and cone that goes cleanly into the mids ie. SB34NRXL - not very common and getting into SS Revalator territory. You can get away with a larger dedicated cone mid for a 3 way if you're using something very high end, such as least a 6" mid with decent linearity ie. Purify paper cone models, ScanSpeak slit cone 18M8631T or 18M4631T. Those are excellent and can go past 2.5k easily to pair with a decent HF dome without a mid dome.
If you want to run a mid dome, you'll need a very good 10" LF or an even better (rare) 12". That will mean using a 3" mid dome, crossing in the sub 600 hz area.
In the case of an 8" LF or dedicated midbass (4 way), I'd cross a 3" mid dome around 600 hz. Some few 8" cones will do 800 - 900 hz cleanly (watch the 1st mode surround reflection FR kink with most 8"s around the 800 - 1000 hz area. It will show in the impedance curve and is sometimes rather audible ie. lower end ScanSpeak discovery stuff, SB Acoustics, etc. That will be noticeable at higher volume levels, otherwise stick with a 6" - 7" and mate with your favorite 2" - 3" mid dome. There are always a few unicorn drivers out there which can bend the rules a tiny bit, but again, very rarely do so.
I don't advocate steeper than 2nd order filters below 1k. Those are audible in transient phase shift playing back piano, percussion and other rhythm instruments. 3rd order is ok between 3k + turnover. Once you go above 5k, you're not going to notice the upper harmonics of those frequencies as a complex connected harmonic structure. Anything higher order HD won't be noticeable anymore. Only the fundamental notes will be noticeable, mainly because the harmonics land above 10k, which depending on youe upper hearing cutoff won't sound like much more than a sine wave to the ear. If you want to hear what this sounds like, try playing a 5k+ sine wave on your speakers and switch back and forth between a square wave. You won't be able to discern the difference between sine and square wave past 5k or so. Thats why tweeter HD isn't that critical past 8k.
The nitpick range of most typical 1" domes will be the 1.5 - 3.5k range, where your ears are VERY sensitive. This is the main area where you should pay very close attention to THD and IMD, as it defines the overall acoustic signature of a dome tweeter. Always watch for a sharp rise in lower order HD when shopping for tweeters. You'd preferably want to cross it at a frequency above the point where THD drops below the 1% point at 90 dB/1M levels. Tweeters like the SS R2604 and Tymphany XT25 will have a sharp rise in HD below 3k. Yes, its predominantly even order HD, but still very noticeable crossing from a good dome mid, not so much from an average 4 - 5" cone mid.
Hopefully that gets the point across.
I've posted previously on the issue of what I call "tweeter scream". Its the horrible sound dome tweeters make when they're crossed too low and play too loud for that crossover point. The fact is most people try to push a tweeter way too low, mainly because they want to maintain directivity from the woofer or mid. They will sight the fact HD is low "enough" to be tolerable for the chosen turnover point. Fact is it can never be low enough, as its much more audible than assumed.
If you have the type of hearing I do, you'll hear how rough an average sized dome tweeter can sound when its forced to play too loud below 2.5 k. It just sounds awful and offensive. That's also due to that weird jump in hearing we have between 1.5 - 2.5 k, which further magnifies the audibility. If it doesn't bother you, than your lucky, but often times the people who don't notice will judge those who do and pass judgment. When people say something is "good enough", that means its not as good as it should or could be. For me, I'd rather compromise in the directivity and at least have cleaner, less fatiguing sound.
Lower odd order HD is always more important than wider dispersion and directivity. Fact is, you're not paying enough attention to the quality of your woofer or midrange if you have to force the tweeter to cover for it. Its ok if the tweeter doesn't have to play too loud, but most of the time you end up with a big deficit in dynamic range in the lower mids.
The beauty of a larger dome mid is the fact it can usually cross higher up to avoid all these above mentioned issues. Most 2" domes can go up to 5 - 6 k and 3".domes can manage 3 - 4k. That's well out of the fussy range where most HF domes loose their composure.
If you have the type of hearing I do, you'll hear how rough an average sized dome tweeter can sound when its forced to play too loud below 2.5 k. It just sounds awful and offensive. That's also due to that weird jump in hearing we have between 1.5 - 2.5 k, which further magnifies the audibility. If it doesn't bother you, than your lucky, but often times the people who don't notice will judge those who do and pass judgment. When people say something is "good enough", that means its not as good as it should or could be. For me, I'd rather compromise in the directivity and at least have cleaner, less fatiguing sound.
Lower odd order HD is always more important than wider dispersion and directivity. Fact is, you're not paying enough attention to the quality of your woofer or midrange if you have to force the tweeter to cover for it. Its ok if the tweeter doesn't have to play too loud, but most of the time you end up with a big deficit in dynamic range in the lower mids.
The beauty of a larger dome mid is the fact it can usually cross higher up to avoid all these above mentioned issues. Most 2" domes can go up to 5 - 6 k and 3".domes can manage 3 - 4k. That's well out of the fussy range where most HF domes loose their composure.
I would like to shine some light on this old classic Seas DOME 87 H Alnico which is not really a dome mid range or a dome tweeter but a little of both. Many years ago i had some great results with this dome speaker in a two way with a 6 inch bass driver. The mid band was simply great and i believe, a clever way to go for a two way which combines a seamless mid band at the cost of some very high frequencies which i did not really miss because the over all sound was that good!
Attachments
what size is that one ? that's an interesting idea ,even if you had to put a 0.75inch tweeter above that wasn't doing much it still sounds like it could work
@Henk Haring That Seas mid-tweet is a lost gem. Its funny how they did things more right back then for the sake of midrange purity and clarity instead of inaccurate, boomy bass and sizzling treble. Later on, the people selling us audio equipment were trying make their products desireable by focusing on specs and the bragging rights they garnered.
Fact is, consumers used to be more vocal of what they wanted from a product rather than manufacturers defining the customer and the taste they should have. Its basically being told what to like and buy. That's really what's essentially wrong today with advertising and marketing.
Those older speakers did a much better job at conveying the music at the time than the junk they sold us later on based on advertised hype. American speaker manufacturers marketed specs more than relying on the impression the customer got from the actual performance of the products.
Science is an important foundation to define performance and push the boundaries to improve products. It isn't however a sole valid substitute of process which should define how the actual quality and performance are perceived. That's like someone trying to tell you what kind of food you should like to eat based on a third party scientific chemical analysis of its taste.
Fact is, consumers used to be more vocal of what they wanted from a product rather than manufacturers defining the customer and the taste they should have. Its basically being told what to like and buy. That's really what's essentially wrong today with advertising and marketing.
Those older speakers did a much better job at conveying the music at the time than the junk they sold us later on based on advertised hype. American speaker manufacturers marketed specs more than relying on the impression the customer got from the actual performance of the products.
Science is an important foundation to define performance and push the boundaries to improve products. It isn't however a sole valid substitute of process which should define how the actual quality and performance are perceived. That's like someone trying to tell you what kind of food you should like to eat based on a third party scientific chemical analysis of its taste.
This is a 1.5 inch dome. https://www.diyaudio.com/community/threads/seas-87h-1968-test.417092/what size is that one ? that's an interesting idea ,even if you had to put a 0.75inch tweeter above that wasn't doing much it still sounds like it could work
@profiguy I agree that the race and marketing for extended highs and pumping low's took over, similar to over saturated high contrast pictures with a overdose sharpening. Seas made some larger (expensive) 1.5 inch domes but it never took of in the main stream DIY field considering that most people have a two way speaker in their living room and could benefit from this concept.
@Henk Haring The modern equivalents to that big Seas are these -
T35 X3-06 (alnico version)
T35C002 (neo version)
The later one is my favorite large HF dome. It does pretty much everything you ask it above 2k 2nd order. I prefer this one with the SS D7608 and Volt VM753. You can use both tweeters with a 1st order network ie. Dynaco A25 clones
T35 X3-06 (alnico version)
T35C002 (neo version)
The later one is my favorite large HF dome. It does pretty much everything you ask it above 2k 2nd order. I prefer this one with the SS D7608 and Volt VM753. You can use both tweeters with a 1st order network ie. Dynaco A25 clones
- Home
- Loudspeakers
- Multi-Way
- The dome midrange thread