I am beginning a project to make a good sounding speaker system, I am looking for products that hit way above their weight and while I cannot audition every tweeter and mid on the planet I can look at the measurements that some amazing people put online for us to benefit from! I have started a spreadsheet with what I am considering some good metrics for a potential mid/tweeter. Spreadsheet
The project is meant to be full range, or darn close to full range so there will be a sub woofer, it is currently chosen as the Morel MW 1058. I bought a couple of these on sale for a great price so I'm using these. If there are better subs I am open to options but I like them so, so far so good!
In terms of Mids and Tweeters I have listed some of the best measuring speakers and catalogued them, I will be adding to this list but its a start for me to keep track of different available speakers that have better measurements.
Obvisouly measurements are not everything, but I would assume that of all the drivers in all the world the best ones also measure good. So I am eliminating drivers that measure poorly off axis, have bad step response and have higher than normal distortion measurements.
At this time for mids I would like to spend less than $200, maybe $220 or so. For tweeter less than $200, ideally less than $150. Many of my list are much higher than this but I just went by measurements and figured the price out later to as to not bias my decision. For instance the Morel CAT tweeter looks bad on measurements but its on the list, I was biased!
So if you have any driver suggestions I will add them to the list! I went to hificompass.com for measurements and will be seeking other resources, do you know of any resources that have comprehensive measurements? I will look there too!
This will be an active system using the Dayton 100x4 w/DSP amp. At least for now until I find out it sucks, if it does. Crossovers would likely be in the 80-120hz range and 1800hz range. At least that seems to be a fairly safe range on most speakers graphs. Each speaker will get 100 watts and the sub 200 watts. 100x4 amp for each speaker. I know it likely won't push 100 watts and thats fine. I have the 36v 10a PS for the amp as well.
Notes for the spreadsheet:
Mids: I am looking for fairly flat speakers, they don't have to be super flat but flat enough, I looked at the 80hz to 2000 hz for the mids and measured peaks and troughs.
Step response to find drivers that recover quickly, likely not the perfect measurement but its something.
Harmonic distortion measurements, I looked in the 80-2000hz range and guaged it.
Tweeters are similar, but the range changes. From 2k - 10k is usually fine, but from 10k-20k things get weird most of the time. I would assume the better tweeters do a better job and these first data points are all off axis dips in response.
Step response for the same reason and also Harmonic Distortion.
I also took liberty to rank tweeters based on 3db down like we do with subs. Just on the high end! So I listed the frequency that they go down 3db. At 0 and at 45 degrees.
Any suggestions would be appreciated, I know speakers don't sound like their graphs, its just the starting point, I will buy several drivers to test them out and see if they have a good sound after this selection process.
The project is meant to be full range, or darn close to full range so there will be a sub woofer, it is currently chosen as the Morel MW 1058. I bought a couple of these on sale for a great price so I'm using these. If there are better subs I am open to options but I like them so, so far so good!
In terms of Mids and Tweeters I have listed some of the best measuring speakers and catalogued them, I will be adding to this list but its a start for me to keep track of different available speakers that have better measurements.
Obvisouly measurements are not everything, but I would assume that of all the drivers in all the world the best ones also measure good. So I am eliminating drivers that measure poorly off axis, have bad step response and have higher than normal distortion measurements.
At this time for mids I would like to spend less than $200, maybe $220 or so. For tweeter less than $200, ideally less than $150. Many of my list are much higher than this but I just went by measurements and figured the price out later to as to not bias my decision. For instance the Morel CAT tweeter looks bad on measurements but its on the list, I was biased!
So if you have any driver suggestions I will add them to the list! I went to hificompass.com for measurements and will be seeking other resources, do you know of any resources that have comprehensive measurements? I will look there too!
This will be an active system using the Dayton 100x4 w/DSP amp. At least for now until I find out it sucks, if it does. Crossovers would likely be in the 80-120hz range and 1800hz range. At least that seems to be a fairly safe range on most speakers graphs. Each speaker will get 100 watts and the sub 200 watts. 100x4 amp for each speaker. I know it likely won't push 100 watts and thats fine. I have the 36v 10a PS for the amp as well.
Notes for the spreadsheet:
Mids: I am looking for fairly flat speakers, they don't have to be super flat but flat enough, I looked at the 80hz to 2000 hz for the mids and measured peaks and troughs.
Step response to find drivers that recover quickly, likely not the perfect measurement but its something.
Harmonic distortion measurements, I looked in the 80-2000hz range and guaged it.
Tweeters are similar, but the range changes. From 2k - 10k is usually fine, but from 10k-20k things get weird most of the time. I would assume the better tweeters do a better job and these first data points are all off axis dips in response.
Step response for the same reason and also Harmonic Distortion.
I also took liberty to rank tweeters based on 3db down like we do with subs. Just on the high end! So I listed the frequency that they go down 3db. At 0 and at 45 degrees.
Any suggestions would be appreciated, I know speakers don't sound like their graphs, its just the starting point, I will buy several drivers to test them out and see if they have a good sound after this selection process.
By fullrange people at this forum mean something else than you are planning. Thats ok, you should post in multiway.
Good luck.
Good luck.
Your way of selecting and qualifying good drivers seem completely unusable for your goal, I for one cannot find out if a driver is good based on the given parameters and ideals for your application.
For example to quantify step response like that is essentially completely useless. We are talking about a band limited transducer without filtering. It is way better too look at waterfall plots and frequency response to look for resonances etc. For example one tweeter gets the label ragged when that is just because the tweeter has a breakup at 25khz which shows up as a ragged step response, looks bad but in practice not a problem as it is not or almost not audible. The nuance of why something leads to a bad parameter/score/measurement is more important in choosing than one number, it is often, in this case too, even impossible to do it like that.
Same goes for frequency response divinations, what is the point of presenting it like this when you have a dsp to equalize it flat for almost no added cost. Even than how do you know a driver is really worse than another if you don’t know why it diviates and if it is even possible or not to fix it by eq/filtering. When you take that into account you cannot score drivers anymore like this as the deviation number on itself says nothing anymore. Same goes for tweeters why would a tweeter that extend more be better? Often one cannot even hear up that high and even than is it even desirable to have good off axis extension? (too much highs can lead to siblant and unpleasant sound).
Same story for HD distortion let’s say a driver has a distortion where the origin of the distortion has a much more audible effect but a lower overall score on one driver than the other. Should one than choose the lower driver or the one with least audible impact? Also how can you take the audibility of distortion in to account when you only have a range. When a driver has 6db lower mid distortion than another it may sound way better despite having similar ranged end points or average distortion number. You have to look at the origin of the distortion in relation to the frequency it occurs and than judge if is an (audible) problem. Sometimes a distortion has a so small impact on a narrow frequency band that it may not be even audible it thus could be still a very good driver despite having a ‘bad’ score.
Regarding comparing tweeter frequency response I personally would worry more about the lowend performance. Crossing a tweeter to a 6 inch is not a trivial task. It is way more important if that can be done properly without too much distortion (way more critical at 2khz!!). So look at a decently low fs and simulate max spl so it can hit reasonable numbers >100db at that frequency and distortion doesn’t rise excessively high. Just because it can make sound at 2khz doesn’t make it a good idea to cross it at that crossover frequency. In most cases recommendations can be found on this and other fora which would help easing that search for rough guidelines if the tweeter could be used.
I really recommend redefining a new set of ideals and goals and doing that without trying to narrow it to one number or parameter. It also would be very beneficial too look at the Thiele smalle parameters to see if the driver can hit the required max volume, can it fit a reasonable box etc. Also one should take in mind that matching the drivers (radiation patterns for example) with each is very important for the end result. Maybe 2 drivers work good together but not at 1.8khz. Just randomly picking a frequency point limits driver options and might be less than ideal.
For example to quantify step response like that is essentially completely useless. We are talking about a band limited transducer without filtering. It is way better too look at waterfall plots and frequency response to look for resonances etc. For example one tweeter gets the label ragged when that is just because the tweeter has a breakup at 25khz which shows up as a ragged step response, looks bad but in practice not a problem as it is not or almost not audible. The nuance of why something leads to a bad parameter/score/measurement is more important in choosing than one number, it is often, in this case too, even impossible to do it like that.
Same goes for frequency response divinations, what is the point of presenting it like this when you have a dsp to equalize it flat for almost no added cost. Even than how do you know a driver is really worse than another if you don’t know why it diviates and if it is even possible or not to fix it by eq/filtering. When you take that into account you cannot score drivers anymore like this as the deviation number on itself says nothing anymore. Same goes for tweeters why would a tweeter that extend more be better? Often one cannot even hear up that high and even than is it even desirable to have good off axis extension? (too much highs can lead to siblant and unpleasant sound).
Same story for HD distortion let’s say a driver has a distortion where the origin of the distortion has a much more audible effect but a lower overall score on one driver than the other. Should one than choose the lower driver or the one with least audible impact? Also how can you take the audibility of distortion in to account when you only have a range. When a driver has 6db lower mid distortion than another it may sound way better despite having similar ranged end points or average distortion number. You have to look at the origin of the distortion in relation to the frequency it occurs and than judge if is an (audible) problem. Sometimes a distortion has a so small impact on a narrow frequency band that it may not be even audible it thus could be still a very good driver despite having a ‘bad’ score.
Regarding comparing tweeter frequency response I personally would worry more about the lowend performance. Crossing a tweeter to a 6 inch is not a trivial task. It is way more important if that can be done properly without too much distortion (way more critical at 2khz!!). So look at a decently low fs and simulate max spl so it can hit reasonable numbers >100db at that frequency and distortion doesn’t rise excessively high. Just because it can make sound at 2khz doesn’t make it a good idea to cross it at that crossover frequency. In most cases recommendations can be found on this and other fora which would help easing that search for rough guidelines if the tweeter could be used.
I really recommend redefining a new set of ideals and goals and doing that without trying to narrow it to one number or parameter. It also would be very beneficial too look at the Thiele smalle parameters to see if the driver can hit the required max volume, can it fit a reasonable box etc. Also one should take in mind that matching the drivers (radiation patterns for example) with each is very important for the end result. Maybe 2 drivers work good together but not at 1.8khz. Just randomly picking a frequency point limits driver options and might be less than ideal.
Last edited:
@admin
The DSP does give me options but I would like to start with good drivers and rely less on DSP, but it does give plenty of options when building and crossing over. A properly built driver would sound better than a DSP'd driver, at least I would think. Not to mention there are only so many DSP settings you can put on a driver, software limited, so it would be better to save those for other things rather than just bad driver design.
I will do more work on the low part of the tweeter, you are definitly correct about harshness in tweeters, though I would likely attribute that to dome material rather than frequency response. Starting out with a really good tweeter can't be a bad thing!
In regards to the waterfall vs step response thing, I hear arguments all the time about waterfalls can be manipulated to look like whatever they want by raising the floor of it. So in reality there is no perfect measurement, the step response was there on hificompass and I was looking for drivers that were able to be controlled more easily so this seemed to fit the bill since you apparently can't use T/S parameters for this.
Funny thing about the audio world there is always someone saying "you can't do it this way" but never tells you how to get it done! I would like to find a good framework for finding good drivers so I am going down this path because there is not a better one. Distortion could add good things to sound, I agree but without testing every driver you have no way of knowing what is good/bad.
Over time I have heard several things though, well performing drivers sound better, low distortion helps a driver sound more neutral and crossover frequencies are super important. So I am looking for these things, you are welcome to add to the list! I just want some scientific way to go through all the hundreds of choices!
The other thing, I went into this blind, just looking at drivers sizes and I looked at every driver on HIFICOMPASS, the ones that are "known" to be good drivers are on this list. Sure some that are known to be good drivers didn't make it. But some of the really good ones are. Purifi is on here, Morel is on here, Relevator is on here. This not on purpose, They all have better off axis, distortion responses, not all of them do good in the step but it was interesting that the good ones made this list, so it can't be BAD to have these properties!
I'll go more into depth on the tweeter low end, thats a great point.
The DSP does give me options but I would like to start with good drivers and rely less on DSP, but it does give plenty of options when building and crossing over. A properly built driver would sound better than a DSP'd driver, at least I would think. Not to mention there are only so many DSP settings you can put on a driver, software limited, so it would be better to save those for other things rather than just bad driver design.
I will do more work on the low part of the tweeter, you are definitly correct about harshness in tweeters, though I would likely attribute that to dome material rather than frequency response. Starting out with a really good tweeter can't be a bad thing!
In regards to the waterfall vs step response thing, I hear arguments all the time about waterfalls can be manipulated to look like whatever they want by raising the floor of it. So in reality there is no perfect measurement, the step response was there on hificompass and I was looking for drivers that were able to be controlled more easily so this seemed to fit the bill since you apparently can't use T/S parameters for this.
Funny thing about the audio world there is always someone saying "you can't do it this way" but never tells you how to get it done! I would like to find a good framework for finding good drivers so I am going down this path because there is not a better one. Distortion could add good things to sound, I agree but without testing every driver you have no way of knowing what is good/bad.
Over time I have heard several things though, well performing drivers sound better, low distortion helps a driver sound more neutral and crossover frequencies are super important. So I am looking for these things, you are welcome to add to the list! I just want some scientific way to go through all the hundreds of choices!
The other thing, I went into this blind, just looking at drivers sizes and I looked at every driver on HIFICOMPASS, the ones that are "known" to be good drivers are on this list. Sure some that are known to be good drivers didn't make it. But some of the really good ones are. Purifi is on here, Morel is on here, Relevator is on here. This not on purpose, They all have better off axis, distortion responses, not all of them do good in the step but it was interesting that the good ones made this list, so it can't be BAD to have these properties!
I'll go more into depth on the tweeter low end, thats a great point.
Typically, SB Acoustics products are considered to hit way above their weight. And you there are plenty of different drivers to choose from. For your purposes the Satori series would probably fit best.
For midranges, i would select ones with least amount of breakups. Since legendary Vifa is no longer available, new ones are pretty close, peerles has really nice close replacement.
Search for thread 'best 5" midrange driver' or something like that.
Otherwise one aurum cantus performs really well.
When it comes to tweeter, nice planar beston is really cheap but sounds as good as expensive ribbons.
Search for thread 'best 5" midrange driver' or something like that.
Otherwise one aurum cantus performs really well.
When it comes to tweeter, nice planar beston is really cheap but sounds as good as expensive ribbons.
The Satori line seems to do very well. I hear a lot about the Dayton Reference series as well. But the DR are very inexpensive, much less than Satori's and not a ton of measurements. But they may be cheap enough to buy and measure a few of them.
How do you determine breakups? Peerless does look pretty good. I am no opposed to going with 3-5" mids at all, its just about performance, it could be an 8" mid for all I care! My concern was trying to get the crossover at a discreet point and 80hz is the agreed upon frequency for subs by many. I'll check out those names, thanks!
Crossover points are dictated by the size of the drivers you select. If it can be 8" or whater mid, then you have to select massive tweeter which can be crossed low. Or small fullrange.
Whats breakup? Are you serious?
Next you ask whats beaming?
Whats breakup? Are you serious?
Next you ask whats beaming?
A better comparison to the Dayton Reference series is probably the SB Acoustics standard line. Not the premium Satori brand.
You didn't specify whether your budget amounts were for a single driver or for a pair. If for a single driver, then the Satoris should fit into the budget. If for a pair, then you probably need to step down to the standard line.
I was asking on the responses, can you see the breakups in the measurements. Is that where the response gets crazy higher up? I would rather sound stupid then misunderstand.
Yes. Thats it, roughly.
Its where cone, from normal operation, pistonic movement, following the audio signal closely, stops being pistonic and becomes too big and parts of the cone move forward, parts move back. Sort of like flapping flag. This results is sharp peaks and dips. Obviously, its does not follow audio signal accurately and has nothing to do with hifi.
Some speakers, like metal cone, aluminium, magnesium, titanium for example, can have massive breakup. Other materials, like paper, have many smaller breakups. Some plastic based cones can have almost nonexistent breakups, suppressed. But many materials are combined, breakups vary. You do not want to listen to breakups. Up and down response is only part of the story. There is time domain, these resonances exist longer than audio signal, they ring and smear the signal. Best to look at waterfall plots. Thats the best way the see speaker behaviour when it comes to breakups.
https://www.waterfallaudio.com/en/technology/
https://forums.audioholics.com/forums/threads/a-couple-questions-about-cone-breakup.81696/
Its where cone, from normal operation, pistonic movement, following the audio signal closely, stops being pistonic and becomes too big and parts of the cone move forward, parts move back. Sort of like flapping flag. This results is sharp peaks and dips. Obviously, its does not follow audio signal accurately and has nothing to do with hifi.
Some speakers, like metal cone, aluminium, magnesium, titanium for example, can have massive breakup. Other materials, like paper, have many smaller breakups. Some plastic based cones can have almost nonexistent breakups, suppressed. But many materials are combined, breakups vary. You do not want to listen to breakups. Up and down response is only part of the story. There is time domain, these resonances exist longer than audio signal, they ring and smear the signal. Best to look at waterfall plots. Thats the best way the see speaker behaviour when it comes to breakups.
https://www.waterfallaudio.com/en/technology/
https://forums.audioholics.com/forums/threads/a-couple-questions-about-cone-breakup.81696/
Last edited:
@tim405's advice is spot on.
OK, so you'll be using DSP so you can EQ the eff out of anything. But will it sound good?
I'd recommend taking a step back and simplifying your design parameters. Think about acoustics in your listening room, listening distance, room size, distance to boundaries, how loud you need to go.
Then look at measurements in respect of those constraints, then look at drivers, and how to match something to these particular 10"s (sensitivity, dispersion, distortion performance in a given bandwidth), then how to match something to a given mid.
OK, so you'll be using DSP so you can EQ the eff out of anything. But will it sound good?
I'd recommend taking a step back and simplifying your design parameters. Think about acoustics in your listening room, listening distance, room size, distance to boundaries, how loud you need to go.
Then look at measurements in respect of those constraints, then look at drivers, and how to match something to these particular 10"s (sensitivity, dispersion, distortion performance in a given bandwidth), then how to match something to a given mid.
Last edited: