Hi Serkan, Thank you for uploading all those amazing videos. I have a question/clarification about the phase correction in your video. On 6:40 you said the phase is corrected and on the screen is visible that the phase is wraping 26 times from +180 to -180 between 50 HZ and 10kHz what should be 9360 deg difference. So, I do not understand what phase is corected and on which way?
You did really count the discontinuities 😂 Every 360 degrees shift results in the same phase response because the waveforms become similar and do not counteract each other. You can also think of the unit circle (pole-zero conversion of the frequency domain) in the complex plane and every 2π addition to the angle will result pointing to the same point in the circle.
Obviously I misunderstand something. I do not want to bother or troll you but my understanding is different and possibly completely wrong so I would like to have some more explanation. If it is like you said, why to do phase correction for LR4 as it is in phase, acctually 360 deg out of phase, same point on the circle as you point out. This is how I see the phase . That circle and every 360 deg same point theory is aplicable (maybe) for cotinuous tone but what is going on with the guitar pluck. From your video we can see that phase diference between 500 and 1000 Hz is 5 times 360 deg so 5 periods what is for 1kHz 5 ms. My understanding is, 2nd harmonic of the 500 Hz guitars pluck will arrive 5ms earlier and smear tone attack and change the pitch sense As I wrote, maybe I am completely wrong so please correct me in that case
Hello Serkan, thanks for your nice videos! Your explanation help me to understand the whole topic so much better! :) I want to measure a 2.1 Setup. Shall i point the Umik-1 up to ceiling or straight to the stereo speakers? My subwoofer is place on the opposite side near to my couch.
Uploaded biquad to my output peq and my dbfs scale meters on those channels go to a constant 3 and minidsp is non functioning on all configurations not just the one I uploaded the biquad too. To fix I have to factory reset unit and unplug.
Hi ! thanks for your tutorials ! I watched the video and at the end you import the filters in the cross over section of the mini dsp. I'm a bit confused as i got a 2.1 system and the cross over are already fitted. How can i apply those corrections with a sub ? And do i have to take the mesurement with the sub or each speaker alone ? And last question : as far as i know my sub doesn't have any filter, so is this tutorial make sense for it ? Many thanks again, your vids are passionating, i'm new in the digital correction and i really want to get the most of it !
It doesn't matter where you upload the biquad filters. They do what they are supposed to do when processed. For subs normal PEQ filters are very effective.
@@ocaudiophile This seems very powerful but I’ve only worked with PEQ to this point. I’m a little confused as to weather or not there is any benefit for biquads when working with retail speakers. You mentioned in another comment that this is something like what Dirac does but Dirac may not get it correct. But, Dirac is working on complete speakers and multiple speakers where their crossovers are fixed. I’m just wondering if biquads are any use to me in a normal 2.1 or 5.1 system. I’d really like to dive into it if there is but not sure what I would be doing it for. Can you shed any light on this? Thanks.
@@williamhicks2763 Thanks. Firstly, the time inverted biquads explained in the video don't work in recent MiniDSP models (I know from subscriber feedback) but you can still use normal allpass filter (non-time inverted phase filters that don't alter frequency response). These are still quite useful to match speaker and sub phase responses at the crossover region for example. MSO uses them for subwoofer/speaker alignment. So does Dirac.
Thanks for the great video. I understand that in your video you are phase aligning a pair of speakers that already has a passive crossover in them. I was thinking if this can be used on active speakers where the MiniDSP is also used as active crossover. As far as I know IIR crossover filters have identical behavior to capacitors/inductors in passive crossovers, so when using an IIR Crossover filter, I guess this might be a way to correct active speakers as well? For example. Most MiniDSP's have advanced biquad programming available with 8 biquads per channel. If I wanted to create a bandpass filter for a midrange in a multi-way speaker with for example LR24 for 2500hz low pass and 500hz high-pass could I then do something like this: biquad #1-4 - crossovers: - HPF 500hz LR 24db with coeficients from MiniDSP spreadsheet (cascading if needed) - LPF 2500hz LR 24db with coeficients from MiniDSP spreadsheet (cascading if needed) biquad #5-8 - phase linearization of crossovers: - 500hz second order time inverted all pass filter - 2500hz second order time inverted all pass filter Would above actually make a phase linear 500Hz HPF + 2500Hz LPF from the digital domain point of view, where the all pass filter reverse the phase shift effect of the IIR Crossover filters? Here I'm not trying to factor in cabinet, room modes, woofers/drivers FR response/phase shift behaviour - just whether the MiniDSP will output a linear phase crossover? And if so should there by any concerns in regards to digital artifacts created by all pass filters (like ringing and all other kind of terms I cannot remember right now)?
In theory it could be done however the phase filters need to be time-reversed and most MiniDSP models refuse to accept the biquad coefficents of time-reversed all pass filters. A better option is to use FIR filters in MiniDSP for phase correction. You can dedicate each speaker 2042 taps of FIR filters even in 2x4HD and all bar the lowest frequency box phase shifts can be corrected with that. Use repHase, 99-100% centering and rectangular windows to achieve good approximations with low tap counts. You can design your crossover filters outside the FIR slots and they will work in parallel. This video can help you: ua-cam.com/video/ChPu0u3nZxc/v-deo.html
Do you know how the MiniDSP plugin will act if it does not support biquad coefficents of time-reversed all pass filters? Will it refuse to apply or pop-up an error if trying to configure them, or will it accept the commands, but output distorted sound? None of my MiniDSP platforms supports FIR, unfortunately (I currently have SHD and C-DSP 8x12 v2.0). I've used the C-DSP 8x12 in the car, but now it's just sitting on the shelf, so was planning to use it for active speakers in home audio, which I've seen a few people do in various forums with success (just have to be cautious about ground loops as it was not designed for home use). It lacks FIR filters (but have DIRAC live instead as an option). Need to test if these time reversed all pass works on it... It uses an older plugin, so it might. My goal is to make a 4-way active speaker. I guess I could make a setup where my SHD is the main device and then use the 4x digital outputs for two 2x4HD, Flex or MiniSHARC/DRC-DI. It would give me 4 channels each, for left speaker and right speaker, but with a dedicated DSP for each speaker. It would not be that clean to have 1xSHD + 2xDSP FIR crossover processors (one for each speaker), but guess that is how I could get FIR filter crossovers with enough taps and outputs. I also know there is a Flex Eight, but that only supports FIR filters on the 2x inputs, and not the outputs like the normal Flex. But right now I'm not sure FIR filters is worth the investment. Many great speakers have been made using classic IIR filters or passive crossovers that behave similarly. So guess I will try out the C-DSP 8x12 first and listen to the results, before considering FIR compatible devices to replace it. But these time reversed all-pass filters using biquads is still interesting, and is definitely something to try out to improve the phase of IIR crossovers. Just hope it works with C-DSP 8x12 :)
Your comment was left in UA-cam spam folders for some reason, only just saw it... I think, it will accept them in the plugin but just mute the MiniDSP when applied.
@@ocaudiophile I tried on the MiniDSP C-DSP 8x12 V2.0, and when entering inverted all pass filters into the EQ biquad section they are definitely not stable (overload of output channels that puts channel into error state and requires reboot to restore unit, so the outputs work again). In biquad section for crossovers I experienced that output level would decrease like 80-100db, but music would play. But I do not believe the lower output level is related to filters generated by your spreadsheet. I tried filter from MiniDSP's spreadsheet as well and got the same result of 80-100db reduction in output level. Not sure what Im doing wrong, but I didn't really bother to try and fix it. Im sure if I figure out what Im doing wrong in defining all 8x biquads for the xover I would get same error as when entering your biquads into EQ biquad section (overload of output channel). For reference, my approach was to use biquad1+biquad2 with values generated by spreadsheets and biquad3-8 was just 0's (except b0=1). I think we can conclude they are not supported for the C-DSP 8x12 at least, and that platform is not stable with them. Im thinking about buying a MiniDSP Flex Eight instead to get FIR filters at the input channels to create inverted all pass filters for crossover linearization, and just be done with it :D But thanks for taking an interest and replying :)
@@Salkcin87, Minidsp 2x4hd didn't work either. Caused my dbfs meter to peg to 3 and stay there until factory reset and unplug. I tried everything I could think of and same result.
Another great Video! Your starting point here again is a combined measurement from left and right speaker? I understand I could also use just Left OR Right and apply the same (!) filtering to the other speaker in the end...? For phase correction with a 2x4HD I understand I would only have to follow the first steps you described and just upload the FIR-Filter (bin-file) to the 2x4HD, correct (so no need for manual all pass filtering and bi-quads....)?
Thank you for your kind words. You should always use same xo and box phase correction for left & right speakers, it's sometimes easier to see the phase response with a stereo measurement. Your ONLY option with 2x4 HD is using FIR filters. It cannot process time-inverted allpass filters (causes a buzzing sound). Unless you use low frequency port corrections, 1024 taps will be enough if you use rectangular windowing and 99%-100% centering.
@@ocaudiophile okay, will try....😉 I think I can even use up to approx 2k taps if I only use two channels of the 2x4HD. What about full range speakers (single drivers) that use a small correction network but no real x-over...? Will there be anything to correct? Haven't measured phase response yet...🤔
@@ocaudiophile okay...👍 l now tried to do it in my own. After I created a correction file in rePhase and exported it in the wav format for REW I uploaded it to REW and did the multiplication. Unfortunately I can't do the estimate IR delay calculation for A times B and thus I can't time align the before and after measurements/calculation... REW says A Times B does not have an impulse response and so cannot be time aligned.... Any idea what the mistake was here....? Any help would be appreciated!😃
Hello OCA, I have a question about a general approach to optimize my 7.1.6 Atmos/Auro setup. I'm thinking about to the following: - external amps and miniDSP for all speaker channels > precise time alignment of all speakers and subs for the MLP > what would be the optimal combination of filters and convolution that can be implemented with the miniDSP? > would that introduce a delay grater than 500ms to the audio signal of a video? (my favorite player for movies is PowerDVD and I can only set a maximum of -500ms there as negativ delay) - 4 subs with one miniDSP, optimized with MSO biquads for input (10 for all subs), output and through crossover function (18 for each sub) > use target curve within MSO > after that no EQ through audyssey anymore, only time alignment with the other speakers with the other miniDSPs > makes it sense to use the FIR filter for each sub on top to do some phase alignment? > should I not use the input BQ on top of output BQ to prevent phase issues? > what number and combination of filters, MSO, and miniDSP is the best approach for 4 subs? Would it even be reasonable in regards to sound improvement to go this way in comparison to your guide for Atmos/Auro calibration with audyssey hack? Or to put it in another way: If you had the possibility's of audyssey App, REW, MSO, UMIK-1, pre-amp and external-amp, miniDSP for each channel and sub > what would you do to get the best result for movies in Atmos/Auro?
MiniDSP 2x4HD FIR filters are uselss for subs but MSO can make use of some "allpass" filters with biquads to fix the phase response. 500ms PowerDVD delay would be enough. Your bass response was already quite nice so I wouldn't expect a major improvement but a slight improvement in bass can always be obtained with a lot of effort. External amps will certainly improve the sound comapred to class D power supplied by the AVR. The speaker crossovers and EQ will still be in the AVR though. You could use 2x8 channel MiniDSP units for bypassing even that but you would be better off with an AVR with Dirac ART if you are willing to invest that much.
@@ocaudiophile Thanks for the feedback. I wonder if ART will be at some point be integrate to existing DIRAC devices like the miniDSP 2x4 HD. But I don't know if this if even possible. And I would wait for even considering an ART AVR until it is available in the below 10K price range. The main reason for my question from the beginning was to bypass the EQ from AVR or pre-amp completely with miniDSP to use your methods of filter creation and convolution. And as you said the bass can already be made pretty good without ART. Do you have a source where I could learn more about "allpass" filters with MSO? A benefit of miniDSP would also be the ability to precisely dial in the delay between all speakers and the subs. So if I understand you right I could use convolution filters with miniDSP and would be able to have synchronized video and sound? So the question for me is still to choose between your method with audyssey or your method with miniDSP. With kind regards Frank
@---vn2kv avoid gain filters and EQ with MSO, just use allpass filters (you can use them in minidsp in biquad form) and time delay. Then EQ MSO optimised combibination's 'minimum phase' response in REW and add these final EQ filters to minidsp.
You mentioned that the minidsp can't invert the all pass for crossover time correction, but then you explain how your biquad calculator can make a normalized time reversed APF in biquad form that is comparable with the minidsp because of the a0 being 1. Does that mean you can use rephrase to find the APF needed for linearization, then make and input the biquad into the mini DSP to allow for crossover phase correction without the use of FIR?
I could use time reversed biquads in my old MiniDSP 2x4 balanced but some people had problems entering them with the newer hardware/firmware. We asked MiniDSP and they gave conflicting answers. It's possible that these biquads are not possible to implement in most machines. It seems your best option is to generate FIR taps in rephase for speaker phase correction. In rePhase, use rectangular windowing with 99%/100% centering to achieve satisfactory results with 1024 taps and avoid box correction.
@@picassoimpaler3243 Dirac corrects phase response to an extent but being an automated system, it's not optimal (i.e. it has no way of knowing every speaker and its crossover frequencies and filter orders). Still, symmetrical speaker setup and carefully taken multiple measurements will improve its results.
I downloaded your excel and noticed that your a1 and a2 are different on video. A1 number is negative on video, when on a file the A1 is positive number. Am i missing something?
Hi sir, I been learning alot from your videos. Im ready to purchase but find minidsp a bit over my budget and I require 4 channel in and out. How about if I use thomann t.racks DSP 4x4 Mini? Thank you in advance.
Thomann has only parametric equalizer no fir filter capability. But Minidsp only has 1024 fir taps capacity so, it is not even a real advantage. You can do a lot better DSP than anything in the market for free by using your computer as source (for up to 7.1 channels - no Atmos) and free dsp engines like Equalizer APO or Camilla
How to digital delay on eqapo or camila pc to hdmi? I even tried to use avr manual distance setting but not work. It does sound good though. Im using smaart v9. I can not seem to find many camiladsp tutorials do you have any or please make one kind sir, thankyou. @@ocaudiophile
I use JRiver since a long time now and I used Equliaser APO for a short time in the past but it certainly has speaker delay features for adjusting distances: sourceforge.net/p/equalizerapo/wiki/Configuration%20reference/#delay-since-version-09:~:text=%2D1.84776%200.729402-,Delay%20(since%20version%200.9),by%20480%20samples%20(10%20ms%20at%2048%20kHz)%0ADelay%3A%20480%20samples,-Copy%20(since%20version
@@ocaudiophile now I’m doing this I’m very happy. Smaart v9 capture live transfer file send over to REW then save untouched file and make new file with curves both files open in rephase linear phase for each 4500 5.1 audio channels export to eqapo.. imho way better than Dirac. I just ordered Iems measuring kit and about to do the same for moondrop starfields and AirPod pros 2. Later I will get a headphone measuring rig and try this method. Ath ad900x. Not sure if I can use live transfer don’t know if the sound card can loop back.
Nice tutorial. I have a questions. For multiple subs using minidsp, should I time align then eq the subs using minidsp amd after that running audyssey? Or just time align the subs using minidsp the run audyssey? Thank you
MiniDSP introduces delays which frequently makes it impossible to properly time align the subs with the rest of the system with the distance limits in the receiver. But if you're using it anyway, then phase align subs with each other - ua-cam.com/video/ga2eOwJRtXo/v-deo.html and EQ the combined response in DSP. You can then apply a final touch if necessary during Audyssey calibration - ua-cam.com/video/g26gbFdAIxE/v-deo.html
@@ocaudiophile other than mso, Just wondering, is it possible after finished align 2 subwoofers as your tutorial, after that I phase align the 3rd subwoofer with the previous 2 subwoofers phase alignment result. Thank you.
Thanks, Can you make video on Phase correction and Alignment where we face Subwoofer or speaker delay more than the limit of Audyssey or any other calibration software?
@@ocaudiophile ohh, thanks and great. But it will be helpfull if you make small video for those who are looking to experiment manual calibration. Because when I add miniDsp 2x4 HD for my subwoofer, it adds extra delay and I want to make it correct with phase alignment, not time alignment.
Hi, to do phase correction in Subwoofer frequency to match with mains, If I apply FDW of 15 cycle (AVR Denon X4700H) and create Minimum Phase virson copy and do EQ of that FDW+MP virson to upload in MiniDsp 2x4 HD, will this work for phase alignment? Without causing additional delay?
Dear Serkan, is there any chance to get in direct contact with you? Do you have a Homepage? Are you located in Germany? Could need your help for a High End speaker Project. All the best & thnk you for all these interesting Video´s - Markus
@@ocaudiophile I prefer your accent much more than AI, besides, you will just sound the same as the swamp of other AI videos out at the moment, I avoid them as I am sure many will.
amazing !
Cheers
Unbelievable!! I had no idea my MiniDSP 2x4 could do phase correction with bi-quads. Incredible knowledge, thank you!
Your assistant is also very knowledgeable
😂😂
Thanks mate,really helpful information.
Glad it helped
❤ thanks my friend for dis nice Video, you ar the best Man of UA-cam ❤
Thanks 😂
Hi Serkan, Thank you for uploading all those amazing videos. I have a question/clarification about the phase correction in your video. On 6:40 you said the phase is corrected and on the screen is visible that the phase is wraping 26 times from +180 to -180 between 50 HZ and 10kHz what should be 9360 deg difference. So, I do not understand what phase is corected and on which way?
You did really count the discontinuities 😂
Every 360 degrees shift results in the same phase response because the waveforms become similar and do not counteract each other. You can also think of the unit circle (pole-zero conversion of the frequency domain) in the complex plane and every 2π addition to the angle will result pointing to the same point in the circle.
Obviously I misunderstand something. I do not want to bother or troll you but my understanding is different and possibly completely wrong so I would like to have some more explanation.
If it is like you said, why to do phase correction for LR4 as it is in phase, acctually 360 deg out of phase, same point on the circle as you point out.
This is how I see the phase .
That circle and every 360 deg same point theory is aplicable (maybe) for cotinuous tone but what is going on with the guitar pluck. From your video we can see that phase diference between 500 and 1000 Hz is 5 times 360 deg so 5 periods what is for 1kHz 5 ms. My understanding is, 2nd harmonic of the 500 Hz guitars pluck will arrive 5ms earlier and smear tone attack and change the pitch sense
As I wrote, maybe I am completely wrong so please correct me in that case
An XO phase correction like LR4 corrects phase over a frequency range (ie 24 dB/octave). It's not a correction on a single frequency.
The waves are continuous and periodic. You can think of 360 deg shift laying on the next wave. It will still not change the waves shape.
Hello Serkan, thanks for your nice videos! Your explanation help me to understand the whole topic so much better! :)
I want to measure a 2.1 Setup. Shall i point the Umik-1 up to ceiling or straight to the stereo speakers? My subwoofer is place on the opposite side near to my couch.
Ceiling if you're gonna measure multiple positions
Uploaded biquad to my output peq and my dbfs scale meters on those channels go to a constant 3 and minidsp is non functioning on all configurations not just the one I uploaded the biquad too.
To fix I have to factory reset unit and unplug.
Damn you beating the crap out of that keyboard! Otherwise good job lol
Hi ! thanks for your tutorials ! I watched the video and at the end you import the filters in the cross over section of the mini dsp. I'm a bit confused as i got a 2.1 system and the cross over are already fitted.
How can i apply those corrections with a sub ?
And do i have to take the mesurement with the sub or each speaker alone ?
And last question : as far as i know my sub doesn't have any filter, so is this tutorial make sense for it ?
Many thanks again, your vids are passionating, i'm new in the digital correction and i really want to get the most of it !
It doesn't matter where you upload the biquad filters. They do what they are supposed to do when processed. For subs normal PEQ filters are very effective.
Unclear, are you correcting a single speaker? (bookshelf woofer/tweeter) or is this a single driver
single speaker with 3 drivers
But the box/port correction could be applied to a subwoofer as well...
@@ocaudiophile This seems very powerful but I’ve only worked with PEQ to this point. I’m a little confused as to weather or not there is any benefit for biquads when working with retail speakers. You mentioned in another comment that this is something like what Dirac does but Dirac may not get it correct. But, Dirac is working on complete speakers and multiple speakers where their crossovers are fixed. I’m just wondering if biquads are any use to me in a normal 2.1 or 5.1 system. I’d really like to dive into it if there is but not sure what I would be doing it for. Can you shed any light on this? Thanks.
@@williamhicks2763 Thanks. Firstly, the time inverted biquads explained in the video don't work in recent MiniDSP models (I know from subscriber feedback) but you can still use normal allpass filter (non-time inverted phase filters that don't alter frequency response). These are still quite useful to match speaker and sub phase responses at the crossover region for example. MSO uses them for subwoofer/speaker alignment. So does Dirac.
Amazing work. Do you offer virtual tuning/calibration
I help when people have difficulty getting their calibration right usually to test new methods but I've limited time to dedicate
Very informative! Do you know of any resources to learn the basics of filters? I'm grasping some of the concepts but not all.
Check the "DSP related" playlist in my channel 👍
Thanks for the great video.
I understand that in your video you are phase aligning a pair of speakers that already has a passive crossover in them. I was thinking if this can be used on active speakers where the MiniDSP is also used as active crossover.
As far as I know IIR crossover filters have identical behavior to capacitors/inductors in passive crossovers, so when using an IIR Crossover filter, I guess this might be a way to correct active speakers as well?
For example.
Most MiniDSP's have advanced biquad programming available with 8 biquads per channel.
If I wanted to create a bandpass filter for a midrange in a multi-way speaker with for example LR24 for 2500hz low pass and 500hz high-pass could I then do something like this:
biquad #1-4 - crossovers:
- HPF 500hz LR 24db with coeficients from MiniDSP spreadsheet (cascading if needed)
- LPF 2500hz LR 24db with coeficients from MiniDSP spreadsheet (cascading if needed)
biquad #5-8 - phase linearization of crossovers:
- 500hz second order time inverted all pass filter
- 2500hz second order time inverted all pass filter
Would above actually make a phase linear 500Hz HPF + 2500Hz LPF from the digital domain point of view, where the all pass filter reverse the phase shift effect of the IIR Crossover filters? Here I'm not trying to factor in cabinet, room modes, woofers/drivers FR response/phase shift behaviour - just whether the MiniDSP will output a linear phase crossover?
And if so should there by any concerns in regards to digital artifacts created by all pass filters (like ringing and all other kind of terms I cannot remember right now)?
In theory it could be done however the phase filters need to be time-reversed and most MiniDSP models refuse to accept the biquad coefficents of time-reversed all pass filters. A better option is to use FIR filters in MiniDSP for phase correction. You can dedicate each speaker 2042 taps of FIR filters even in 2x4HD and all bar the lowest frequency box phase shifts can be corrected with that. Use repHase, 99-100% centering and rectangular windows to achieve good approximations with low tap counts. You can design your crossover filters outside the FIR slots and they will work in parallel. This video can help you:
ua-cam.com/video/ChPu0u3nZxc/v-deo.html
Do you know how the MiniDSP plugin will act if it does not support biquad coefficents of time-reversed all pass filters? Will it refuse to apply or pop-up an error if trying to configure them, or will it accept the commands, but output distorted sound?
None of my MiniDSP platforms supports FIR, unfortunately (I currently have SHD and C-DSP 8x12 v2.0). I've used the C-DSP 8x12 in the car, but now it's just sitting on the shelf, so was planning to use it for active speakers in home audio, which I've seen a few people do in various forums with success (just have to be cautious about ground loops as it was not designed for home use). It lacks FIR filters (but have DIRAC live instead as an option). Need to test if these time reversed all pass works on it... It uses an older plugin, so it might.
My goal is to make a 4-way active speaker. I guess I could make a setup where my SHD is the main device and then use the 4x digital outputs for two 2x4HD, Flex or MiniSHARC/DRC-DI. It would give me 4 channels each, for left speaker and right speaker, but with a dedicated DSP for each speaker.
It would not be that clean to have 1xSHD + 2xDSP FIR crossover processors (one for each speaker), but guess that is how I could get FIR filter crossovers with enough taps and outputs.
I also know there is a Flex Eight, but that only supports FIR filters on the 2x inputs, and not the outputs like the normal Flex.
But right now I'm not sure FIR filters is worth the investment. Many great speakers have been made using classic IIR filters or passive crossovers that behave similarly. So guess I will try out the C-DSP 8x12 first and listen to the results, before considering FIR compatible devices to replace it.
But these time reversed all-pass filters using biquads is still interesting, and is definitely something to try out to improve the phase of IIR crossovers. Just hope it works with C-DSP 8x12 :)
Your comment was left in UA-cam spam folders for some reason, only just saw it...
I think, it will accept them in the plugin but just mute the MiniDSP when applied.
@@ocaudiophile
I tried on the MiniDSP C-DSP 8x12 V2.0, and when entering inverted all pass filters into the EQ biquad section they are definitely not stable (overload of output channels that puts channel into error state and requires reboot to restore unit, so the outputs work again).
In biquad section for crossovers I experienced that output level would decrease like 80-100db, but music would play. But I do not believe the lower output level is related to filters generated by your spreadsheet. I tried filter from MiniDSP's spreadsheet as well and got the same result of 80-100db reduction in output level. Not sure what Im doing wrong, but I didn't really bother to try and fix it. Im sure if I figure out what Im doing wrong in defining all 8x biquads for the xover I would get same error as when entering your biquads into EQ biquad section (overload of output channel). For reference, my approach was to use biquad1+biquad2 with values generated by spreadsheets and biquad3-8 was just 0's (except b0=1).
I think we can conclude they are not supported for the C-DSP 8x12 at least, and that platform is not stable with them.
Im thinking about buying a MiniDSP Flex Eight instead to get FIR filters at the input channels to create inverted all pass filters for crossover linearization, and just be done with it :D
But thanks for taking an interest and replying :)
@@Salkcin87, Minidsp 2x4hd didn't work either. Caused my dbfs meter to peg to 3 and stay there until factory reset and unplug. I tried everything I could think of and same result.
Another great Video! Your starting point here again is a combined measurement from left and right speaker? I understand I could also use just Left OR Right and apply the same (!) filtering to the other speaker in the end...? For phase correction with a 2x4HD I understand I would only have to follow the first steps you described and just upload the FIR-Filter (bin-file) to the 2x4HD, correct (so no need for manual all pass filtering and bi-quads....)?
Thank you for your kind words. You should always use same xo and box phase correction for left & right speakers, it's sometimes easier to see the phase response with a stereo measurement. Your ONLY option with 2x4 HD is using FIR filters. It cannot process time-inverted allpass filters (causes a buzzing sound). Unless you use low frequency port corrections, 1024 taps will be enough if you use rectangular windowing and 99%-100% centering.
@@ocaudiophile okay, will try....😉 I think I can even use up to approx 2k taps if I only use two channels of the 2x4HD.
What about full range speakers (single drivers) that use a small correction network but no real x-over...? Will there be anything to correct? Haven't measured phase response yet...🤔
@@andreasheiden7122 I have no experience with single driver speakers :(
@@ocaudiophile okay...👍 l now tried to do it in my own. After I created a correction file in rePhase and exported it in the wav format for REW I uploaded it to REW and did the multiplication. Unfortunately I can't do the estimate IR delay calculation for A times B and thus I can't time align the before and after measurements/calculation... REW says A Times B does not have an impulse response and so cannot be time aligned.... Any idea what the mistake was here....? Any help would be appreciated!😃
@@andreasheiden7122 you don't need to time align the phase correction, you need to remove its SPL offset though.
Hello OCA, I have a question about a general approach to optimize my 7.1.6 Atmos/Auro setup.
I'm thinking about to the following:
- external amps and miniDSP for all speaker channels
> precise time alignment of all speakers and subs for the MLP
> what would be the optimal combination of filters and convolution that can be implemented with the miniDSP?
> would that introduce a delay grater than 500ms to the audio signal of a video? (my favorite player for movies is PowerDVD and I can only set a maximum of -500ms there as negativ delay)
- 4 subs with one miniDSP, optimized with MSO biquads for input (10 for all subs), output and through crossover function (18 for each sub)
> use target curve within MSO
> after that no EQ through audyssey anymore, only time alignment with the other speakers with the other miniDSPs
> makes it sense to use the FIR filter for each sub on top to do some phase alignment?
> should I not use the input BQ on top of output BQ to prevent phase issues?
> what number and combination of filters, MSO, and miniDSP is the best approach for 4 subs?
Would it even be reasonable in regards to sound improvement to go this way in comparison to your guide for Atmos/Auro calibration with audyssey hack?
Or to put it in another way: If you had the possibility's of audyssey App, REW, MSO, UMIK-1, pre-amp and external-amp, miniDSP for each channel and sub > what would you do to get the best result for movies in Atmos/Auro?
MiniDSP 2x4HD FIR filters are uselss for subs but MSO can make use of some "allpass" filters with biquads to fix the phase response. 500ms PowerDVD delay would be enough. Your bass response was already quite nice so I wouldn't expect a major improvement but a slight improvement in bass can always be obtained with a lot of effort. External amps will certainly improve the sound comapred to class D power supplied by the AVR. The speaker crossovers and EQ will still be in the AVR though. You could use 2x8 channel MiniDSP units for bypassing even that but you would be better off with an AVR with Dirac ART if you are willing to invest that much.
@@ocaudiophile Thanks for the feedback. I wonder if ART will be at some point be integrate to existing DIRAC devices like the miniDSP 2x4 HD. But I don't know if this if even possible. And I would wait for even considering an ART AVR until it is available in the below 10K price range.
The main reason for my question from the beginning was to bypass the EQ from AVR or pre-amp completely with miniDSP to use your methods of filter creation and convolution. And as you said the bass can already be made pretty good without ART. Do you have a source where I could learn more about "allpass" filters with MSO? A benefit of miniDSP would also be the ability to precisely dial in the delay between all speakers and the subs.
So if I understand you right I could use convolution filters with miniDSP and would be able to have synchronized video and sound? So the question for me is still to choose between your method with audyssey or your method with miniDSP.
With kind regards
Frank
www.beisammen.de/index.php?thread/121932-multi-sub-optimizer-mso/
Hello! I also have MSO and miniDSP HD for 4 subwoofers. Have you found the optimal settings for MSO?
@---vn2kv avoid gain filters and EQ with MSO, just use allpass filters (you can use them in minidsp in biquad form) and time delay. Then EQ MSO optimised combibination's 'minimum phase' response in REW and add these final EQ filters to minidsp.
You mentioned that the minidsp can't invert the all pass for crossover time correction, but then you explain how your biquad calculator can make a normalized time reversed APF in biquad form that is comparable with the minidsp because of the a0 being 1.
Does that mean you can use rephrase to find the APF needed for linearization, then make and input the biquad into the mini DSP to allow for crossover phase correction without the use of FIR?
I could use time reversed biquads in my old MiniDSP 2x4 balanced but some people had problems entering them with the newer hardware/firmware. We asked MiniDSP and they gave conflicting answers. It's possible that these biquads are not possible to implement in most machines. It seems your best option is to generate FIR taps in rephase for speaker phase correction. In rePhase, use rectangular windowing with 99%/100% centering to achieve satisfactory results with 1024 taps and avoid box correction.
@@ocaudiophile great, thank you!
Unfortunately I am using a box with Dirac capabilities, so the fir is locked down.
@@picassoimpaler3243 Dirac corrects phase response to an extent but being an automated system, it's not optimal (i.e. it has no way of knowing every speaker and its crossover frequencies and filter orders). Still, symmetrical speaker setup and carefully taken multiple measurements will improve its results.
Is there any reason to update MiniDSP 2x4HD to the latest firmware version, or just stick with the older version that it shipped with?
I haven't used both but my friend is quite happy with the latest version. Safer to ask in the forum first though.
I'd stick with the old firmware, new one has bugs with time inverted biquads
Thanks, love your videos. Is this basically what Dirac does correcting the phase?
That's what it attempts to but can't say it gets it right at all times
I downloaded your excel and noticed that your a1 and a2 are different on video. A1 number is negative on video, when on a file the A1 is positive number. Am i missing something?
MiniDSP biquads are designed like that, they invert the sign of a1. I must have mentioned that in the video.
Hi sir, I been learning alot from your videos. Im ready to purchase but find minidsp a bit over my budget and I require 4 channel in and out. How about if I use thomann t.racks DSP 4x4 Mini? Thank you in advance.
Thomann has only parametric equalizer no fir filter capability. But Minidsp only has 1024 fir taps capacity so, it is not even a real advantage. You can do a lot better DSP than anything in the market for free by using your computer as source (for up to 7.1 channels - no Atmos) and free dsp engines like Equalizer APO or Camilla
How to digital delay on eqapo or camila pc to hdmi? I even tried to use avr manual distance setting but not work. It does sound good though. Im using smaart v9.
I can not seem to find many camiladsp tutorials do you have any or please make one kind sir, thankyou.
@@ocaudiophile
I use JRiver since a long time now and I used Equliaser APO for a short time in the past but it certainly has speaker delay features for adjusting distances:
sourceforge.net/p/equalizerapo/wiki/Configuration%20reference/#delay-since-version-09:~:text=%2D1.84776%200.729402-,Delay%20(since%20version%200.9),by%20480%20samples%20(10%20ms%20at%2048%20kHz)%0ADelay%3A%20480%20samples,-Copy%20(since%20version
@@ocaudiophile now I’m doing this I’m very happy. Smaart v9 capture live transfer file send over to REW then save untouched file and make new file with curves both files open in rephase linear phase for each 4500 5.1 audio channels export to eqapo.. imho way better than Dirac. I just ordered Iems measuring kit and about to do the same for moondrop starfields and AirPod pros 2. Later I will get a headphone measuring rig and try this method. Ath ad900x. Not sure if I can use live transfer don’t know if the sound card can loop back.
Nice tutorial. I have a questions. For multiple subs using minidsp, should I time align then eq the subs using minidsp amd after that running audyssey? Or just time align the subs using minidsp the run audyssey? Thank you
MiniDSP introduces delays which frequently makes it impossible to properly time align the subs with the rest of the system with the distance limits in the receiver. But if you're using it anyway, then phase align subs with each other - ua-cam.com/video/ga2eOwJRtXo/v-deo.html and EQ the combined response in DSP. You can then apply a final touch if necessary during Audyssey calibration - ua-cam.com/video/g26gbFdAIxE/v-deo.html
@@ocaudiophile thank you. What about if I have more than 2 subwoofer. How to phase align it?
MSO
@@ocaudiophile other than mso, Just wondering, is it possible after finished align 2 subwoofers as your tutorial, after that I phase align the 3rd subwoofer with the previous 2 subwoofers phase alignment result. Thank you.
Yes İ guess you could do that using vector average of each aligned pair with the next one.
Can I use Re Phase to correct phase and use that measurements to REW for EQ and upload it in Mini Dsp 2x4 HD
Yes but you'll be limited to 2042 FIR taps per channel. The video shows how.
Thanks, Can you make video on Phase correction and Alignment where we face Subwoofer or speaker delay more than the limit of Audyssey or any other calibration software?
@ManCaveAudio I'm working on a new A1 version which will solve insufficient sub delay issue.
@@ocaudiophile ohh, thanks and great. But it will be helpfull if you make small video for those who are looking to experiment manual calibration. Because when I add miniDsp 2x4 HD for my subwoofer, it adds extra delay and I want to make it correct with phase alignment, not time alignment.
Hi, to do phase correction in Subwoofer frequency to match with mains, If I apply FDW of 15 cycle (AVR Denon X4700H) and create Minimum Phase virson copy and do EQ of that FDW+MP virson to upload in MiniDsp 2x4 HD, will this work for phase alignment? Without causing additional delay?
What to do when your speaker filters are unknown? Can you eyeball or measure them?
You can see their whereabouts in the phase graph but you cannot know the exact frequencies
@@ocaudiophile so then I cannot use this tutorial. too bad
@@l.s.1709 I am afraid, that's correct.
Dear Serkan, is there any chance to get in direct contact with you? Do you have a Homepage? Are you located in Germany? Could need your help for a High End speaker Project. All the best & thnk you for all these interesting Video´s - Markus
AI is a real problem these days, it makes videos far worse. HOWEVER, at least you made some contribution.
True, unfortunately this video was shot during the early days of AI and my accent is pretty ugly, too :)
@@ocaudiophile I prefer your accent much more than AI, besides, you will just sound the same as the swamp of other AI videos out at the moment, I avoid them as I am sure many will.
uhhh we didn't learn a damn thing
?