Optimizing subwoofers and integration with mains: multi sub optimizer - Page 66 - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 691Likes
Reply
 
Thread Tools
post #1951 of 1981 Old 06-19-2020, 10:11 AM
AVS Forum Special Member
 
Join Date: Feb 2013
Posts: 1,654
Mentioned: 50 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 841
Quote:
Originally Posted by Sam Ash View Post
1. As long as a processor uses the RBJ convention it can be used with MSO and REW, right ?
That's sufficient, but not necessary. The processor can use the legacy Q definition (e.g. Behringer DCX2496) or the RBJ Q definition (e.g. Behringer iNuke). As long as you know which convention is used, and pick the Q value in MSO according to that convention, you'll be fine.

Quote:
Originally Posted by Sam Ash View Post
2. Ability to upload biquad coefficient files (a text file) into a DSP is much more versatile. Would entering these manually, into a non-supported DSP that used the RJB convention, be a tedious and time consuming process ?
I'm not aware of any commercial DSP boxes other than miniDSP that allow entering biquad coefficients in any form, be it file or manual entry. For manual data entry, the usual approach is to enter 3 parameters for a PEQ: center frequency in Hz, boost/cut at center frequency in dB, and Q.

Quote:
Originally Posted by Sam Ash View Post
Just saw post no. 29 of this thread, that's a lot of data for just one PEQ filter. After reading your response, I am under the notion that using the MiniDSP 2x4HD would be much more efficient.
That's actually an entire channel, which in the case of that particular miniDSP model happens to be 6 PEQs. For the 2x4 HD, it's 10 PEQs (= 10 biquads). The biquad file export and import does an entire channel in one file, not just one filter - another advantage of that approach.

Quote:
Originally Posted by Sam Ash View Post
I'm not sure whether the software on the NST processor would be able to import the biquad coefficient files generated by MSO.
As I mentioned, this seems to be a capability that's unique to miniDSP. I'm cheating a bit here. Analog Devices, the vendor of the DSP chips used by miniDSP, has software with this capability, but that's for people building their own DSP boards, and the files it uses aren't compatible with the miniDSP format.

Quote:
Originally Posted by Sam Ash View Post
This is an odd question Andy but would you know how to convert a 2x4HD kit into a 2x4HD balanced kit ? - A kind and knowledgeable member on the MiniDSP forums said that it is possible. I cannot find any step by step references that explain the procedure.
That question belongs in the DIY forum.
Sam Ash likes this.
andyc56 is online now  
Sponsored Links
Advertisement
 
post #1952 of 1981 Old 06-19-2020, 10:28 AM
AVS Forum Club Gold
 
giomania's Avatar
 
Join Date: Nov 2001
Location: Chicago, Illinois
Posts: 4,279
Mentioned: 44 Post(s)
Tagged: 0 Thread(s)
Quoted: 918 Post(s)
Liked: 730
Quote:
Originally Posted by Sam Ash View Post
Hi Andy, I appreciate your detailed and interesting explanations.

So let me ask you some questions:-

1. As long as a processor uses the RBJ convention it can be used with MSO and REW, right ?

2. Ability to upload biquad coefficient files (a text file) into a DSP is much more versatile. Would entering these manually, into a non-supported DSP that uses the RJB convention, be a tedious and time consuming process ?

Just saw post no. 29 of this thread, that's a lot of data for just one PEQ filter. After reading your response, I am under the notion that using the MiniDSP 2x4HD would be much more efficient.

I'm not sure whether the software on the NST processor would be able to import the biquad coefficient files generated by MSO.

This is an odd question Andy but would you know how to convert a 2x4HD kit into a 2x4HD balanced kit ? - A kind and knowledgeable member on the MiniDSP forums said that it is possible. I cannot find any step by step references that explain the procedure.
Assuming the person was not simply referencing the conversion of connections; XLR to RCA (miniDSP input) RCA to XLR (miniDSP output), I would suggest you ask that member.

If you need to know how to convert the cables, I offer this information I collected:

There are guides to wire balanced audio adapters that do not take into account which direction (XLR -> RCA -or- RCA -> XLR) the adapter is intended to be used. Using an incorrectly wired XLR -> RCA adapter can stress or damage XLR output circuitry on some equipment. It is common for the XLR -> RCA adapters to connect both pins 1 (ground) and Pin 3 (NEG) to shield, which is fine for use from RCA -> XLR. However, adapters wired like this should be avoided for use from XLR -> RCA, as it can stress or damage XLR output circuitry on some equipment. The best source for the specific pinout requirements based upon the connection type are from Rane Corporation’s Note 110 (Sound System Interconnection), located here.

If you just want to buy the cables, see below for those that comply with the Rane Note.

Premade XLR (Source) to RCA (Destination) Adapters
This adapter from Emotiva is wired correctly for XLR -> RCA use (No Transformer / Cross-Coupled Stage):
Pin 1 (GND) to RCA barrel/shield
Pin 2 (POS) to RCA center
Pin 3 (NEG) not connected

Premade RCA (Source) to XLR (Destination) Adapters

Below are some pre-made cable options and information for making your own adapters.

Monoprice 6ft Premier Series XLR Male to RCA Male Cable, 16AWG ~$6.99
This cable is wired correctly for RCA -> XLR:
Pin 1 (GND) to RCA barrel/shield
Pin 2 (POS) to RCA center
Pin 3 (NEG) to Pin 1 (GND)

Hosa Technology RCA Male to 3-Pin XLR Male Audio Cable (Metal) - 2' ~$7.45
This cable is wired correctly for RCA -> XLR:
Pin 1 (GND) to RCA barrel/shield
Pin 2 (POS) to RCA center
Pin 3 (NEG) to Pin 1 (GND)

Note that most people don't seek this information out, purchase the (technically) incorrectly wired conversion cables, and they work just fine. It all depends upon your level of tolerance for wanting to use the correct thing or not.

Mark
andyc56 likes this.

Lumagen Radiance Pro HDR Tone Mapping Guide: Click Here

Lumagen Radiance Pro Setup & Calibration Tips: Click Here

UHD-BD Backup & Playback Guidance Link
giomania is offline  
post #1953 of 1981 Old 06-20-2020, 09:57 AM
Advanced Member
 
Sam Ash's Avatar
 
Join Date: May 2008
Posts: 893
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 525 Post(s)
Liked: 73
Quote:
Originally Posted by giomania View Post
Assuming the person was not simply referencing the conversion of connections; XLR to RCA (miniDSP input) RCA to XLR (miniDSP output), I would suggest you ask that member.

If you need to know how to convert the cables, I offer this information I collected:

There are guides to wire balanced audio adapters that do not take into account which direction (XLR -> RCA -or- RCA -> XLR) the adapter is intended to be used. Using an incorrectly wired XLR -> RCA adapter can stress or damage XLR output circuitry on some equipment. It is common for the XLR -> RCA adapters to connect both pins 1 (ground) and Pin 3 (NEG) to shield, which is fine for use from RCA -> XLR. However, adapters wired like this should be avoided for use from XLR -> RCA, as it can stress or damage XLR output circuitry on some equipment. The best source for the specific pinout requirements based upon the connection type are from Rane Corporation’s Note 110 (Sound System Interconnection), located here.

If you just want to buy the cables, see below for those that comply with the Rane Note.

Premade XLR (Source) to RCA (Destination) Adapters
This adapter from Emotiva is wired correctly for XLR -> RCA use (No Transformer / Cross-Coupled Stage):
Pin 1 (GND) to RCA barrel/shield
Pin 2 (POS) to RCA center
Pin 3 (NEG) not connected

Premade RCA (Source) to XLR (Destination) Adapters

Below are some pre-made cable options and information for making your own adapters.

Monoprice 6ft Premier Series XLR Male to RCA Male Cable, 16AWG ~$6.99
This cable is wired correctly for RCA -> XLR:
Pin 1 (GND) to RCA barrel/shield
Pin 2 (POS) to RCA center
Pin 3 (NEG) to Pin 1 (GND)

Hosa Technology RCA Male to 3-Pin XLR Male Audio Cable (Metal) - 2' ~$7.45
This cable is wired correctly for RCA -> XLR:
Pin 1 (GND) to RCA barrel/shield
Pin 2 (POS) to RCA center
Pin 3 (NEG) to Pin 1 (GND)

Note that most people don't seek this information out, purchase the (technically) incorrectly wired conversion cables, and they work just fine. It all depends upon your level of tolerance for wanting to use the correct thing or not.

Mark
Thank you Mark, I really appreciate the detailed information. It shows that you've spent considerable time researching the topic.

I am more interested in technically upgrading the 2x4HD (kit version) to a proper balanced version - not just a conversion from RCA to XLR using adapters and converters. I'm of the opinion that an advanced electronics engineer on this forum may have achieved it. My room has a slight ground loop issue which causes a light hum in my sub. However, both my processor and sub have balanced interfaces and replacing the RCA cable with an XLR cable alleviated the problem.
giomania and tjambro like this.
Sam Ash is offline  
Sponsored Links
Advertisement
 
post #1954 of 1981 Old 06-20-2020, 10:16 AM
Advanced Member
 
Sam Ash's Avatar
 
Join Date: May 2008
Posts: 893
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 525 Post(s)
Liked: 73
Quote:
Originally Posted by andyc56 View Post
That's sufficient, but not necessary. The processor can use the legacy Q definition (e.g. Behringer DCX2496) or the RBJ Q definition (e.g. Behringer iNuke). As long as you know which convention is used, and pick the Q value in MSO according to that convention, you'll be fine.



I'm not aware of any commercial DSP boxes other than miniDSP that allow entering biquad coefficients in any form, be it file or manual entry. For manual data entry, the usual approach is to enter 3 parameters for a PEQ: center frequency in Hz, boost/cut at center frequency in dB, and Q.



That's actually an entire channel, which in the case of that particular miniDSP model happens to be 6 PEQs. For the 2x4 HD, it's 10 PEQs (= 10 biquads). The biquad file export and import does an entire channel in one file, not just one filter - another advantage of that approach.



As I mentioned, this seems to be a capability that's unique to miniDSP. I'm cheating a bit here. Analog Devices, the vendor of the DSP chips used by miniDSP, has software with this capability, but that's for people building their own DSP boards, and the files it uses aren't compatible with the miniDSP format.



That question belongs in the DIY forum.
Thank you for those clarifications Andy. So if one had 4 subs and was resorting to the manual method, then MSO would eventually provide a set of 4 results for the 4 sub channels where each result would be defined by 10 biquads. Hence, these would then have to be entered manually in the processor for each channel along side the timing and the gain values.

I suppose uploading a file is a much more faster. especially when the process needs to be repeated as part of the tweaking process.

I think it's time to read through the MSO tutorial

FORGOT TO MENTION - It would be nice to have MSO as a native Apple Mac app if possible and if it does not entail tons of work.

Last edited by Sam Ash; 06-20-2020 at 10:20 AM.
Sam Ash is offline  
post #1955 of 1981 Old 06-20-2020, 10:45 AM
AVS Forum Special Member
 
Join Date: Feb 2013
Posts: 1,654
Mentioned: 50 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 841
Quote:
Originally Posted by Sam Ash View Post
Thank you for those clarifications Andy. So if one had 4 subs and was resorting to the manual method, then MSO would eventually provide a set of 4 results for the 4 sub channels where each result would be defined by 10 biquads. Hence, these would then have to be entered manually in the processor for each channel along side the timing and the gain values.
Yes, that's right, although you generally won't use all 10 biquads. In MSO, you define filter channels in the software, then add PEQs to each channel. Each PEQ is implemented by one biquad.

Quote:
Originally Posted by Sam Ash View Post
I suppose uploading a file is a much more faster. especially when the process needs to be repeated as part of the tweaking process.
In the case of 4 subs, this would be 4 files, one for each channel. The gain and delay for each channel still need to be entered manually, as these aren't implemented by biquads.

Quote:
Originally Posted by Sam Ash View Post
FORGOT TO MENTION - It would be nice to have MSO as a native Apple Mac app if possible and if it does not entail tons of work.
That would require completely rewriting it, so it isn't going to happen. It took me a year and a half to originally write, and I'm retired. It was still in a fairly primitive state at the time this thread was started, and I've spent lots of time on it since then. You might be able to get it to run under WINE.
giomania likes this.
andyc56 is online now  
post #1956 of 1981 Old 06-20-2020, 02:27 PM
Advanced Member
 
Sam Ash's Avatar
 
Join Date: May 2008
Posts: 893
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 525 Post(s)
Liked: 73
Quote:
Originally Posted by andyc56 View Post
Yes, that's right, although you generally won't use all 10 biquads. In MSO, you define filter channels in the software, then add PEQs to each channel. Each PEQ is implemented by one biquad.
That makes sense, I've started reading the tutorial.


Quote:
In the case of 4 subs, this would be 4 files, one for each channel. The gain and delay for each channel still need to be entered manually, as these aren't implemented by biquads.
Understood thanks to your clear and concise explanations

Quote:
That would require completely rewriting it, so it isn't going to happen. It took me a year and a half to originally write, and I'm retired. It was still in a fairly primitive state at the time this thread was started, and I've spent lots of time on it since then. You might be able to get it to run under WINE.
You've done an amazing job developing MSO and I think we all appreciate your efforts immensely. I will definitely look into WINE.
tjambro likes this.
Sam Ash is offline  
post #1957 of 1981 Old 06-21-2020, 05:00 AM
AVS Forum Club Gold
 
giomania's Avatar
 
Join Date: Nov 2001
Location: Chicago, Illinois
Posts: 4,279
Mentioned: 44 Post(s)
Tagged: 0 Thread(s)
Quoted: 918 Post(s)
Liked: 730
Quote:
Originally Posted by Sam Ash View Post
Thank you Mark, I really appreciate the detailed information. It shows that you've spent considerable time researching the topic.

I am more interested in technically upgrading the 2x4HD (kit version) to a proper balanced version - not just a conversion from RCA to XLR using adapters and converters. I'm of the opinion that an advanced electronics engineer on this forum may have achieved it. My room has a slight ground loop issue which causes a light hum in my sub. However, both my processor and sub have balanced interfaces and replacing the RCA cable with an XLR cable alleviated the problem.
I would also be interested in a 2x4 HD balanced version without Phoenix connectors. That said, the 2x4 HD combined with MSO has been the best bang for the buck upgrade to my system EVER. I am not even sure that the $199 DIRAC live 2.0 upgrade to the 2x4 HD would provide better results than MSO?

Mark

Lumagen Radiance Pro HDR Tone Mapping Guide: Click Here

Lumagen Radiance Pro Setup & Calibration Tips: Click Here

UHD-BD Backup & Playback Guidance Link
giomania is offline  
post #1958 of 1981 Old 06-21-2020, 08:26 AM
AVS Forum Special Member
 
Join Date: Feb 2013
Posts: 1,654
Mentioned: 50 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 841
Quote:
Originally Posted by giomania View Post
I am not even sure that the $199 DIRAC live 2.0 upgrade to the 2x4 HD would provide better results than MSO?
It doesn't make sense to have Dirac Live on a piece of hardware that's only used for subs and not full-range. I think the Dirac live option for the 2x4 HD is intended for people using it as an active crossover in an active 2-way stereo speaker system (hence its default crossover configuration of a high-pass at ~2 kHz in two of its channels and a low-pass at ~2 kHz in the other two). It would be interesting to know how many 2x4 HD users actually use it that way. My guess is that it's a small percentage.

The analogous Dirac arrangement to MSO would be one of the latest AVRs with Dirac Live Bass Control (PDF). There's a lot of discussion about this in the Dirac Live thread. It offers some features that MSO doesn't, such as simultaneous "optimization of the splice" between the subs and all satellites. They're experiencing some growing pains at the moment, as can be seen from the last few pages of the Dirac Live thread. I'm hopeful that they'll get it all worked out.
giomania and pepar like this.
andyc56 is online now  
post #1959 of 1981 Old 06-21-2020, 09:07 AM
AVS Forum Club Gold
 
giomania's Avatar
 
Join Date: Nov 2001
Location: Chicago, Illinois
Posts: 4,279
Mentioned: 44 Post(s)
Tagged: 0 Thread(s)
Quoted: 918 Post(s)
Liked: 730
Quote:
Originally Posted by andyc56 View Post
It doesn't make sense to have Dirac Live on a piece of hardware that's only used for subs and not full-range. I think the Dirac live option for the 2x4 HD is intended for people using it as an active crossover in an active 2-way stereo speaker system (hence its default crossover configuration of a high-pass at ~2 kHz in two of its channels and a low-pass at ~2 kHz in the other two). It would be interesting to know how many 2x4 HD users actually use it that way. My guess is that it's a small percentage.



The analogous Dirac arrangement to MSO would be one of the latest AVRs with Dirac Live Bass Control (PDF). There's a lot of discussion about this in the Dirac Live thread. It offers some features that MSO doesn't, such as simultaneous "optimization of the splice" between the subs and all satellites. They're experiencing some growing pains at the moment, as can be seen from the last few pages of the Dirac Live thread. I'm hopeful that they'll get it all worked out.


Good point, Andy.

I’m actually thinking about getting the Emotiva RMC-1 because it’s the only device (in a reasonable price range) that can go above 16 channels, theoretically. But even if they never released the add-on module with four additional XLR connections, I would still have the miniDSP 2x4 HD managing my subs.

Mark


Sent from my iPhone using Tapatalk Pro

Lumagen Radiance Pro HDR Tone Mapping Guide: Click Here

Lumagen Radiance Pro Setup & Calibration Tips: Click Here

UHD-BD Backup & Playback Guidance Link
giomania is offline  
post #1960 of 1981 Old 06-21-2020, 09:58 AM
AVS Forum Special Member
 
Join Date: Feb 2013
Posts: 1,654
Mentioned: 50 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 841
Quote:
Originally Posted by giomania View Post
I’m actually thinking about getting the Emotiva RMC-1 because it’s the only device (in a reasonable price range) that can go above 16 channels, theoretically. But even if they never released the add-on module with four additional XLR connections, I would still have the miniDSP 2x4 HD managing my subs.
I'd be extremely cautious about Emotiva pre-pros, as they have a long history of introducing very buggy pre-pros way behind schedule, then failing to fix the bugs. I've got Emotiva power amps that have worked flawlessly for years, but the UMC-200 pre-pro I used to have was buggy, then died an early death.

Also, if you use the 2x4 HD to drive your subs in a pre-pro with Dirac Live Bass Control (DLBC), then the optional DLBC software version with multi-sub optimization can't be used. That DLBC version can only be used in conjunction with the internal channels of the pre-pro.
sdurani, giomania, shepP and 1 others like this.
andyc56 is online now  
post #1961 of 1981 Old 06-21-2020, 03:25 PM
AVS Forum Club Gold
 
giomania's Avatar
 
Join Date: Nov 2001
Location: Chicago, Illinois
Posts: 4,279
Mentioned: 44 Post(s)
Tagged: 0 Thread(s)
Quoted: 918 Post(s)
Liked: 730
Quote:
Originally Posted by andyc56 View Post
I'd be extremely cautious about Emotiva pre-pros, as they have a long history of introducing very buggy pre-pros way behind schedule, then failing to fix the bugs. I've got Emotiva power amps that have worked flawlessly for years, but the UMC-200 pre-pro I used to have was buggy, then died an early death.

Also, if you use the 2x4 HD to drive your subs in a pre-pro with Dirac Live Bass Control (DLBC), then the optional DLBC software version with multi-sub optimization can't be used. That DLBC version can only be used in conjunction with the internal channels of the pre-pro.
I have read the horror stories about Emotiva, and am leery, but I am considering it for several reasons:

1) I have a Lumagen Radiance Pro to handle the video switching, so that mitigates the HDMI issues.
2) They seem to have stabilized their firmware recently.
3) It is possible they may be able to do DLBC with four subs via the expansion module. If this never comes to fruition, I just continue along with the miniDSP managing the subs, and the processor sees just the "one" sub.
4) I am not sure it is likely that mainstream affordable processors will exceed 16 channels in the future, and I need 19 channels (9 bed, 6 heights, 4 subs).

All that said, I will check out the DIRAC live document and thread to continue my research.

Thank you.

Mark

Lumagen Radiance Pro HDR Tone Mapping Guide: Click Here

Lumagen Radiance Pro Setup & Calibration Tips: Click Here

UHD-BD Backup & Playback Guidance Link
giomania is offline  
post #1962 of 1981 Old 06-23-2020, 02:14 AM
Advanced Member
 
Sam Ash's Avatar
 
Join Date: May 2008
Posts: 893
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 525 Post(s)
Liked: 73
Quote:
Originally Posted by andyc56 View Post
It offers some features that MSO doesn't, such as simultaneous "optimization of the splice" between the subs and all satellites.
Hi Andy, can you kindly explain that feature. How is it advantageous and what benefit does it provide ?
Sam Ash is offline  
post #1963 of 1981 Old 06-23-2020, 12:03 PM
AVS Forum Special Member
 
Join Date: Feb 2013
Posts: 1,654
Mentioned: 50 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 841
Quote:
Originally Posted by Sam Ash View Post
Hi Andy, can you kindly explain that feature. How is it advantageous and what benefit does it provide ?
You've probably heard of the "sub distance tweak" in conjunction with Audyssey, which uses a flawed method in its automatic calculation of sub distance. In this tweak, you manually adjust sub distance for the smoothest response of, say Center + Subs in the crossover region as measured using REW. But unless you're very lucky, you'll find that the sub distance that optimizes Center + Subs does not necessarily optimize L + R + Subs. Once you adjust the sub distance, you've run out of things you can adjust, so you can't optimize both simultaneously. For that, you need more ways to adjust relative phase of satellites and subs in the crossover region. The appropriate level of DLBC gives more available phase adjustments in the form of all-pass filters. This allows the response in the crossover region of C + Subs, L + R + Subs, and (edit: I think) various combinations of surrounds and subs to all be optimized at once.
sdurani, hdaudio, galonzo and 2 others like this.

Last edited by andyc56; 06-23-2020 at 02:27 PM.
andyc56 is online now  
post #1964 of 1981 Old 06-23-2020, 05:38 PM
AVS Forum Club Gold
 
giomania's Avatar
 
Join Date: Nov 2001
Location: Chicago, Illinois
Posts: 4,279
Mentioned: 44 Post(s)
Tagged: 0 Thread(s)
Quoted: 918 Post(s)
Liked: 730
Yeah, I definitely need a processor with DLBC. Thanks for that, Andy. The challenge is that I need one that can handle 21 channels (9.6.6) or 19 channels (9.4.6) with two subs paired like I run them now. Something like this will probably never exist in an affordable price range. In the interim, I will continue chugging along with the miniDSP and the MSO magic.

Mark
tjambro likes this.

Lumagen Radiance Pro HDR Tone Mapping Guide: Click Here

Lumagen Radiance Pro Setup & Calibration Tips: Click Here

UHD-BD Backup & Playback Guidance Link
giomania is offline  
post #1965 of 1981 Old 06-25-2020, 12:20 AM
Advanced Member
 
Sam Ash's Avatar
 
Join Date: May 2008
Posts: 893
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 525 Post(s)
Liked: 73
Quote:
Originally Posted by andyc56 View Post
You've probably heard of the "sub distance tweak" in conjunction with Audyssey, which uses a flawed method in its automatic calculation of sub distance. In this tweak, you manually adjust sub distance for the smoothest response of, say Center + Subs in the crossover region as measured using REW. But unless you're very lucky, you'll find that the sub distance that optimizes Center + Subs does not necessarily optimize L + R + Subs. Once you adjust the sub distance, you've run out of things you can adjust, so you can't optimize both simultaneously. For that, you need more ways to adjust relative phase of satellites and subs in the crossover region. The appropriate level of DLBC gives more available phase adjustments in the form of all-pass filters. This allows the response in the crossover region of C + Subs, L + R + Subs, and (edit: I think) various combinations of surrounds and subs to all be optimized at once.
Thanks Andy, there is a dual approach to this, either C+Subs or L+R+Subs. Would it be possible to achieve L + C + R + Subs ?
Sam Ash is offline  
post #1966 of 1981 Old 06-28-2020, 08:30 AM
Senior Member
 
anothermib's Avatar
 
Join Date: Mar 2017
Posts: 239
Mentioned: 6 Post(s)
Tagged: 1 Thread(s)
Quoted: 203 Post(s)
Liked: 71
Quote:
Originally Posted by andyc56 View Post
...
But unless you're very lucky, you'll find that the sub distance that optimizes Center + Subs does not necessarily optimize L + R + Subs. Once you adjust the sub distance, you've run out of things you can adjust, so you can't optimize both simultaneously. For that, you need more ways to adjust relative phase of satellites and subs in the crossover region.
...
The answer may be obvious, but to be sure I ask anyway : This can be achieved with MSO by just adding individual measurement sets for all three channels (LCR) in the mains section, can’t it?
anothermib is offline  
post #1967 of 1981 Old 06-28-2020, 10:01 AM
AVS Forum Special Member
 
Join Date: Feb 2013
Posts: 1,654
Mentioned: 50 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 841
Quote:
Originally Posted by anothermib View Post
The answer may be obvious, but to be sure I ask anyway : This can be achieved with MSO by just adding individual measurement sets for all three channels (LCR) in the mains section, can’t it?
That's part of the answer, but not the whole answer. MSO minimizes an aggregate measure of the error. For each measurement group, the RMS error from the target in dB over frequency is computed, then all these RMS errors are in turn RMS'ed together over all the measurement groups to compute the total RMS error displayed in the Optimization Status window. Because of the individual EQ of each sub (or satellite as applicable), the act of minimizing the aggregate error also ends up minimizing the individual error over frequency of each measurement group by itself. That is, it ends up being a simultaneous minimization of the errors of each measurement group individually, and it's the added degrees of freedom made possible by the individual sub EQ that makes this possible. If you were to take away these degrees of freedom, say by using only shared EQ on the subs, you'd end up minimizing some compromise of the error over all the measurement groups, not each measurement group individually.

To state it more concisely, in order to achieve this simultaneous minimization of errors, you need two things:
1) The ability to minimize an aggregate measure of the error (e.g. the RMS error in MSO)
2) Sufficient degrees of freedom in the available adjustments such that in doing so, the individual errors that make up the aggregate error are all minimized by themselves simultaneously

Simply adding L, R and C to the mains section in MSO would achieve (1) above, but not (2). Requirement (2) is met in systems like DLBC and Harman's system by using a shared all-pass filter on the subs, and individual all-pass filters on the satellites.

It gets worse though. Suppose you had different crossover frequencies, say f1 and f2, for L+R and for C. Then with MSO you'd end up doubling the number of required sub measurements, one set for the crossover frequency f1 and the other for f2. What's really needed is the ability to measure the sub responses without any crossover at all, and a knowledge of the specific types of LPF and HPF used in the crossover so the desired data for all the different crossover frequencies can just be computed from the sub data measured without crossover. This is best done by hardware and optimization software that are made for one another. MSO tries to eke out the best possible performance with the lowest common denominator of hardware, so such esoteric solutions become impractical.
andyc56 is online now  
post #1968 of 1981 Old 06-28-2020, 12:12 PM
Senior Member
 
anothermib's Avatar
 
Join Date: Mar 2017
Posts: 239
Mentioned: 6 Post(s)
Tagged: 1 Thread(s)
Quoted: 203 Post(s)
Liked: 71
Optimizing subwoofers and integration with mains: multi sub optimizer

I am not 100% sure I follow the issue with (2). Isn’t the number of measurement groups to optimize over the same? Having delays, levels and a number of peqs per sub seems to be quite a few variables already. How many degrees of freedom would be required? Wouldn’t MSO always find an optimum for all components within the constraints? Is the issue that this optimum may be inferior to the result just considering LR?
Is the method of choice then optimizing with a combined LR measurement rather than individual channels followed by another optimization that keeps everything fixed and just looks for the best crossover frequency for the center?
I guess many may just focus on LR and leave the rest to luck, but I am wondering how much bass ends up in the center - particularly when upmixing e.g. with DSU.

Last edited by anothermib; 06-28-2020 at 12:19 PM.
anothermib is offline  
post #1969 of 1981 Old 06-28-2020, 12:51 PM
AVS Forum Special Member
 
Join Date: Feb 2013
Posts: 1,654
Mentioned: 50 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 841
Let me try to explain it this way. Suppose you wanted to optimize the integration of (L and Subs), (R and Subs) and (C and Subs) simultaneously (which is different from optimizing some compromise of the three). Let's forget about PEQ for a moment and just consider relative phase adjustment. Suppose you adjusted the sub distance to optimize the integration of (C and Subs). Once you've made that adjustment, you must then leave it as-is, because if you later change it, then the integration of (C and Subs) is by definition no longer optimized. But you want to separately optimize (L and Subs) and (R and Subs) in the crossover region as well. That means you must have the ability, in the crossover region to:

1) Adjust the relative phase of L and Subs without affecting the relative phase of (R and Subs) and (C and Subs) at all.
2) Adjust the relative phase of R and Subs without affecting the relative phase of (L and Subs) and (C and Subs) at all.

You can't use L or R distance for this, as that's already been set by impulse response measurements that match up the times of arrival. You can't use sub distance either, as you've already used that to optimize (C and Subs). You are now out of adjustments. That's where the all-pass filters come in. They provide adjustable phase shift in the crossover region, but their phase shift and group delay approach zero at high frequencies, so they don't affect the timing measurements that were used to set the L, R and C distances to begin with.
andyc56 is online now  
post #1970 of 1981 Old 06-28-2020, 01:39 PM
Senior Member
 
anothermib's Avatar
 
Join Date: Mar 2017
Posts: 239
Mentioned: 6 Post(s)
Tagged: 1 Thread(s)
Quoted: 203 Post(s)
Liked: 71
Ok, if I understand this correctly, then this is basically describing the general issue with the sub distance tweak. I am familiar with that.
The things I am less sure about are probably the following:

1. Wouldn’t MSO normally find the best compromise of integration between LCR and the subs? And isn’t that a more desirable configuration for everything that uses the c speaker than just optimized LR?

2. Isn’t there a reasonable chance that the other variables will have a favorable effect beyond just the distance tweak even if it falls short of the direct approach with a dedicated filter.

3. Is there any general drawback in using individual measurements for the channels rather than a combined LR or LCR response?

4. If I were prepared to allow another variable delay just for the center channel - sacrificing the alignment with LR, wouldn’t that help the situation?

... It seemed straight forward at first. However, I have to admit there are questions the longer I think about it ;-)
anothermib is offline  
post #1971 of 1981 Old 06-30-2020, 01:57 AM
Member
 
mcb61's Avatar
 
Join Date: Mar 2008
Location: West Oz (Au)
Posts: 48
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 25 Post(s)
Liked: 26
I have previously used MSO, with great success, to optimise my four subwoofers in a difficult open plan area.

A recent change to one of my subwoofers has resulted in the need to remeasure the subs and to re-run MSO.

When looking at the results, I noticed that the delays for subs 3 and 4 where 9.81 and 6.26 ms, where I was expecting them to be the same, as the two subs are two couch kickers which sit side by side and directly behind the the couch, with the join between them being in line with the MLP. Further, the delay for sub 2 was 10.46ms even though subs 1 and 2 are further away than the couch kickers. The actual predicted response looked fairly good.

Whilst trying to understand what was causing the significant difference between subs 3 & 4, I was reviewing the MSO help files and noted at the Tips and Tricks section the comment about near field subs. As suggested in the notes, I then used both REW impulse method and the Audyssey delay method, referred to in the guide by AustinJerry, to see what delays they both proposed.

Using the REW impulse method, the required delays where Sub 1 – 0.7ms, Sub 2 - Nil and Subs 3&4 - 9ms.

Using the Audyssey method, after converting the distance to milliseconds the timings equated to Sub 1 – 0.7ms, Sub 2 - nil, Sub 3 - 9.09ms and Sub 4 – 9.0 ms. The difference between subs 3 and 4 in distance is 3cm (1 and a bit inches), which I put down to not getting the Audyssey mic exactly in the middle of the MLP.

The difference in the distances also equates pretty closely to the physical distances, with Subs 3&4 being the closet and Sub 1 being the furthest.

Given the clear pattern with the two sets of alternative delay measurements, I then ran several MSO’s with fixed delays of 0.7ms for Sub 1 and 9.0ms for Subs 3 and 4.

The results from these fixed delays do not look as good as the earlier results, as can be seen below.

This leads to the question, should I be looking to use the better looking response with the MSO calculated delays, or the fixed delays?

I would appreciate any thoughts that Andy or other users may have.
Attached Thumbnails
Click image for larger version

Name:	Sub2 NI 92 R1.jpg
Views:	25
Size:	327.7 KB
ID:	2751054   Click image for larger version

Name:	200624 200603 New Alt Fixed Delays T90 and alt s34 gains R4.jpg
Views:	22
Size:	336.7 KB
ID:	2751056  
mcb61 is offline  
post #1972 of 1981 Old 06-30-2020, 09:42 AM
AVS Forum Special Member
 
Join Date: Feb 2013
Posts: 1,654
Mentioned: 50 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 841
Quote:
Originally Posted by mcb61 View Post
A recent change to one of my subwoofers has resulted in the need to remeasure the subs and to re-run MSO.

When looking at the results, I noticed that the delays for subs 3 and 4 where 9.81 and 6.26 ms, where I was expecting them to be the same, as the two subs are two couch kickers which sit side by side and directly behind the the couch, with the join between them being in line with the MLP. Further, the delay for sub 2 was 10.46ms even though subs 1 and 2 are further away than the couch kickers. The actual predicted response looked fairly good.

Quote:
Originally Posted by mcb61 View Post
The difference in the distances also equates pretty closely to the physical distances, with Subs 3&4 being the closet and Sub 1 being the furthest.

Given the clear pattern with the two sets of alternative delay measurements, I then ran several MSO’s with fixed delays of 0.7ms for Sub 1 and 9.0ms for Subs 3 and 4.

The results from these fixed delays do not look as good as the earlier results, as can be seen below.

This leads to the question, should I be looking to use the better looking response with the MSO calculated delays, or the fixed delays?
In general I'd say that if the difference between the two results were small, then pick the one that sounds best to you. But the difference between these two plots is pretty significant, so if it were a binary choice, I'd say to use the one that shows the better computed results. There is another thing you might try, though. Having two near-field subs equidistant from the MLP suggests that, if they are identical, their delays should be the same. With version 1.41 or later of MSO, you can force that condition. The technique involves having no delay blocks for the two near-field subs at all. The delay blocks for all other subs should be allowed to have negative delay values. After running the optimization, do a "Normalize Delays" operation. This will shuffle the delay blocks around, giving equal positive delays to the two near-field subs, and you'll end up with one more delay block than you started with.
giomania and Dick_Lai like this.
andyc56 is online now  
post #1973 of 1981 Old 06-30-2020, 02:44 PM
Member
 
mcb61's Avatar
 
Join Date: Mar 2008
Location: West Oz (Au)
Posts: 48
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 25 Post(s)
Liked: 26
Quote:
Originally Posted by andyc56 View Post
In general I'd say that if the difference between the two results were small, then pick the one that sounds best to you. But the difference between these two plots is pretty significant, so if it were a binary choice, I'd say to use the one that shows the better computed results. There is another thing you might try, though. Having two near-field subs equidistant from the MLP suggests that, if they are identical, their delays should be the same. With version 1.41 or later of MSO, you can force that condition. The technique involves having no delay blocks for the two near-field subs at all. The delay blocks for all other subs should be allowed to have negative delay values. After running the optimization, do a "Normalize Delays" operation. This will shuffle the delay blocks around, giving equal positive delays to the two near-field subs, and you'll end up with one more delay block than you started with.
Thanks Andy, I now recall reading that you had introduced this feature, but didn't come close to thinking about trying it. I will give it a run later today.
mcb61 is offline  
post #1974 of 1981 Old 07-01-2020, 02:58 PM
Member
 
mcb61's Avatar
 
Join Date: Mar 2008
Location: West Oz (Au)
Posts: 48
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 25 Post(s)
Liked: 26
Thanks once again Andy, I am now getting a better result.

When reviewing my other settings, I noted that at Optimisation Options / Optimisation Methods, I had "As flat as possible..." and not "Best match of MLP....." selected (for each of my recent optimisations). Mutter, mutter mutter.

For those that aren't aware, the "Best match..." option results in better final results, once Dirac or Audyssey etc has been run.

Last edited by mcb61; 07-01-2020 at 03:06 PM. Reason: Clarity
mcb61 is offline  
post #1975 of 1981 Old 07-01-2020, 06:45 PM
AVS Forum Special Member
 
Join Date: Nov 2016
Posts: 1,585
Mentioned: 49 Post(s)
Tagged: 0 Thread(s)
Quoted: 1299 Post(s)
Liked: 397
@andyc56 , need a bit help. I'm trying to help someone on AVS to use MSO. I asked him to take readings but running into weird issue I can't figure out the reason of. My friend has a total of 6 subs spread in 4 different locations (front left, front right, back left, back right). 2 subs on front left are stacked and also on front right. So I asked him to take the readings of front left subs as 1 sub and same for front right subs. So we end up having a total of 4 subs as far as MSO is concerned. So far so good. Just to start off, I asked him to take reading at MLP by first taking reading of all subs together and then each sub individually. We end up having following readings (P1 refers to position 1)
P1 ALL
P1 FL
P1 FR
P1 BL
P1 BR

He sent me the REW file. Now the issue is that when I create the group for Position 1, it doesn't match P1 ALL. I have setup projects like 100 of times for myself so I know how MSO works and needs to be setup. I don't have this issue with my readings but in his case, it doesn't match properly. He did take each reading with acoustic timing reference on. It seems like a blocker at this point since I can't visit him personally since he lives in a different state. Any change you can spot the issue if I provide MSO/REW files?

For Sale (2 Mini Marty with UXL Drivers)" https://www.avsforum.com/forum/209-a...arty-2-nj.html
harrisu is offline  
post #1976 of 1981 Old 07-01-2020, 07:07 PM
AVS Forum Special Member
 
Join Date: Nov 2016
Posts: 1,585
Mentioned: 49 Post(s)
Tagged: 0 Thread(s)
Quoted: 1299 Post(s)
Liked: 397
Quote:
Originally Posted by mcb61 View Post
For those that aren't aware, the "Best match..." option results in better final results, once Dirac or Audyssey etc has been run.
Can you elaborate a bit more on this? What do you mean by above statement? As far as I know, Best match for MLP means that MSO will prefer to make MLP better as compare d to other seats. I have never used "As flat as possible" but anytime I use it, MSO throws all the FR all over the place . Not sure when someone would use As flat as possible.

For Sale (2 Mini Marty with UXL Drivers)" https://www.avsforum.com/forum/209-a...arty-2-nj.html
harrisu is offline  
post #1977 of 1981 Old 07-01-2020, 07:37 PM
AVS Forum Club Gold
 
appelz's Avatar
 
Join Date: May 2004
Location: A Hilton property near you!
Posts: 1,111
Mentioned: 50 Post(s)
Tagged: 0 Thread(s)
Quoted: 855 Post(s)
Liked: 1007
Quote:
Originally Posted by andyc56 View Post
I'm not aware of any commercial DSP boxes other than miniDSP that allow entering biquad coefficients in any form, be it file or manual entry. For manual data entry, the usual approach is to enter 3 parameters for a PEQ: center frequency in Hz, boost/cut at center frequency in dB, and Q.

QSC QSYS platform is able to import both custom IIR and FIR filter coefficients.

Looks like quite a few others also have the ability to import custom FIR filters, although not all are explicitly using biquads.

https://eclipseaudio.com/fir-capable-products/

Can import files into their software using SMAART which I use, as well as REW apparently.

Adam Pelz ,Acoustic Mafia - Hear No Evil
Level III Trinnov Altitude Calibrator, JBL Master ARCOS Calibrator, CEDIA Designer, Home Acoustics Alliance Instructor LIII, THX HT1+ HT2+ Video
Mercenary Calibrator for Manufacturers, Integrators and System Owners
appelz is online now  
post #1978 of 1981 Old 07-01-2020, 07:51 PM
Member
 
mcb61's Avatar
 
Join Date: Mar 2008
Location: West Oz (Au)
Posts: 48
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 25 Post(s)
Liked: 26
Quote:
Originally Posted by harrisu View Post
Can you elaborate a bit more on this? What do you mean by above statement? As far as I know, Best match for MLP means that MSO will prefer to make MLP better as compare d to other seats. I have never used "As flat as possible" but anytime I use it, MSO throws all the FR all over the place . Not sure when someone would use As flat as possible.
AndyC probably needs to comment on this, but my understanding is that As flat as possible, would be used by someone who is not using any subsequent Audy/Dirac correction. Whereas Best Match just tries to get each seat as even as possible, when compared to the MLP, however the overall results are not necessarily very flat. The subsequent running of Audy/Dirac will then flatten the overall response.

One difference in my understanding, from yours is that the Best Match, does not try to get MLP the best, it is just used as the reference. If you want to favour the MLP, you need to go into the weighting section of Optimisaton Options.

Like you, I have used the Best Match in the past and got good results. For some reason, as noted in my earlier post, I had had been inadvertently using the As flat as possible in my earlier runs with this set of measurements and my predicted results weren't looking anywhere as good as those I had got in the past. Once I found the incorrect setting, my results started looking much more consistent.

Interestingly, the distance between the two halves of the couch kickers (subs 3 and 4) also seem to be yielding more consistent delays.

If I get a chance later today (currently mid morning here), I will see if I can dig out a couple of graphs that use the same optimisation settings, with the only change being the Optimisation Method, which bears out you observation.

As background, in case anybody is confused as to how my system is configured, I have a 5.4.4 AV system in an open plan family room. Two cylinder subs located at the front (one each side) and the two couch kickers seated directly behind the couch.

I use MSO to develop the delays, gains and PEQ's which in then feed into a MiniDSP 2x4HD, to take the single sub input and get the subs all working to their best.

Once the subs are correctly configured, I then run a Dirac calibration on my MiniDSP 88A. The 88A feeds the LCR and surrounds, plus the subs (1 channel) and two of the tops.
Attached Thumbnails
Click image for larger version

Name:	P1030104.JPG
Views:	18
Size:	159.2 KB
ID:	2751814  
mcb61 is offline  
post #1979 of 1981 Old 07-01-2020, 08:21 PM
AVS Forum Special Member
 
Join Date: Nov 2016
Posts: 1,585
Mentioned: 49 Post(s)
Tagged: 0 Thread(s)
Quoted: 1299 Post(s)
Liked: 397
Quote:
Originally Posted by mcb61 View Post

One difference in my understanding, from yours is that the Best Match, does not try to get MLP the best, it is just used as the reference. If you want to favour the MLP, you need to go into the weighting section of Optimisaton Options.
I agree with your understanding.

Quote:
I use MSO to develop the delays, gains and PEQ's which in then feed into a MiniDSP 2x4HD, to take the single sub input and get the subs all working to their best.
Same here except I'm not using PEQ. My four subs do the work good enough to make 2 seats work fine. 3rd not so much but no one sits there so no worries. If I add 1 PEQ per channel, I get response within 1-2 across seats.

For Sale (2 Mini Marty with UXL Drivers)" https://www.avsforum.com/forum/209-a...arty-2-nj.html
harrisu is offline  
post #1980 of 1981 Old 07-06-2020, 11:15 AM
AVS Forum Club Gold
 
giomania's Avatar
 
Join Date: Nov 2001
Location: Chicago, Illinois
Posts: 4,279
Mentioned: 44 Post(s)
Tagged: 0 Thread(s)
Quoted: 918 Post(s)
Liked: 730
A friend offered a miniDSP 2x4 balanced unit to use as a way to have a shared filter and still use the miniDSP 2x4 HD for the MSO biquads and easy loading of Bass EQ filters.

I can daisy-chain the miniDSP units, and would have the balanced version first, so the only question is if the shared filter results in better performance. I know the goal of the shared filters (from the MSO Tips below) is to improve MLP response, but my response is very good now.

However, I am using six subs in my setup, configured as three pairs. While the sub pairs are fairly equidistant, perhaps the balanced unit can be used to eliminate any slight differences in these distances by time align two of the sub pairs that are not exactly equidistant from the MLP? Perhaps the differences are so small that it would not result in any improvement? So I thought I would ask before I attempt the experiment, as I really don't have the time to waste if it won't make much difference. Hopefully my millisecond conversion math is correct.

The front pair are 5" (~0.4 ms) different in distance from the MLP.
The side pair are 14" (~1.5 ms) different in distance from the MLP.
The rear pair are 1" (~0.07 ms) different in distance from the MLP.

Given these measurements, I was thinking the front and side pair would be good candidates to time align each pair with delays in the miniDSP 2x4 balanced before sending the signal on to the miniDSP 2x4 HD. The balanced unit only has a maximum delay of 7 ms, and while all my subs have delays greater than 7 ms, I was thinking I would only need to delay the closer sub in each pair, so that limitation does not matter.

Edit: I realized this time alignment of pairs will not work, as I cannot go from four outputs in the miniDSP 2x4 balanced to two inputs on the miniDSP 2x4 HD. I know that using multiple miniDSP 2x4 HDs is possible, but too cumbersome to load Bass EQ filters in two devices.

I would appreciate any input.

Thanks.

Mark



Reference from MSO Tips:

Improving Response Flatness at the Main Listening Position

The corrected responses of Figure 30 (page 42) show improved frequency response flatness for all listening positions compared to the uncorrected responses of Figure 31. Yet the response at the main listening position may not be as flat as you would like it to be. Since one side effect of adjusting the individual channel EQs is to reduce the seat-to-seat variation in frequency response, altering these individual EQ settings is not a good way to flatten the response at the main listening position any further. Instead, a better idea is to keep the existing individual channel EQs and add shared filters that affect the responses of all subs simultaneously so that the seat-to-seat variation of the frequency response is not affected. Keep in mind that in doing so, the flatness of the frequency response at positions other than the main listening position may be degraded relative to Figure 30, as was mentioned earlier in the tutorial when discussing the pitfalls of shared EQ. However, should you choose to use this technique, MSO gives you the needed tools.

To perform this task, you'll clone the Optimized configuration of tutorial_4.msop. Open up tutorial_4.msop and in the Config View, select the root node of the Optimized configuration (the node labeled “Optimized” just underneath Configurations. Right-click and choose Clone Configuration from the context menu. This will cause the Clone Configuration dialog to be launched. In this dialog, enter MLP Cleanup in the Choose a name for the new configuration box. Make sure the checkbox labeled Clone associated graphs is checked. This will create an identical copy of the Optimized configuration, and replicate its graphs, with the traces of the new graph referring to the new configuration. Right-click on the MLP Cleanup node and choose Lock All Filter Parameters of this Configuration. This will prevent the optimizer from changing the individual subwoofer EQ settings of MLP Cleanup. You'll see that the icons for all the filters in this channel are dark gray in color. You'll want to add some shared sub filters and run a new optimization on MLP Cleanup so that only the parameters of these shared filters are adjusted.

Lumagen Radiance Pro HDR Tone Mapping Guide: Click Here

Lumagen Radiance Pro Setup & Calibration Tips: Click Here

UHD-BD Backup & Playback Guidance Link

Last edited by giomania; 07-06-2020 at 12:23 PM.
giomania is offline  
Sponsored Links
Advertisement
 
Reply Subwoofers, Bass, and Transducers

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off