-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
miniDSP 2.1 solution finder
- Thread starter dom_st
- Start date
- Thread Starter
- #42
OK again only limited time, but did two measurements with the speakers only.
this is left and right speaker, 40+cm distance from the wall behind.
left = green
right = blue
This is in the "soft" speaker setting, damping the highs a bit. i can select the switch between soft, linear and brilliant. I think i will stick with linear for the further tests.
This is the same setup but with the switch for the highs set to linear response instead of the soft setting. Seems to have a effect, as there is some slower roll off agt the end of the frequency range and maybe a 1-2dB boost.
left = green
right = blue
By the way the hypex amp is now connected to the pre-outs of the yamaha.
So Wiim --> miniDSP --> Yamaha --> Amp
Didn´t touch the subwoofer so far, will do that last as i probably won´t need it if i make the floor standers sound great ... but on the other hand, a sub is a sub
I will follow up with measurements and auto EQ filter settings and their results and screenshots maybe later, but definitely tomorrow as i think i have much more time then.
this is left and right speaker, 40+cm distance from the wall behind.
left = green
right = blue
This is in the "soft" speaker setting, damping the highs a bit. i can select the switch between soft, linear and brilliant. I think i will stick with linear for the further tests.
This is the same setup but with the switch for the highs set to linear response instead of the soft setting. Seems to have a effect, as there is some slower roll off agt the end of the frequency range and maybe a 1-2dB boost.
left = green
right = blue
By the way the hypex amp is now connected to the pre-outs of the yamaha.
So Wiim --> miniDSP --> Yamaha --> Amp
Didn´t touch the subwoofer so far, will do that last as i probably won´t need it if i make the floor standers sound great ... but on the other hand, a sub is a sub
I will follow up with measurements and auto EQ filter settings and their results and screenshots maybe later, but definitely tomorrow as i think i have much more time then.
I like to think of EQ as a subtractive process. It only eliminates tonal problems. Adding the sub is the big deal, letting the speakers relax a bit and broadening the range of your system can bring a lot of life back into what EQ removes. Even those bass towers are probably better off handling midbass where I think they cross (around 300-500hz) and giving the rest to a sub.
It looks like the graphs have dips and peaks that oppose each other left to right. This may mean they sum well, but that would only be for mono content. The MiniDSP can allow correction of each speaker, a huge selling point. But it may not be worth it to fix these room dips. Peaks can be clipped, but dips are hard to fill. That's where room treatment and placement come in. I wouldn't really worry about it too much. Psychoacoustics states that the peaks are more noticeable anyways. Setting up a sub in a different part of the room (like a side wall) may allow you to fill some of those dips in the sub bass. There are actually a ton of ways to handle a subwoofer, from matching the natural rolloff of the mains, to actively crossing over, or shelving. I won't get too far into it, but allowing a full range speaker to be supplemented by a subwoofer is on the table if the mains can keep up and all the dips and peaks level out at the end. Most sub-bass content ends up being mono anyways, so getting it from three sources might be one move. You could always blow up your towers, though. I only read so much about them.
When a sub is set up right, the results in a well recorded film can be downright shocking. Sudden and loud noises can make me jump out of my seat a bit or wince (in a good way), and the same kind of energy can be applied to music.
This setup is super cool, by the way. Wish I could hear it.
It looks like the graphs have dips and peaks that oppose each other left to right. This may mean they sum well, but that would only be for mono content. The MiniDSP can allow correction of each speaker, a huge selling point. But it may not be worth it to fix these room dips. Peaks can be clipped, but dips are hard to fill. That's where room treatment and placement come in. I wouldn't really worry about it too much. Psychoacoustics states that the peaks are more noticeable anyways. Setting up a sub in a different part of the room (like a side wall) may allow you to fill some of those dips in the sub bass. There are actually a ton of ways to handle a subwoofer, from matching the natural rolloff of the mains, to actively crossing over, or shelving. I won't get too far into it, but allowing a full range speaker to be supplemented by a subwoofer is on the table if the mains can keep up and all the dips and peaks level out at the end. Most sub-bass content ends up being mono anyways, so getting it from three sources might be one move. You could always blow up your towers, though. I only read so much about them.
When a sub is set up right, the results in a well recorded film can be downright shocking. Sudden and loud noises can make me jump out of my seat a bit or wince (in a good way), and the same kind of energy can be applied to music.
This setup is super cool, by the way. Wish I could hear it.
- Thread Starter
- #44
Appreciate your input, thanks a lot!
My idea of a proper approach was try to tune the speakers so they work great and then add the bass if required, but yeah maybe the other way around makes more sense.
Here some info on the woofer, my sub is the old Nubert AW-1000.
It´s probably better suited for movies than music but it does go low and loud if required, so i understand what you describe there
Personally i like if the music has some nice foundation, doesn´t need to be very loud but yeah nice smooth bass to support the speakers is the goal, now.
I have repeated the measurement with both speakers active, this is the red line.
The dip between 50 and 100Hz is quite big as it seems, but bass seems to be quite present.
Not sure what the UMIK is measuring at 20Hz, as the sub is not active right now.
My idea of a proper approach was try to tune the speakers so they work great and then add the bass if required, but yeah maybe the other way around makes more sense.
Here some info on the woofer, my sub is the old Nubert AW-1000.
It´s probably better suited for movies than music but it does go low and loud if required, so i understand what you describe there
Personally i like if the music has some nice foundation, doesn´t need to be very loud but yeah nice smooth bass to support the speakers is the goal, now.
I have repeated the measurement with both speakers active, this is the red line.
The dip between 50 and 100Hz is quite big as it seems, but bass seems to be quite present.
Not sure what the UMIK is measuring at 20Hz, as the sub is not active right now.
- Thread Starter
- #47
And this is the left speaker with the filter.
green = without any EQ
blue = with EQ
So it seems the linearity is definitely improved. But the area from 20-100Hz is untouched.
So i would conclude that i could put my crossover for the left speaker at maybe around 95Hz and let the rest do the subwoofer?
green = without any EQ
blue = with EQ
So it seems the linearity is definitely improved. But the area from 20-100Hz is untouched.
So i would conclude that i could put my crossover for the left speaker at maybe around 95Hz and let the rest do the subwoofer?
- Thread Starter
- #48
I hope it´s not hurting anyone if i spam this thread. Maybe my journey helps someone in the same situation.
This is the suggested filter for the right speaker. I wonder why there is no filter, trying to close the gap at 100 Hz, could i try and fix that manually?
This is the suggested filter for the right speaker. I wonder why there is no filter, trying to close the gap at 100 Hz, could i try and fix that manually?
- Thread Starter
- #49
And this is the right speaker, with filtering (purple) and without (blue). Really nice to see the effect.
Eventhough there are still some few dB peaks and drops here and there, but maybe that is irrelevant?
And now i guess i would measure both speakers together and see what i get.
Eventhough there are still some few dB peaks and drops here and there, but maybe that is irrelevant?
And now i guess i would measure both speakers together and see what i get.
- Thread Starter
- #50
ok this is unexpected.
red was the old measurement, where left and right speaker were measured without any EQ.
green is with EQ, it looks better throughout, but not a massive difference.
i will cut off the main speakers now at 90Hz and introduce the sub.
red was the old measurement, where left and right speaker were measured without any EQ.
green is with EQ, it looks better throughout, but not a massive difference.
i will cut off the main speakers now at 90Hz and introduce the sub.
- Thread Starter
- #52
Ok i seem to have a problem here. Not sure why the bass is so super loud even though i put it´s volume to maybe just 25%, everything was shaking here.
The issue i have is, the sub has a line in and a line out, as well as high level inputs.
How do i connect the balanced miniDSP flex output properly.
I have TS to RCA converters, TRS to RCA converters what would be correct to use and
from there ???
The issue i have is, the sub has a line in and a line out, as well as high level inputs.
How do i connect the balanced miniDSP flex output properly.
I have TS to RCA converters, TRS to RCA converters what would be correct to use and
from there ???
There's a lot to unpack here, so I'm just gonna point out what I think is the most important and overlooked aspect here.
The "Target level" is a tool that can help or hurt your outcome here tremendously. It's where you sacrifice sensitivity for flatness. REW will attempt to fill dips and further cut peaks until the output is ruler flat. Some dips shouldn't or can't be fixed. Boosting bass this way can make speakers explode. Ensure you have stiff cutoffs when boosting ported bass or things will shatter and amplifiers will in unfortunate ways. One of the reasons to buy the wonder Class-D amps is their ability to shrug off high power usage and how well they can handle failures when designed well.
Setting the gain in the miniDSP should keep the sub from tearing your house down. Letting the DSP handle it should be better overall. You can also run a compressor for safety. It may make things sound worse while it's setting up, but it will keep electronics from breaking and may save drivers from being overworked.
I don't have a ton of time right now, so I hope this leads you on the right path.
The "Target level" is a tool that can help or hurt your outcome here tremendously. It's where you sacrifice sensitivity for flatness. REW will attempt to fill dips and further cut peaks until the output is ruler flat. Some dips shouldn't or can't be fixed. Boosting bass this way can make speakers explode. Ensure you have stiff cutoffs when boosting ported bass or things will shatter and amplifiers will in unfortunate ways. One of the reasons to buy the wonder Class-D amps is their ability to shrug off high power usage and how well they can handle failures when designed well.
Setting the gain in the miniDSP should keep the sub from tearing your house down. Letting the DSP handle it should be better overall. You can also run a compressor for safety. It may make things sound worse while it's setting up, but it will keep electronics from breaking and may save drivers from being overworked.
I don't have a ton of time right now, so I hope this leads you on the right path.
Last edited:
I have another minute so I'm going to try to help you make sense of the REW EQ box I marked up.
In blue, the treble and bass can be adjusted to match your preferences. This can include full-blown house curves, like a harman curve or a line on the equal loudness countours. Simply put, the rise or cut can start on any selected frequency. My desktop nearfield monitors have -1.5dB/octave starting at 1500hz, and a 3dB/octave rise starting at 150hz. I have it set up where I can turn those boosts off easily, because I set up a "flat" EQ and a "boosted" profile. That's for another time.
The red section is where all of the sacrifices happen for on-axis response. In a passive speaker, the amplifier sends full power for each frequency to the crossover, which divides the energy across each driver. These components dissipate extra energy as heat, which allows them to play more accurately each frequency they are asked to. This is different from DSP, which tells the amplifier to do less or more per frequency. This is important as it will change the instructinos your amplifier is given, unlike passives. Understanding this difference is key to utilizing DSP to it's fullest extent.
As I suggested in the photo, every three decibels of gain increases the wattage by approximately double. The reality is more complicated, as are logarithms, but we'll stick with the easy explanation. This means changing the "sensitivity" of the speaker to allow more room for corrections by changing the Target Level. If the mic calibrations are correct, 72dB takes only a few hundred miliwatts from an 85dB sensitivity tower. Bringing that sensitivity down 3dB requires so little extra power when bringing volumes back up to listening levels, only another 100mW or so. The best advantage of our new amps is having a whole 500 or so watts available, which makes the issue of trading lower sensitivity for accuracy trivial. Any driver that is also brought down this way has had it's power handling superficially increased. There will be less power flowing at those frequencies. Likewise, any driver that is boosted or stays the same receives the full brunt of the amplifier power. Since your speakers are ported, I would always test them with a high-pass filter on the miniDSP (where you do the crossover now) around 45hz to protect the woofers. Porting is fascinating, and I recommend learning about it. Pushing speakers below their port tuning can damage them, as they are in their greatest physical extension. You could also easily lower the subwoofer crossover to 60-80hz. The specs of the tower justify that. And that subwoofer is measuring incredibly well up to 80hz.
In purple are the peak settings. An EQ "Peak" can be up or down, positive or negative. The width is the "Q", inverted for your convenience with smaller numbers making wider bigger peaks. "Gain" is up or down, increasing or decreasing volume. In REW, they are calling this "Boost". What REW is asking you is "how much do you want to push your amplifier, bro?" I am no amplification expert, but if you have 6dB of gain on a filter (pretend one of your eq peaks is inverted, raising the volume of a frequency band) it will demand four times more power from the amplifier than if the gain were 0, while the amplifier is playing those notes, or frequencies.
Here is my EQ on my monitors currently. Not the same software, but the same idea. Note the gain in the bass. I would demand a lot from my amplifier to run like this, and I would likely cross the line level threshold (in your case, 4V) and cause an amplifier shutdown. Note the global gain of 0dB.
Here, I have lowered the global gain (also known as preamp volume, preline vol, etc.) to never pass above 0dB. On the MiniDSP, this is "Gain (dB)" as I circled above. This lowers the sensitivity of my speakers considerably, but extends their range. I can't play the speakers as loudly this way, as the small diameter woofer is always going to have a maximum SPL of about 86dB down at 50hz. I am simply lowering the volume of the rest of the speaker to meet the weak link. If I were to use a subwoofer on this setup, there wouldn't be a boost down low. No negative global gain, no volume limited by bass output.
To relate this back to our REW settings, you can select the range you would like to correct with the Match Range section. If I were setting up a set of mains for a subwoofer, I would pick something like 80hz to 20000hz. Then to set up the sub, I might say 0hz (or whatever is the lowest) to 80hz when adjusting the subwoofer. When designing speakers, you can choose ranges like 1.2khz to 20khz for a tweeter, for example. Individual Max Boost is how much you are willing to have a single band of EQ boost a range. There are negatives to overboosting certain frequencies, but you can safely put this as high as you want. Overall Max Boost is important because multiple peaks can come together to create a larger "overall boost" at a certain frequency. In the case of my monitors, this is 8dB, as I am willing to apply -8dB of gain to my speakers to prevent damage. On the miniDSP, this is the "Gain (dB)" setting I circled in red before. Flatness Target is how accurate you would like the speaker to be. Most speakers are judged as good when they are flat to 3dB (+/-). Though lower is better, I guess. It depends on how many bands the new MiniDSP's let you have.
You may notice some peaks and dips that don't change no matter how much EQ you throw at them. These are the room and the listening position in the room. Any EQ related to these can be safely ignored. Sometimes a speaker will reveal quirks and resonances when pushed at certain frequencies. Isolating these issues can be difficult.
I think this is the best I'm going to be able to do for the next few days. I'm sure someone else can help you whlie I'm away. This is a really fun time to get into projects like this and I'm sure you'll be satisfied with the results. There are many more qualified than me who can help you over the weekend if you choose to ask. Good luck!
In blue, the treble and bass can be adjusted to match your preferences. This can include full-blown house curves, like a harman curve or a line on the equal loudness countours. Simply put, the rise or cut can start on any selected frequency. My desktop nearfield monitors have -1.5dB/octave starting at 1500hz, and a 3dB/octave rise starting at 150hz. I have it set up where I can turn those boosts off easily, because I set up a "flat" EQ and a "boosted" profile. That's for another time.
The red section is where all of the sacrifices happen for on-axis response. In a passive speaker, the amplifier sends full power for each frequency to the crossover, which divides the energy across each driver. These components dissipate extra energy as heat, which allows them to play more accurately each frequency they are asked to. This is different from DSP, which tells the amplifier to do less or more per frequency. This is important as it will change the instructinos your amplifier is given, unlike passives. Understanding this difference is key to utilizing DSP to it's fullest extent.
As I suggested in the photo, every three decibels of gain increases the wattage by approximately double. The reality is more complicated, as are logarithms, but we'll stick with the easy explanation. This means changing the "sensitivity" of the speaker to allow more room for corrections by changing the Target Level. If the mic calibrations are correct, 72dB takes only a few hundred miliwatts from an 85dB sensitivity tower. Bringing that sensitivity down 3dB requires so little extra power when bringing volumes back up to listening levels, only another 100mW or so. The best advantage of our new amps is having a whole 500 or so watts available, which makes the issue of trading lower sensitivity for accuracy trivial. Any driver that is also brought down this way has had it's power handling superficially increased. There will be less power flowing at those frequencies. Likewise, any driver that is boosted or stays the same receives the full brunt of the amplifier power. Since your speakers are ported, I would always test them with a high-pass filter on the miniDSP (where you do the crossover now) around 45hz to protect the woofers. Porting is fascinating, and I recommend learning about it. Pushing speakers below their port tuning can damage them, as they are in their greatest physical extension. You could also easily lower the subwoofer crossover to 60-80hz. The specs of the tower justify that. And that subwoofer is measuring incredibly well up to 80hz.
In purple are the peak settings. An EQ "Peak" can be up or down, positive or negative. The width is the "Q", inverted for your convenience with smaller numbers making wider bigger peaks. "Gain" is up or down, increasing or decreasing volume. In REW, they are calling this "Boost". What REW is asking you is "how much do you want to push your amplifier, bro?" I am no amplification expert, but if you have 6dB of gain on a filter (pretend one of your eq peaks is inverted, raising the volume of a frequency band) it will demand four times more power from the amplifier than if the gain were 0, while the amplifier is playing those notes, or frequencies.
Here is my EQ on my monitors currently. Not the same software, but the same idea. Note the gain in the bass. I would demand a lot from my amplifier to run like this, and I would likely cross the line level threshold (in your case, 4V) and cause an amplifier shutdown. Note the global gain of 0dB.
Here, I have lowered the global gain (also known as preamp volume, preline vol, etc.) to never pass above 0dB. On the MiniDSP, this is "Gain (dB)" as I circled above. This lowers the sensitivity of my speakers considerably, but extends their range. I can't play the speakers as loudly this way, as the small diameter woofer is always going to have a maximum SPL of about 86dB down at 50hz. I am simply lowering the volume of the rest of the speaker to meet the weak link. If I were to use a subwoofer on this setup, there wouldn't be a boost down low. No negative global gain, no volume limited by bass output.
To relate this back to our REW settings, you can select the range you would like to correct with the Match Range section. If I were setting up a set of mains for a subwoofer, I would pick something like 80hz to 20000hz. Then to set up the sub, I might say 0hz (or whatever is the lowest) to 80hz when adjusting the subwoofer. When designing speakers, you can choose ranges like 1.2khz to 20khz for a tweeter, for example. Individual Max Boost is how much you are willing to have a single band of EQ boost a range. There are negatives to overboosting certain frequencies, but you can safely put this as high as you want. Overall Max Boost is important because multiple peaks can come together to create a larger "overall boost" at a certain frequency. In the case of my monitors, this is 8dB, as I am willing to apply -8dB of gain to my speakers to prevent damage. On the miniDSP, this is the "Gain (dB)" setting I circled in red before. Flatness Target is how accurate you would like the speaker to be. Most speakers are judged as good when they are flat to 3dB (+/-). Though lower is better, I guess. It depends on how many bands the new MiniDSP's let you have.
You may notice some peaks and dips that don't change no matter how much EQ you throw at them. These are the room and the listening position in the room. Any EQ related to these can be safely ignored. Sometimes a speaker will reveal quirks and resonances when pushed at certain frequencies. Isolating these issues can be difficult.
I think this is the best I'm going to be able to do for the next few days. I'm sure someone else can help you whlie I'm away. This is a really fun time to get into projects like this and I'm sure you'll be satisfied with the results. There are many more qualified than me who can help you over the weekend if you choose to ask. Good luck!
Last edited:
- Thread Starter
- #57
I am just experimenting with the switch settings for highs and bass on my mains and how they interact with my room.
There is a bass boost switch at the back of my mains that i probably never used, result is visible in red, that seems to partially fill the gap between 50 and 100Hz, for my left speaker.
The small variations along the chart are maybe due to my mic being slightly moved by a few cm, but should be more or less where it was yesterday.
the right speaker in linear mode (highs) and bass (boosted) measures like this (brownish). boosting bass here probably doesn´t make sense.
and this is l+r together, bass boosted. looks pretty flat to me (but +/-5dB or even more probably isn´t great). again this is without any PEQ activated.
so this doesn´t look too bad, if i apply a harman target curve.
I think the target curve would look even better if s set the highs to soft instead of linear and leave the bass boosted.
And now finally with sub (EQed and gains -28dB)
green line only mains full range, no sub.
blue line mains with xo at 60Hz, no sub.
purple line, mains with xo at 60Hz, with sub cut-off at 60Hz.
There is a bass boost switch at the back of my mains that i probably never used, result is visible in red, that seems to partially fill the gap between 50 and 100Hz, for my left speaker.
The small variations along the chart are maybe due to my mic being slightly moved by a few cm, but should be more or less where it was yesterday.
the right speaker in linear mode (highs) and bass (boosted) measures like this (brownish). boosting bass here probably doesn´t make sense.
and this is l+r together, bass boosted. looks pretty flat to me (but +/-5dB or even more probably isn´t great). again this is without any PEQ activated.
so this doesn´t look too bad, if i apply a harman target curve.
I think the target curve would look even better if s set the highs to soft instead of linear and leave the bass boosted.
And now finally with sub (EQed and gains -28dB)
green line only mains full range, no sub.
blue line mains with xo at 60Hz, no sub.
purple line, mains with xo at 60Hz, with sub cut-off at 60Hz.
Attachments
The whole thing looks top of the line! It must sound incredible. If I were to do anything, it would be playing with the target level and max gain. Setting max gain to 3 and subtracting 3 from the target level, then subtracting 3dB from each line’s global gain inside the miniDSP will put you on the path to perfect. If the volumes are still loud enough and if you can still hear an improvement, those numbers can get cranked pretty high. Though you have had a stellar start with your choices, so there’s a lot less to fix.
Last edited:
- Thread Starter
- #59
I will read that again and try that these days. (thumbs up)
And i found one mayor issue, as i was working with the MiniDSP default "96kHZ model" instead of the 2x4 HD type. That´s why REWs auto EQ only provided 5 filters instead of 10 possible ones.
This is the result of a random position of the right speaker, so twice the amount of filters
Similar threads
