Great video, I need to track down your REW setup video referenced in this one to ensure I have it all setup properly. I am about to redo my MSO readings because I did not use the timing feature. I plan to try to use the Java drivers, set speakers to small with the highest x-over possible. Send the timing to the left speaker, and send my sweep to the right speaker, and unplug the right speaker. As a result the sweep should not play on the timing speaker as it will specifically target the opposite speaker, and the dead speaker unplugged should not play the bass sweep but will pass over to energize the subs via the x-over.
That will work perfectly well! I was going to suggest that as an option, but didn't want to complicate an already complicated conversation on AVS. Feel free to keep that thread going and update us on your progress :-)
Thank you for the video. I have struggled and struggled but can not get any ASIO drivers to work. I read below that Java can still be used similarly to aligning PEQ my 3 subs with the minidsp. I was told to do the following: "You can get your MSO measurements using the JAVA drivers by: 1. TURN OFF THE POWER TO SUBS, MINIDSP AND AVR WHILE CHANGING THE CONNECTIONS 2. Connect the miniDSP to either the right or left speaker PRE-AMP OUTPUT AND DISCONNECT THAT SPEAKER on AVR. (the other will be used for timing reference.) 3. Set the fronts to "Large" in the AVR. 4. Follow MSO instructions from here." Can I accomplish the same goal but raising the crossover? That is certainly easier. Thank you
For a sub only optimization, the process above will work. However, you are measuring a normal channel, not the LFE channel. As a result, raising the crossover won't work, that would actually remove more of the bandwidth we want to measure. As a results you do need to set the L&R channels to "large". Just write down the current crossover value so it can be properly set when you're done.
No. When measured using the LFE channel, the subs will be 10 dB louder than the mains. This is by design and is part of the specification for the LFE channel. When you set your crossovers and sweep a non-LFE channel, the subs will be at the correct level (i.e., those redirected frequencies are not coming from the LFE channel so they do not receive the 10 dB uplift).
I'm having trouble getting a nice response in MSO. I'm thinking it might be because I have combined nearfield and far field subs depending on seat position and this is an overly complicated setup. But I'm also wondering if it is simply because my REW measurements have such varied SPL. I gain-matched my 4 subs which means each sub has different SPL at all positions, including MLP. I'm wondering if level-matching each sub (at the MLP) would produce different results in MSO. I plan to try the switch tomorrow but wanted to comment in case others might be in a similar spot.
Again, apologies on the delay. I did respond to your other question, but catching it here as well. Yes, I agree that near field subs will muck things up a bit. Adding/removing gain is adding/removing gain no matter where you do it. If the subs are identical, I prefer gain matching over level matching as it's just easier. We can then trim things out using the optimization. I also recommend the response on the other thread to limit the cut allowed per sub. Just know that with near-field, it's probably going to be difficult to get all the subs working equally hard. That's really just something to be aware of as you start to optimize. You could end up in a situation where you start to push the far field subs too hard before the near field subs have run out of steam. I supposed you might also be able to localize some of the sounds to at least the near field subs. On that last comment, I'm thinking along the lines of how PEQ for each sub will be different. You could hear more of certain notes from some subs vs others. Especially true for a big gain discrepancy between them.
Is it better to use the center channel instead of LFE? On my Yamaha AVR the LFE seems to have a LPF so there is a drop off from about 100-120hz even when in straight or direct modes.
Yes, in my opinion it's better to use C or R instead of LFE. Set the C or L/R channels to "Large" in your AVR to bypass the crossover. Then you will get a true full-range measurement of the subs with no LPF in place.
Jeff, thanks again for the tutorial, much appreciated. One thing that I would ask for clarification on is what volume setting are you using on the AVR and what SPL are you setting each sub to prior to taking the measurements? Since you are not using the test tones from the AVR to make your measurements, this means that you must be using the volume knob on the AVR to set a specific SPL volume for the sweeps.
Check out the first 5-10 mins of this video for how I set levels: ua-cam.com/video/odsMMYXW1y4/v-deo.html Basically using REW-generated noise with the REW SPL meter and the AVR volume knob to get things setup. Remember to zero out all of your existing trims/gains/delays/room EQ/etc before starting this! Regarding individual subs, this gets into the debate of gain matching vs level matching. My 4 subs are identical so it's trivial to gain match them. The gain knob on the plate amp has detents and I just set them to all to the same setting. I prefer gain matching with MSO for a couple of reasons. First, as mentioned, it's very easy for me and I'm lazy ;-). Second, MSO is going to recommend differing gain levels no matter what method I use to match my subs. I'd rather have them all starting from the same point of reference than have no idea where they really are.
@@jeffmery thanks. I do the same. A problem that I'm running into while setting up my buddy's system which only has 2 subs monolith monolith THX tens is that M So is pretty much killing everything below 60hz with the PEQs. So when I do a sub only set up with MSN and then feed the results to to the MiniDSP and then run the auto calibration the AVR... the AVR ends up setting the sub gain to + 10db because MSO pulled so much energy out of the subs.
@@wells2671 I'd be interested in seeing the raw REW measurements as well as the before/after of MSO. I assume measurements were all taken with an acoustic timing reference?
My reply is going to counter some of what you read online. It also (somewhat) is counter to what I mention in my videos. That's not to say that other folks or my own videos are wrong. It's just that I've found a *much* better way to get good sweeps while eliminating specific level discussions. The MV really doesn't matter. What matters is that the level of the sweep is 30 - 40 dB above the noise floor of the room. 30 dB is good for most things. Use closer to 40 dB when you need things like the waterfall or spectrogram that require more energy in the room (so for subs I use a higher value). "Cool, Jeff. How do I do this?" Take a sweep of any speaker. Above the chart area, select "Distortion". For the data traces at the bottom, select *only* "Fundamental" and "Noise Floor". You want the difference between these two measurements to be between 30 - 40 dB for the entire range of the sweep. If it's too narrow, turn things up a little and run a sweep to verify. You don't really need to turn things down if it's > 40 dB gap. However, I've found that sweeps don't need to be as loud as you think. Let me know if this helps. I've answered this question a lot lately so I clearly need to do a video on it.
@@jeffmery When setting the volume, noise floor + 40dB, using the LFE channel, to measure subs for time alignment and EQ, do you do this with only one channel turned on or with all subs turned on? In my case I have 4 subs and I'm using a miniDSP 2X4 HD. Thanks in advance for your videos. I have learned so much from them.
@@martinmares8998 Ideally you want to be +30 to +40 dB above the noise floor for all measurements. That way all the data used by MSO is good and not noise or something that can't be fixed.
I like them both. A lot. :-). I really think that it doesn't get any easier than DLBC. It's about as "push-button" as you can get. The resulting sound is fantastic. The downside is that it requires a processor that supports Dirac and DLBC (some from Emotiva support Dirac, but not DLBC. I'm not picking on them, but know they have this issue. There may be others). Additionally, on many AVRs/Processors, the multi-sub DLBC requires an add-on license somewhere between $350/$500 USD. That license is included on some models (StormAudio, JBL Synthesis, others), but something to be aware of if you're shopping around (HTP-1, Arcam, others). MSO also produces excellent results. However, MSO requires more work on your part. You have to create measurements, feed them into MSO, run a couple of optimizations with different parameters, then program the miniDSP and verify results (then run room EQ). That being said, it will work with *any* AVR/Processor, even one with Dirac and DLBC. Another benefit is that the sub alignment is separate from the room correction. You don't have to change or recalculate the alignment every time you change the room correction. You can even change from Audyssey to YPAO to Dirac to ARC and your subs are always aligned (this is a benefit of any external method, not specific to MSO). I used to use MSO and DLBC together. I recently moved to just pure DLBC with my 4 subs. DLBC seems to need direct control of >1 sub to get the best sub-speaker integration results among all speaker pairs. The crossover points are smoother with 4 subs under DLBC than they are with 4 under MSO presented as 1 to DLBC. I *think* this makes sense given the phase manipulation that DLBC does to make the crossovers smooth. It just has more to work with. It would be interesting to try something like 4 subs integrated as 2 pairs with MSO and presented as 2 subs to DLBC. I bet it's better than the 4 -> 1, but i'm not sure how it would compare to 4 direct.
@Joe N Tell - Just came across the short MSO video you did with Youthman. Feel free to ping me when you're ready to go down the MSO rabbit hole. Happy to help out.
The specific volume (i.e., the number displayed on the AVR/processor) doesn't matter. What matters is making sure the volume is sufficiently above the noise floor of the room. This will vary so here's how I find a good volume level. We want our sweeps to be 30 - 40 dB above the noise floor of the room. This is really easy to figure out! Take a sweep. Switch to the "Distortion" view in REW. Change the scale to SPL or dBFS Select the "Fundamental" and "Noise Floor" traces You want the fundamental and noise floor to be 30-40 dB apart over the frequency range that you care about. Where you have nulls, the noise floor typically will rise a bit. It usually rises as you get very low in frequency as well (e.g., below 20 Hz). If you're 30-40 dB above, you're good to go. If you're not, just turn the volume up a bit and take another sweep to verify.
@@robotag95 Yep. You only need to implement the high-pass filter side of the crossover. Where you implement that filter somewhat depends on how you're going to optimize your subs and if you plan to implement BEQ in your theater. It could go on the input or output side of the miniDSP. I'm happy to help figure out what I think would be best if you want to give some additional details.
@@jeffmery i have subs that have freq ranges at the low end @ 18, 32, 45hz. My question is should i put a high-pass filter at those values to not damage the speakers?
Great video, I need to track down your REW setup video referenced in this one to ensure I have it all setup properly. I am about to redo my MSO readings because I did not use the timing feature.
I plan to try to use the Java drivers, set speakers to small with the highest x-over possible. Send the timing to the left speaker, and send my sweep to the right speaker, and unplug the right speaker.
As a result the sweep should not play on the timing speaker as it will specifically target the opposite speaker, and the dead speaker unplugged should not play the bass sweep but will pass over to energize the subs via the x-over.
That will work perfectly well! I was going to suggest that as an option, but didn't want to complicate an already complicated conversation on AVS. Feel free to keep that thread going and update us on your progress :-)
Thank you for the video. I have struggled and struggled but can not get any ASIO drivers to work. I read below that Java can still be used similarly to aligning PEQ my 3 subs with the minidsp.
I was told to do the following:
"You can get your MSO measurements using the JAVA drivers by:
1. TURN OFF THE POWER TO SUBS, MINIDSP AND AVR WHILE CHANGING THE CONNECTIONS
2. Connect the miniDSP to either the right or left speaker PRE-AMP OUTPUT AND DISCONNECT THAT SPEAKER on AVR. (the other will be used for timing reference.)
3. Set the fronts to "Large" in the AVR.
4. Follow MSO instructions from here."
Can I accomplish the same goal but raising the crossover? That is certainly easier. Thank you
For a sub only optimization, the process above will work. However, you are measuring a normal channel, not the LFE channel. As a result, raising the crossover won't work, that would actually remove more of the bandwidth we want to measure.
As a results you do need to set the L&R channels to "large". Just write down the current crossover value so it can be properly set when you're done.
Thanks for a great series with tutorials. Do you level match fronts and sub before taking measurements?
No. When measured using the LFE channel, the subs will be 10 dB louder than the mains. This is by design and is part of the specification for the LFE channel. When you set your crossovers and sweep a non-LFE channel, the subs will be at the correct level (i.e., those redirected frequencies are not coming from the LFE channel so they do not receive the 10 dB uplift).
I'm having trouble getting a nice response in MSO. I'm thinking it might be because I have combined nearfield and far field subs depending on seat position and this is an overly complicated setup. But I'm also wondering if it is simply because my REW measurements have such varied SPL. I gain-matched my 4 subs which means each sub has different SPL at all positions, including MLP. I'm wondering if level-matching each sub (at the MLP) would produce different results in MSO. I plan to try the switch tomorrow but wanted to comment in case others might be in a similar spot.
Again, apologies on the delay.
I did respond to your other question, but catching it here as well. Yes, I agree that near field subs will muck things up a bit. Adding/removing gain is adding/removing gain no matter where you do it. If the subs are identical, I prefer gain matching over level matching as it's just easier. We can then trim things out using the optimization.
I also recommend the response on the other thread to limit the cut allowed per sub. Just know that with near-field, it's probably going to be difficult to get all the subs working equally hard. That's really just something to be aware of as you start to optimize. You could end up in a situation where you start to push the far field subs too hard before the near field subs have run out of steam. I supposed you might also be able to localize some of the sounds to at least the near field subs.
On that last comment, I'm thinking along the lines of how PEQ for each sub will be different. You could hear more of certain notes from some subs vs others. Especially true for a big gain discrepancy between them.
Is it better to use the center channel instead of LFE? On my Yamaha AVR the LFE seems to have a LPF so there is a drop off from about 100-120hz even when in straight or direct modes.
Yes, in my opinion it's better to use C or R instead of LFE. Set the C or L/R channels to "Large" in your AVR to bypass the crossover. Then you will get a true full-range measurement of the subs with no LPF in place.
@@jeffmery thanks for taking the time to reply 👍🏼
My old aventage receiver seems to have some sort of fixed filter
@@jeffmery Would you then have to connect the subwoofer RCA cable to the center channel pre-out and physically disconnect the center from the AVR?
@@chrislukowski1825 Correct! A bit of extra work, but worth it IMO.
Jeff, thanks again for the tutorial, much appreciated. One thing that I would ask for clarification on is what volume setting are you using on the AVR and what SPL are you setting each sub to prior to taking the measurements? Since you are not using the test tones from the AVR to make your measurements, this means that you must be using the volume knob on the AVR to set a specific SPL volume for the sweeps.
Check out the first 5-10 mins of this video for how I set levels: ua-cam.com/video/odsMMYXW1y4/v-deo.html
Basically using REW-generated noise with the REW SPL meter and the AVR volume knob to get things setup. Remember to zero out all of your existing trims/gains/delays/room EQ/etc before starting this!
Regarding individual subs, this gets into the debate of gain matching vs level matching. My 4 subs are identical so it's trivial to gain match them. The gain knob on the plate amp has detents and I just set them to all to the same setting. I prefer gain matching with MSO for a couple of reasons.
First, as mentioned, it's very easy for me and I'm lazy ;-). Second, MSO is going to recommend differing gain levels no matter what method I use to match my subs. I'd rather have them all starting from the same point of reference than have no idea where they really are.
@@jeffmery thanks. I do the same. A problem that I'm running into while setting up my buddy's system which only has 2 subs monolith monolith THX tens is that M So is pretty much killing everything below 60hz with the PEQs. So when I do a sub only set up with MSN and then feed the results to to the MiniDSP and then run the auto calibration the AVR... the AVR ends up setting the sub gain to + 10db because MSO pulled so much energy out of the subs.
@@wells2671 I'd be interested in seeing the raw REW measurements as well as the before/after of MSO. I assume measurements were all taken with an acoustic timing reference?
@@jeffmery I'll share those later when I get access to them.
Let me know where I can send or share the file.
I notice the REW input level is set to -12 dBFS, and the gains in the miniDSP are 0 dB. What level should the AVR master volume be set to?
My reply is going to counter some of what you read online. It also (somewhat) is counter to what I mention in my videos. That's not to say that other folks or my own videos are wrong. It's just that I've found a *much* better way to get good sweeps while eliminating specific level discussions.
The MV really doesn't matter. What matters is that the level of the sweep is 30 - 40 dB above the noise floor of the room. 30 dB is good for most things. Use closer to 40 dB when you need things like the waterfall or spectrogram that require more energy in the room (so for subs I use a higher value).
"Cool, Jeff. How do I do this?"
Take a sweep of any speaker. Above the chart area, select "Distortion". For the data traces at the bottom, select *only* "Fundamental" and "Noise Floor". You want the difference between these two measurements to be between 30 - 40 dB for the entire range of the sweep. If it's too narrow, turn things up a little and run a sweep to verify. You don't really need to turn things down if it's > 40 dB gap. However, I've found that sweeps don't need to be as loud as you think.
Let me know if this helps. I've answered this question a lot lately so I clearly need to do a video on it.
thank you
@@jeffmery When setting the volume, noise floor + 40dB, using the LFE channel, to measure subs for time alignment and EQ, do you do this with only one channel turned on or with all subs turned on? In my case I have 4 subs and I'm using a miniDSP 2X4 HD. Thanks in advance for your videos. I have learned so much from them.
@@martinmares8998 Ideally you want to be +30 to +40 dB above the noise floor for all measurements. That way all the data used by MSO is good and not noise or something that can't be fixed.
What do you think of MSO compared to DLBC?
I like them both. A lot. :-).
I really think that it doesn't get any easier than DLBC. It's about as "push-button" as you can get. The resulting sound is fantastic. The downside is that it requires a processor that supports Dirac and DLBC (some from Emotiva support Dirac, but not DLBC. I'm not picking on them, but know they have this issue. There may be others). Additionally, on many AVRs/Processors, the multi-sub DLBC requires an add-on license somewhere between $350/$500 USD. That license is included on some models (StormAudio, JBL Synthesis, others), but something to be aware of if you're shopping around (HTP-1, Arcam, others).
MSO also produces excellent results. However, MSO requires more work on your part. You have to create measurements, feed them into MSO, run a couple of optimizations with different parameters, then program the miniDSP and verify results (then run room EQ). That being said, it will work with *any* AVR/Processor, even one with Dirac and DLBC. Another benefit is that the sub alignment is separate from the room correction. You don't have to change or recalculate the alignment every time you change the room correction. You can even change from Audyssey to YPAO to Dirac to ARC and your subs are always aligned (this is a benefit of any external method, not specific to MSO).
I used to use MSO and DLBC together. I recently moved to just pure DLBC with my 4 subs. DLBC seems to need direct control of >1 sub to get the best sub-speaker integration results among all speaker pairs. The crossover points are smoother with 4 subs under DLBC than they are with 4 under MSO presented as 1 to DLBC. I *think* this makes sense given the phase manipulation that DLBC does to make the crossovers smooth. It just has more to work with. It would be interesting to try something like 4 subs integrated as 2 pairs with MSO and presented as 2 subs to DLBC. I bet it's better than the 4 -> 1, but i'm not sure how it would compare to 4 direct.
@Joe N Tell - Just came across the short MSO video you did with Youthman. Feel free to ping me when you're ready to go down the MSO rabbit hole. Happy to help out.
@@jeffmery I've been in here for a while now. 🤣 I have my system setup using sub+mains. The best I've ever hear my bass sound ever.
@@joentell That's awesome! I subscribed to your channel and am looking forward to any content you release on this over there.
@@jeffmery contact me. Would love to pick your brain.
Hi. Question, what volume do you set the AVR or processor at?
The specific volume (i.e., the number displayed on the AVR/processor) doesn't matter. What matters is making sure the volume is sufficiently above the noise floor of the room. This will vary so here's how I find a good volume level.
We want our sweeps to be 30 - 40 dB above the noise floor of the room. This is really easy to figure out!
Take a sweep.
Switch to the "Distortion" view in REW.
Change the scale to SPL or dBFS
Select the "Fundamental" and "Noise Floor" traces
You want the fundamental and noise floor to be 30-40 dB apart over the frequency range that you care about. Where you have nulls, the noise floor typically will rise a bit. It usually rises as you get very low in frequency as well (e.g., below 20 Hz).
If you're 30-40 dB above, you're good to go. If you're not, just turn the volume up a bit and take another sweep to verify.
@@jeffmery thank you so much!
Another question, could i put a crossover in Minidsp to protect my subs in the low end because they don't have internal protection?
@@robotag95 Yep. You only need to implement the high-pass filter side of the crossover. Where you implement that filter somewhat depends on how you're going to optimize your subs and if you plan to implement BEQ in your theater. It could go on the input or output side of the miniDSP. I'm happy to help figure out what I think would be best if you want to give some additional details.
@@jeffmery i have subs that have freq ranges at the low end @ 18, 32, 45hz. My question is should i put a high-pass filter at those values to not damage the speakers?