REW: Equalizing Subs
Вставка
- Опубліковано 19 вер 2021
- This tutorial walks through the use of the Room EQ Wizard EQ functionality, how to implement for subs, applying the PEQ settings in the miniDSP, and verifying the response via in-room measurement.
As a bonus, there is a walk through on how to quickly test different house-curves before you build them into your sub EQ (if that's your choice).
Supporting Tutorials:
REW on Mac OS: • REW on MacOS - Introdu...
Taking Measurements: • REW on MacOS - Taking ...
MSO Sub Measurements: • Measurements for MSO: ...
REW Alignment Tool: • Sub Alignment with REW...
The sub measurement video is mainly for MSO, but the process is the same for taking initial measurements to align multiple subs with one another using the alignment tool.
You have the BEST guide for REW and MiniDSP on youtube. Period!
Very concise. Helped me fine tune my understanding of applying eq. Thank you.
Thanks for the detailed videos, was very helpful in setting up my minidsp SHD
Hi Jeff - excellent video - as always. Well done. I now need to tie up the measurement and alignment videos along with this one to complete the picture. Very good work. Thank you. Suggested short cut - to get a flat target curve, one can select the speaker type to 'none'.....
Nice! I guess I had never actually looked at the other settings in that drop down. Thanks for the pointer!
awesome and really helpful video. Thank you!
Thank you!
You know you can click the little arrows to the left halfway down to get the waterfall out of the way, I like to use up the whole screen and really zoom in and out. I used to use the "limits" all the time, but now I just use the +/- buttons, sliders and mouse. even like a 1 or 2 db scale the closer i get to perfect.
Another good tip for doing EQ is importing a room curve txt file in the preferences.. I haven't experimented too much with different ones, but for home theater I use the harmon curve, and car audio the AudioFrog. After that, your imported will always come in first. Just make sure you don't select "add room curve" unless you want to tweak the target even more. Good video, I watched the whole thing even though I knew everything. lol.
I'm not a fan of integrating the room curve into the sub alignment and EQ settings. There's nothing wrong with doing it that way at all. Plenty of people do it that way (I did for a long time as well) and have some most excellent sounding systems. I have a couple of reasons for my personal preference/approach ...
First, for a new user, how do I know what room curve I want or need? It's much easier to EQ to flat, then experiment with various house curves as a low shelf filter on the input side. It's one setting I can play with, realtime as I'm listening. If I have to bake the curve into the EQ'd response to test, I have to import, EQ, and export multiple curves into and out of REW and then into the miniDSP. It's a much more tedious and error-prone process. Once you find what you like, you can absolutely bake it into then curve.
Second, I prefer doing all of my house curve an EQ work in my processor. To me, it's much easier to work with full range or partial range (i.e. below Schroeder frequency) in Dirac than it is in REW. Again, for me this is about effort. Not all processors have this capability though, so it might not work for everyone.
Third, I have a very well established "starting point" for all changes I want to make to my system. Even if I swap out processors (which I'm about to do) I know my sub response is a perfect flat response at all listening positions. I don't have to remember to switch to a flat setting from a curve of some kind in an external system. I can't tell you how many times I thought I had my flat response selected in the miniDSP, but didn't. Then Dirac or YPAO come along and cut the house curve out. There's no choice but to repeat the entire measurement and room EQ process over. That's super time consuming when the channel counts get into the double digits.
To be clear, I'm not arguing at all with your point. It's very valid! There are a lot of different ways to solve problems and good reasons to use them all depending on your particular config or needs. I'm just presenting some alternative points for folks to consider when they read this later. I definitely appreciate your comments! They're always informative and constructive so thank you very much!
@@jeffmery I guess I use the curve during eq as I'm not sure if I have the ability to add it after EQ.. (Yamaha 2070 receiver) I tried flat, and there was NO bass and the highs were a bit much. I've read articles about how and why to use a room curve, human ears etc etc and i tend to agree with them. In a car a curve is even more effective. I could understand doing everything flat and just raising the Subs, You're basically ending up w a curve when you do that. I try to eq everything close to how I like it, so any adjustments after that are minimal. But thats just me! Not telling you how you should do it. Just pointing out capabilities in REW. Half the videos I see, Guys don't know 80% of the features and capabilities of the program. And Im still learning myself.
@@kewlbug Yamaha just added the ability to do a house curve with YPAO in their newest high-end AVRs. Don't quote me on this, but I believe only the RX-A8A, RX-A6A, and RX-A4A have this capability.
I ran Yamaha AVRs for years and still have an RX-A1080 running our living room setup and back patio. I love their sound, but the bass management and integration was always severely lacking. Your experience above mirrors my own.
You are absolutely right that a house curve is a house curve whether it's built in to the EQ curve or implemented somewhere else (input side, AVR software, etc). You are also absolutely right that a house curve is a necessity due to how human hearing works.
Thank you again for pointing out the REW features. It's an amazing piece of software!
Curious - have you done the sub distance tweak with your 2070? I've found that, like Denon/Marantz, there's almost always a cancellation at the crossover point after you run YPAO.
@@jeffmery I just looked up subwoofer distance tweak and can't figure out exactly what it is besides a trick maybe for guys that don't know how to use REW. I'll have to see if there's an update for my 2070, it would be cool to get a curve option in the YPAO. But I only use YPAO as a starting point anyway.
My process is roughly this.
For EQ Run YPAO, copy PEQ from YPAO to manual PEQ. Make manual adjustments to each PEQ. (YPAO gets it close but a human w REW is better)
Using timing reference and REW Align front speakers, Align subs, And then use "Alignment Tool". Speakers measurement at the top and "sub" on the bottom. Click the screen where the crossover is and "click align slopes at cursor" it will usually give you something like 15ft or 30 ft delay on the subs.. This may be "correct" but I feel like if I can get it closer with the alignment tool, that is more healthy than a 30ft delay. So I move the slider around at the bottom to find the best result. I will also click the invert and find the best delay there as well. Always comparing actual changes with real life results. The predicted is usually very close if you are doing everything right.
I also see a lot of people get confused with actual distance vs distance delay. The receiver is using actual distances, so adding distance is actually removing delay from that driver.
I'm really thinking about getting a MiniDSP just for subs. as the Yamaha only gives me 4 bands of PEQ for subs. Even if I just get the regular Mini dsp (with 5 bands?)
@@kewlbug Ahh okay so when you use the alignment tool to align speakers and subs, you're already doing the sub distance tweak. Then you're adding the second sub in after the fact.
I completely agree with that assessment on distance vs delay. I wish the industry had just settled on delay, but it's all over the map.
I came from Yamaha and do think the miniDSP helps a lot with YPAO. As you said, there are more bands of EQ and you can get a much better response. I really think everyone should buy the 2x4 HD, and not the standard 2x4. It's not a single feature that makes it better, but the sum of all the differences (more PEQ filters, larger available delay values [not usually needed though], ability to use BEQ if that's your thing). To be fair, most of the time I don't need the full 10 bands of PEQ to get good results with either REW or MSO. But 8 seems to be the magic spot for me.
This may have been mentioned on another comment. When loading PEQ files for each output, you can link two outputs together so that you only have to load twice, not 4 times. Once for each linked pair of outputs. Just a little time saver.
Good tip! It can certainly save a little time 🙂
I'm a creature of habit and just tend to not link things very often. I can't tell you how many times I've thought I linked something (or thought they were not linked) and gotten goofy results. I just leave them unlinked for my own personal sanity. LOL!
Thanks for the videos on your channel, it explains it in a way that will help many people interested in REW and how to take advantage of their teams.
Do you plan to make a video where you can see the acoustics of your rooms when you see measurements of TR, ETC, Waterfall for example?
Written with translator.
Greetings from Spain
Thank you! I'm glad it is helpful to you. I am not planning to do videos on those topics right now. But I will put them on my list. I have a few other topics I would like to cover first.
@@jeffmery Thanks for the deference in answering, I will be attentive to your new videos.
Greetings
Neat video! thank you for this... so, is a MiniDSP is like the hardware version of making an Equalizer APO filter and applying that instead?
That's correct! In home theater land, the devices are generally black boxes. We can't load arbitrary software on those devices to create and apply filters. We need a device in the signal chain to implement that for us. "miniDSP" is technically a brand/manufacturer of external DSP hardware. They are generally on the inexpensive side of things and very popular, especially for subs though they make different products for different use-cases. QSC is a popular high-end DSP manufacturer often recommended by many professional calibrators.
There are other ways to implement DSP externally as well. It's very common for folks in the DIY area to purchase amps with built-in DSP capabilities. Some speakers like the JBL M2 and Klipsch Jubilee require external DSP to implement the necessary active crossover network for those speakers.
Really loving your videos and earned a subscribe! I do have one question though that has really been consuming me. At what volume level should my AVR be set to when doing all these measurements and calibrations? Should I set the volume to 0 dB? This seems so simple but I just can't get my head wrapped around it.
You want the sweeps to be 30-40 dB above the noise floor of your room. Check out the first few minutes in this video to see how I set this: ua-cam.com/video/odsMMYXW1y4/v-deo.html
One small difference is that I say "40 dB above the noise floor" in the video. In reality, 30-40 dB above is just fine.
We can't just use 0 dB on our AVR/Processor. From a technical perspective, 0 dB is just a specific voltage input to the amplifier. What we hear at the microphone or listening position depends on a lot of things such as the gain structure of the amp(s), the sensitivity of the speakers, and the distance from the speaker. Setting the volume with the signal generator simplifies the process so we can ignore all the super technical stuff.
Wow finally found a vid that's very clear/understand. With all this measurements you are doing is that with audessy on/dynamic eq on?
Iv been playing with rew now near 1 year and really understanding program and this just cleared thing more.
I also just worked out low shelf input peq, makes everything so much easier .
Before I done low shelf adjustments, I got all my subs gained level, little delays just a good clean up before I done low shelf and got very good results :). I also have 4 diff presents.
If you are going to equalize subs your room correction should definitely be off for the raw measurements. I would turn off both Audyssey and any other processing like DEQ. Take care of equalizing the subs. Then recalibrate Audyssey. Be sure to do the house curve *last* as Audyssey will equalize it back down to flat if you implement the house curve first.
Another good alternative to a traditional house curve based on a low-shelf is just trying a normal PEQ @ 25 Hz, Q=1, Gain=Whatever (try starting with 3 dB). The benefit is that you're not asking things to be boosted way down low where the subs can't play. You may or may not like it, but I think it's worth experimenting.
@@jeffmery hi Jeff :). When doing sweeps/measurements, what is your sub trim level? Eg 0db, -5db and master volume?
@@almiragustin2262 In my processor they're 0 dB. On the subs, they're at the mid point of their range. You want the sweeps to be 30-40 dB above the noise floor of your room.
Use the tone generator and SPL meter in REW. Play "speaker cal" pink random noise through your L speaker. Turn up the volume until the SPL meter gets to 75 dB.
Stop the noise generator. Change the noise generator to "sub cal" ping random noise. Play that on your LFE channel. It should read 85 dB on the SPL meter. If it's too high, turn the gain knob on all subs down by the same amount. If it's too low, turn them all up by the same amount. Go slowly so you don't overshoot by too much.
Now take a sweep of something. Look at the sweep in the "Distortion" view of REW. Change the units (top left of the graph) to either dBFS or SPL. On the traces at the bottom select only "Fundamental" and "Noise Floor". The difference between the two should be 30-40 dB. 30 dB is fine for EQ, sub distance tweak, most full range stuff. If you care, or need, to look at the waterfall or spectrogram charts, it might be better to target closer to a 40 dB difference.
Once you know what main volume you need for 30-40 dB - THEN write that down so you can return to it easily :-).
@@jeffmery hi Jeff, thank you again. Il give this another shot.. iv seen you other channel with alignment tool can't wait to try that.
Ok. Here's my understanding.
1st. Turn OFF audessesy/ dynamic eq.
2nd. Have trim level on sub at 0db.
3rd. Turn up L speaker on generator Pink noise til I get 75db.
4th. Don NOT touch master volume at this stage as we now know its set to 75db.
5th. Il gain match both my subs so they both together produce 85db.
6th. Full raw measurements on rew everything FLAT.
7th. Run L sub
8th. Run R sub
9th. Go to alignment tool and get best positive results.(delays/distance)
10th. Go to eq on rew
11th Run best FLAT curve.
12th. Once chosen best results load file on output stage.
13th. Run rew, we should get same results as we have loaded on output stage. Once happy.
14th. Run audessey as normal.
DONE?
LOL is this correct?
@@almiragustin2262 Yep! That's it! One addition ... after you Run audyssey in step 14 ...
Bass will probably sound pretty thin. You will need to implement a house curve and then probably turn the bass up a little. I would do those things *after* turning DEQ on.
Good luck!
Very nice Videos Jeff 👍
Do you make a Video to align an EQ the other Speakers as example FL FR also?
And a Dirac Video would also be very nice
I have not, and don't plan, to make a video on doing full-range EQ. Most people, even the very high-end Trinnov/Storm folks, are going to use an automated room correction system with manual tweaks and tuning for full-range correction.
I thought about doing a full Dirac Live video, but there are quite a few out there already. I'm personally partial to this recent one from StormAudio: ua-cam.com/video/ZwCmMY9j6AA/v-deo.html. There some StormAudio specific stuff at the beginning, but if you jump to about the 2:10 point, it's more generic. The one thing I do NOT like a about that video, is that they load different target curves to different groups of speakers. I believe they are doing this because of the complexity in the setup the are using. For most folks, it's best to load the same target curve to all speaker groups (symmetrical pairs).
That being said, I will probably do a video at some point on working with target curves in Dirac. That seems to be a big point of confusion and contention with folks. What's the difference between a target curve and a crossover? How are/aren't they related? Should I only do partial or full range correction? Why? Those are just some of the things I've seen asked and argued about :-)
@@jeffmery perfect thx 😉
Hi Jeff, Excellent video and presentation of the many features of REW. I have used REW before but still learned a lot of new details and features that I was not familiar with watching your video. I have subscribed and will be watching all of your content. I also have Dirac Live and was hoping you may go over your technique of creating your house curve in that application as well. Thank you.
Follow up question: Would you REW the mains in addition to the subs then use Dirac Live to equalize the entire system? Or just run Dirac Live for the mains and apply a house curve at the Dirac Live level? It would seem you are applying EQ twice to the mains which I am not sure is desired. Also, I believe Dirac live should be turned off during the entire REW measurements right? Thanks
I’ve wanted to go over a video on house curves in general. Work has been crazy for a few months so I haven’t had any time to make new tutorials. I’m hoping to get some downtime in the next few weeks though.
I would integrate and EQ the subs in REW, then present them as a single “virtual” sub to Dirac, then finally run Dirac to correct the entire frequency range (or the range you care about, such as up to the transition frequency). I also recommend doing the house curve somewhere outside of the EQ process (in Dirac for your case). If we do our integration and EQ to a flat response within the miniDSP, we never have to touch our subs again. We know they are always well integrated and always flat.
Implementing the house curve somewhere else gives us more flexibility (again, in my opinion). We don’t have to fiddle with REW or risk messing up our sub alignment just because we want to try a different house curve. With Dirac, I’d do all the curves there so you don’t have to worry about managing curves in more than one location.
There are other options, of course. Hopefully I’ll get a chance to do that house curves video soon.
Tons of good info here. Looking forward to my MiniDSP which should be here Friday. I only have one Listening Position. Since I have one ported sub and one sealed sub, would I be better off using MSO to generate the EQ Filters or does it matter? My ported sub is a PA460 in a 7CU box that I wanted to use mostly for Midbass that I wasn't getting out of my UM-18-22.
With a single listening position, I think MSO is overkill. I would just use REW with the miniDSP to get everything aligned and equalized. Things will go a lot faster for you.
The important thing with the ported PA460 is making sure you properly implement a high-pass filter to protect the driver below the port tuning of the enclosure. The HPF can be implemented in a lot of different places depending on your equipment. Many of the popular amps for DIY subs include DSP. If that's an option, I would probably implement the HPF there, and then do all of the alignment and EQ in the miniDSP. My rationale is simply to prevent accidental removal of the HPF while messing with EQ and house curves and all that. I'd hate to accidentally damage the driver. Having the HPF on an external device that you're NOT messing with would be a little safer. Also I am sometimes an idiot and have to make sure and save me from myself ;-).
Of course, implementing the HPF on the miniDSP for that sub is 100% perfectly acceptable! Just make sure it's there each time you make changes. I would also implement the HPF before taking your initial measurements. No point in measuring something that your sub won't be able to produce once the HPF is in place.
One final consideration is the number of PEQ slots that will be available on the miniDSP after implementing the HPF. The best place to implement it will be on the output side connected to the amp for that sub (since the HPF is unique to that device and not shared by several identical subs). The HPF will consume one of the available PEQ slots on that channel. So with the standard miniDSP 2x4 you'd only have 5 PEQs available for EQ; the 2x4HD would have 9 PEQs available. Again .. just something to keep in mind as you start down the EQ route.
@@jeffmery Thanks for getting back to me. I am using Behringer 3000DSP for the PA460. Using the old INUKE3000DSP for the Ultimax 18. i'll leave that HP Filter in place. Is there a place I can share some of my settings with you like a Forum?
@@Quaddragon I'm active over on avsforum.com (fattire is my username). You can also e-mail to jeff at bisonhta dot com if that's easier. Happy to take a look.
Jeff: Really well done. Thank you for the insights. One thing I dont understand it when the same filter is applied to all 4 subs in MiniDSP. Shouldnt there be a unique filter for each sub?
When using REW to EQ subs, we are EQ-ing all 4 subs as one sub. We time align them first for the most positive summation, then EQ that combined response. The filters can be put on the input side of the miniDSP. In this case they'd only be entered once. They can also be put on the output side of the miniDSP. However, they need to be duplicated onto each sub to be effective.
It is possible to EQ subs individually. It's very difficult to do manually, but there are tools like MSO that do this to even out the response.
Fantastic video again Jeff. Quick one, instead of loading the rew file for each sub, any reason you did not use the "link" feature on the top of the output section? (Ray from the forum...:))
Hey man! The only reason was to make what was being done very obvious. Linking the outputs does save a bit of time for sure though.
@@jeffmery Ah got it. Makes sense. I really enjoy your videos though. Very well done.
Thanks for the help.
Have one question what’s the volume on the reciever when doing measurements?
Some say it should be set to where your front speaker hit 75 dB in a random pink noise test in REW (which is -17,5dB for me)
My subs are turned up to 50% in gain on the knob. But I can’t Seem to figure out the volume of the reciever when running measurements for each sub
The exact number on the AVR really doesn't matter. It's going to vary from person to person depending on the speakers in use. The pink noise method with one of the front main speakers is the best/easiest method. With the mains at 75 dB, the subs (when played together) should measure 85 dB through the LFE channel. If the subs are below 85 dB, turn the gain knobs up by the same amount on each sub. If they're too hot, do the opposite. Remember to make these adjustments with your room correction off and the sub gain in the AVR set to 0.0 dB.
Once you get the subs EQ'd, and you've verified that with a measurement. You're going to need to check levels again. The exact process will depend on the room correction software you're using; they all have slightly different processes and recommended sub level settings. I personally use the miniDSP gain settings if I need to *reduce* the subs output. I will use the gain knobs on the subs to *increase* the output.
@@jeffmery Thank you for the answer!
Hello I got a question. After you EQ something in REW, if you take those filters in Equalizer APO, probably you will see that overall db go over the 0 (if you boosted something in EQ) and you have to apply a preamp filter and decrease gain to match 0 again, or you risk to blow your speakers. The question is, if you export the EQ filters directly in minidsp or an other dsp, how can you know if you need to fix this db issue?
Thanks
The TL;DR answer is "you don't know if you need to fix it". That's the one of the reasons I recommend cutting and limited use of boost. While not in this video, one of the others talks about the initial setup of subs for measurements. In there I specifically talk about setting the gain controls on those subs to the midway point. The main reason for that is to give us room to do only cuts, and then turn the sub gain up to compensate for the loss in output.
It takes time to understand what is causing the dip. Then based on the cause, applying the right solution. Room modes, SBIR, comb filtering, phase interference, etc are all things that cause dips, but the solutions for each are different. There's a fine line to walk and requires good knowledge of your own system and its capabilities.
Thanks so much for these tutorials. I am confused about changes made to the MiniDSP via REW and the changes made by DIRAC 3 Live. Is there an order to make the changes? Is there conflict with the changes made? It seems both affect equalization and timing? Thanks for any feedback.
Glad they're helpful! I'm assuming you're using Dirac via your AVR/Processor and not Direct built-in to miniDSP product ...
The correct procedure is to optimize and EQ the subs first, then run Dirac. There are no conflicts and work very well together. When we take care of the subs first, we're giving Dirac less work to do. Dirac will see a really well performing sub instead of a hot mess.
Dirac will EQ the entire frequency range, not just the subs. It will also correct the timing (they specifically correct the impulse response) of all speakers. If you have the DLBC license, it can help integrate the subs with the speakers as well. Just a reminder that not all Dirac devices are capable (or allowed in some cases) of running DLBC so that may or may not even be an option.
Let me know if this clears things up for you. If not, keep asking questions!
@@jeffmery Thanks! I have the MiniDSP Studio with Dirac 3. I am using it with a two Chanel system. I have two main speakers, two bass panels and two sub woofers. I ordered the DDRC 88D to get the base management add on and so I could have the extra output for the two bass panels. It had to be sent back to repair a bug so I don’t have it in the system now. Hopefully all will work out…….
@@allenanderson1680 Ahh gotcha so studio -> 88D -> amps. That sounds pretty awesome and gives a TON of control. I would still treat the subs as noted above. Feed them a mono signal and get them aligned and working well together. I would do this in the 88D and present them to the studio (and Dirac) as a single mono sub.
IIRC, Dirac in the studio doesn't do anything special to the low end. It's going to EQ and time align the two channels which is great, but it isn't going to do some of the tricks that a DLBC capable system can perform. You'll get all the bass management and speaker integration figured out in the 88D and then let Dirac do the final touches.
Again, this sounds like a killer setup so please report back your results!
Great guide, it's got some very useful info!
Ps. Have you ever tried adding up to 8-9 dB on 80 Hz and below? Would it be audible in a 'negative way'?
I have. I personally don't care for that much low-end boost, but I know plenty of folks that add +10 dB. It really comes down to your personal preference. It's also important to consider where that increase starts. Some curves start the increase too high up in the frequency range. This sounds like really boomy male voices. That doesn't hurt anything, but just something to listen for as you start to build your own house curve.
One more quick reply ... one "negative" thing you may hear with a big boost like that is the room itself. Depending on the frequencies, I get all kinds of rattles in my room with too much bass. I've traced almost all of them down and fixed them, but some can't be fixed (like the lightbulbs in the ceiling fan).
Most purpose-built rooms don't have as much of an issue with this. But in rooms like mine (not purpose built or isolated, built on the second floor with a suspended wood floor) this can be a real issue.
The upside is I don't need to add bass shakers or Crowson's to my seating for tactile response ;-).
@@jeffmery thanks for answering. I need a significant boost due to a open baffle configuration and front and back wave cancellation. It's good to hear that there's no distortions or other sonic artifacts.
@@jeffmery as for the rattles around the room, it is often a nuisance, but I kind of even like to hear that while watching movies on stereo 😁
Hi Jeff, really great tutorial. I do have a question about matching your other speakers with the subs. My EQ went pretty well, got a pretty flat real inroom reponse (target 90db, few drops of 4db) but overall much better vs my other results. But.... When adding my towers to the mix (xover at 80hz) and run a sweep between 15-20khz, I notice huge dips in the low freq response. For example, when sweeping only the subs I get about +91DB at 68hz but when sweeping left tower + subs I get 68db at 68hz. That is a 25db drop. The exact same dips and big drop happen with the right tower. Do I need to watch another video on integrating the subs with the main speakers?
I'd put money on this being normal crossover cancellation. The best way to fix this is the "sub distance tweak". I haven't done a video on this, but there are plenty on that topic.
Be aware that this cancellation "at the crossover" may not happen exactly at the crossover frequency. It could be 10-20 Hz above or below. The fix is the same, though.
@@jeffmery I thought I was done after watching several sub video's haha. I'll have a look, thanks!
@@jeffmery Hi Jeff, the sub distance tweak did work :) I do have another question, I've been trying to let REW EQ to my target just as you showed in this video but my in room response after applying the EQ is even worse vs no EQ. My in room response is not even close to my target in REW creating huge bumps and dips while REW almost perfectly matches the target with just a few filters. I tried many settings but this happens every single time. In all video's I watch the in room response is almost the same as the REW predicted result but in my case it's always off target for miles. I don't understand why this happens. Do you have any idea? P.s. my starting curve is the raw sub response + allignment of the sub with 1/6 smoothing(got two subs).
@@johandeen6096 The in-room response should indeed almost exactly match the REW prediction. The most common mistake is not using an acoustic timing reference when taking the measurements. Since the sub distance tweak worked for you, I assume that is not an issue, but want to be sure.
Another reason could be that you have the incorrect DSP selected in REW. That affects the formatting of the EQ going into the miniDSP. REW supports both the standard 2x4 and the 2x4HD natively. Their formats and requirements are different so ensure you have the correct one selected.
Question; I have a UMIK-1, REW and MiniDSP 2x4 HD that I've been using for years for a 2-channel system. Problem is, my computer doesn't seem to support multichannel output now that I have switched to Atmos. So I'm considering buying a new computer, so that I can measure each speaker seperately. Does it have to be a Mac? Will any Mac with an HDMI-out work?
It does not have to be a Mac. REW supports Windows, Mac, and Linux operating systems. However, I only have Mac computers so I have no direct experience with the other platforms. REW cannot measure more than a 7.1 configuration without some ... creative measurement tricks.
Essentially, you would need to measure each speaker with it connected to a speaker that REW can sweep. As an example, I could hook one of my Atmos speakers into the "R" channel. Then when I sweep "R", it will actually be the Atmos speaker playing. REW doesn't know or care. Just be sure to label the sweep correctly within REW.
Another solid video. Question though, if I don't have a mini-dsp - it looks like I could export a file for my Storm Audio AVR and use that instead to eq the subs to flat. Then Re-run Dirac live for the full room correction? Cheers.
I don't have a Storm, but you absolutely should be able to implement in there. On my Trinnov, it's done using the active crossovers feature. I think this is called "Multiway Speaker" or something similar on Storm. Just add the 4 subs and then apply the PEQ filters, gains, and delays there.
I think Storm even allows for importing filters directly from REW? I wish Trinnov had that! Manual entry is a pain in the butt.
@@jeffmery Thanks Jeff, getting my mic tomorrow and gonna start testing away! Starting with my 2 subs placement. Then I'll move to see if I can flatten out the sub response and add it in somehow. Right now I wasn't happy with my 2 klipsch r115 subs. I recently moved the one in the rear literally right behind my MLP about 20 inches behind me and I love "feeling" the response. But it sucks I can't have the seat to the left and right of me have the same effect. Do I need 2 more subs? Push it back a little further? What do you think this does to the total response in the rest of the room having one sub up front and one right behind the MLP?
@@spencerbigum1309 There are definitely benefits to near-field sub placement. First, less power is needed since the subs are closer. There isn't as much distance attenuation. Second, the placement tends to greatly reduce the influence of the room on the response (i.e., it should be more. naturally flat). The downside is that the bass can sometimes be more easily localized as coming from the sub. There is a dependency on the crossover frequency here of course.
As far as total room response goes that really depends on the room, the subs, and their placement. You can get excellent response from 2 subs at the front of the room (1/4 & 3/4 points of the front wall). That's especially true if you have the flexibility to move the seating positions forwards/backwards in relation to the nulls.
Near field has benefits as well as you've found out. When you do your sub crawl to find the best positions for the subs, try placing them near field, behind the seating, with each sub "between" each seat. If you have three listening positions, this would be about on the arm rests for the center position. If you have 4 seats try them right between the left pair and right pair of seating positions. There aren't any hard/fast rules so you'll have to see what works.
There's a guy over on AVSForum that runs 8 x sealed 18" subs. 5 of them are on the front wall, but 3 of them are near field immediately behind his seats. All the folks that have heard it have been extremely impressed.
@@jeffmery I can't thank you enough for your response. I think I may also know of the guy you're talking about. I saw a video on the Kansas City Theater group and they featured a theater that sounds exactly like the one you described. The downside is that I have 2 rows of seats so I can't just plop some subs behind the first MLP row lol. Maybe having 4 subs, one off-center in front, 2 on the sides, and one in the rear? Kinda like a cross formation. I'll have to test to be sure I guess. If I have 2 subs now, and buy 2 additional subs of a different brand - can't I just level match and use REW/Dirac Live to make them sound good? I've read that you are supposed to have matching subs. Again thanks for any help - I'm having a lot of fun exploring all this stuff.
@@spencerbigum1309 Yep, that's the guy. Really good about helping people and giving advice/input too.
2 rows absolutely makes things more complicated. I don't think most folks realize just how much more complicated it gets.
For 2 rows, it's still possible to get good bass from 2 subs at the front positions as indicated above. However, the ability to precisely place each row in the room is critical. The goal would be to place them in relation to each other and the nulls/peaks so that the rows have a similar response. Then EQ will affect each row in the same manner.
If you are unable to move the seating around enough, then the answer will be 4 subs in the room. There are a lot of placements to consider, but I would start with the 1/4 and 3/4 positions of the front and rear walls. This will cancel the first 3 width modes as well as the odd order length modes.
It is possible to mix different makes/models of subs. Just try to match what you purchase with what you own. Don't mix sealed and ported. Try to find subs that have similar port tuning and output capabilities as what you own. In a perfect world, all 4 subs would be identical. If you have the means, consider selling your current 2 subs and building/buying 4 identical subs. That's definitely not a hard requirement, but man it makes things a lot easier.
As far as brands go, I'm partial to Rythmik and PSA. They both have very complete offerings at a wide range of budgets. I also feel that for a given budget, Rythmk/PSA will be much higher quality with more output than just about anything else. DIY is the king of value, but I totally understand it's not for everyone.
Hey Jeff great videos, thnx for doing them! After a few weeks of experimenting with MSO on my 6 subs, I'm curious of your opinion. Do you prefer MSO over manual REW w/EQ? after MSO are you adding a house curve? I do like the results of MSO, but Im consistently going back into ARC after to boost 20hz and below with my sealed 18s. Not that its really hard to do, and I get all rooms are different, just wondering what tweaks you may do after MSO. Thanks again!
For a single listening position, I prefer REW. For multiple listening positions, I prefer MSO. I do add a house curve after MSO. I've done it a few different ways, but have settled on the exact same methodology as you. I use my room EQ software which is Dirac in my case.
My personal workflow is optimize to a flat target with MSO. Then run my room correction software. Add the house curve in the room correction software if it has that capability (I do this with Dirac). Add the house curve using a low shelf filter on the input side of the miniDSP if the room correction software doesn't allow custom curves (I did this with YPAO).
I'm not experienced with ARC, but Dirac allows for multiple calibrations to be stored on the processor. It's then easy to switch between them. I've got three loaded and the only differences are the house curve. One is pretty flat, one is +3 dB and one is +6 dB. I can switch with the press of a button on the remote. That's really all I do and it isn't nearly as involved as it seems.
My personal rationale is that once my subs are aligned to flat with MSO in the miniDSP, I never have to touch them again. I can tinker in the room EQ software all I want. I can even swap to completely different AVR/Processor hardware and room EQ technolgoy without impacting my subs' alignment and integration with one another.
Does this help you out?
@@jeffmery Absolutely helps.. makes perfect sense. I'm still getting the feel for the AVM70 after my 60, but so far so good. Thanks for the quick reply!
@@534jgm Awesome. I came *so* close to getting an AVM70 myself; it's a great piece of gear. I'm sure you'll get it dialed in no time!
@@jeffmery First-awesome tutorials! Subscribed! I currently use the multi sub optimizer license for Dirac Live Bass Control. I’m planning a 2nd home theater room now and your tutorials have been helpful. I may use MSO in conjunction with the same or different room optimizer platform.
Question regarding your personal workflow. You mentioned you use MSO with MiniDSD 2x4 HD to optimize your 4 sub array, but that you also use Dirac with Dirac Live Bass Control. Isn’t this redundant, since DLBC can jointly optimize the subs and also the splice between main and subs?
Or do you use MSO/miniDSP to optimize the sub array, then have DLBC just “optimize the splice” with the single sub license of DLBC? I’ve been considering going this route to overcome some DLBC limitations I ran into. Thanks!
@@armandosinger Good catch, Armando ;-). Yes, I use MSO via miniDSP 2x4 HD and then use DLBC to take care of the splice with the non-sub speakers. It is indeed a bit redundant. However, I find that there are two advantages (to me anyway). First, MSO gives a *slightly* better response on my two outer listening positions than DLBC alone. DLBC does an excellent job integrating the subs with the non-sub speakers.
Second, DLBC is really not good with multi-row theaters. I help quite a few people out and DLBC just doesn't get the job done. MSO is much, better in this situation in my personal experience. It's also more flexible, since not everyone has a DLBC capable AVR/processor.
Do you keep your subs that "hot" through the whole range, or do you then re-run room eq to drop the SPL down to ~75-78 dB near the crossover frequency? I'm not sure what the best way to go about it to prevent the bass from being too bloated from 70-90 hz.
I'm sweeping using the LFE channel. The spec for that channel is to add 10 dB over the non LFE channels. So if we calibrate the mains to 75 dB, the LFE will measure 85 dB. This is perfectly appropriate for calibrating multiple subs to play nice with one another.
What you're describing is bass management of non-LFE channels. When we set a speaker to "small" (hate that term) and apply a crossover, we are redirecting some sound to the subwoofer. That redirected sound is added to the sub's output. However, that redirected sound does NOT get the 10 dB bump like the LFE channel. Figure 4 at the bottom of this page is a good example: www.soundandvision.com/content/bass-management-and-lfe-channel. Notice that the LFE get's +10 dB, but the incoming signals from the other channels do not.
We can easily verify this. Instead of sweeping the LFE channel, sweep, say, the R channel (assuming the R channel is crossed over). The output will be consistent both above and below the crossover frequency.
So - calibrate the subs with one another using LFE and then your room correction software should take care of the rest (mostly, but that's a whole 'nother ball of wax).
@@jeffmery as far as room correction goes, I'm in an even more challenging situation since my old Denon AVR-4311 doesn't support the new audyssey app so it applies some correction to any room curve I set up. I'd really like to upgrade to something with Dirac or ARC Genesis, but there is no inventory anywhere.
@@cafiveohsrockbandremix9254 I understand. I'm not an Audyssey expert by any means. However, there are some tricks I've learned from others when it comes to subs. Everything below assumes you're going to use a miniDSP for all the sub integration ...
1. Align your subs, but do NOT EQ them. Just figure out the proper delay to get them working together and not fighting one another.
2. Run Audyssey
3. Measure the LFE channel AFTER running Audyssey
4. EQ the result to what you want
When doing the EQ in the process above, it's critical that you apply the SAME PEQ filters to BOTH subs. You're not going to EQ each sub independently like you would with an MSO-style setup. I prefer to apply this EQ on the outputs of the miniDSP. This will mean applying the PEQ filters more than once (once per sub), but you're just applying the same PEQs to each one.
I always recommend EQ-ing to flat, then applying a house curve separately. The easiest way to do this is to use a low-shelf filter on the INPUT side of the miniDSP. Start with a center frequency of 47 Hz, a Q of 0.707, and whatever gain you want (3 - 6 dB is a good starting point).
Let me know if this makes sense or if you have more questions!
@@jeffmery thanks for the tips! This morning I did it reverse in that I aligned, eq'd to flat at around 78dB, and then ran Audyssey. Surprisingly Audyssey barely adjusted my subs at all. I'm going to try your suggested method to see how flat Audyssey will actually get my subs prior to me applying any EQ in the miniDSP. Maybe I'll make a video to show others what Audyssey does.
@@cafiveohsrockbandremix9254 The fact that it didn't change anything is a GOOD thing! Remember that Audyssey, and most other RC software, is optimizing over an area of measurements. There will always be variations when trying to measure just a single position.
More people making videos and showing different approaches and results is always a good thing!
Jeff... great help - my problem is that the resulting test sweep after EQ'ing doesn't come close to the target or predicted lines. those are beautifully flat - but once I apply those filters to my DSP, the next test has a huge 15db null at 62.5K - not sure what I'm doing wrong. thanks for any guidance.
There are a lot of reasons for the results to not match the prediction. I'd say the most common is not using a timing reference when taking the measurements. Use an acoustic timing reference with USB microphones. Another one is exporting the measurements with smoothing. There should be no smoothing on the exports for MSO.
Here's a great post from Andy on what to look for when the results don't match. I refer to it often for folks: www.avsforum.com/threads/optimizing-subwoofers-and-integration-with-mains-multi-sub-optimizer.2103074/post-61216139
@@jeffmery thank you, goodness, I was making several errors found in this list!
@@peterwelpton It's easy to do. Glad you found them!
If you choose to do the house curve (miniDSP LS centered at 47 Hz, not the REW house curve), couldn't you add them after your 6 "advanced" PEQ's on the output side instead of doing it on the input of the miniDSP? You would have to add the LS filter to all 4 sub outs, but then your inputs would be free to use BEQ (bass EQ) since they go on the input of the miniDSP.
Yep. That is absolutely an option. I like starting out on the input side just because it allows for quick iteration. One small change affects all subs so it’s easy to try different settings very quickly.
Should I do this before or after ARC? I got a great flat response and matched right on par with the predicted and then ARC messed it all up. I even changed arc min. correction to just 80hz, to keep what I did with the minidsp, and it still shows different, not what I was measuring pre, room correction.
Before ARC. What you typically see in room correction (doesn't matter which one), is the result of multiple microphone positions averaged together in some way. That's really difficult to compare with a single microphone position within REW. You'd need to measure the same, multiple positions, and then average the results together to approximate what the room correction software is seeing.
Unfortunately, ARC is about the only room correction that I don't have hands-on time with. I recognize your username from over on AVS. It might be worth reading through the ARC Genesis or asking in the REW thread over there to see if someone with ARC experience can help get you sorted out.
@@jeffmery thank you,
when you find what db curve you want with the low shelf. can't you just keep using it that way ? why do you go and add it to the eq house curve setting?
You can absolutely just keep using it that way.
The main reason to put it on the output PEQ filters is for BEQ users. BEQ implements all of its filters on the input side. If there's anything you need, BEQ will overwrite them. Not everyone uses BEQ though (I don't anymore myself).
Hi Jeff. Great content! I own a MiniDSP Flex and am wondering if you could give me a brief overview (high level summary) of how best to set this up in a 2.2 system. I'll take a stab at it - maybe you or someone here tell me what I got wrong? Thanks!
Step 1: time align subs to each other (take individual sub measurements using acoustic timing reference, then use REW's time alignment tool)
Step 2: take a measurement of both L and R speakers playing together (using same acoustic timing reference), then time align that measurement to the time aligned subs?
Step 3: Apply EQ (up to 400 or 500 Hz). Do I do this in two phases? One for the subs (following this video as a guide), and one for the speakers? Or am I supposed to apply EQ to the system as a whole (using a measurement of all 4 speakers playing together)?
Step 4: Find the best crossover point using REW and set that in the MiniDSP app. Any suggestions on this? Should I do this BEFORE step 3?
Thanks, Josh. Unfortunately, I have zero experience with a 2.2 setup. That being said I think I'd try a slightly different process.
Step 1: Time align subs as above
Step 2: EQ subs
Step 3: Measure L+R together as above
Step 4: Implement crossover & time align subs w/ L+R at crossover
Step 5: Fine tune the overall EQ
I would implement the sub EQ on the miniDSP output PEQ. I would implement the overall EQ on the miniDSP input PEQ.
I have no idea what results this would yield, but it's the first way I'd try to do it using just REW.
If you're feeling adventurous, you can give MSO a shot. Someone just created a video showing an integration like this. He's using a mic that takes a ton of measurements at once, but the process with a single measurement point is really the same: www.avsforum.com/threads/optimizing-subwoofers-and-integration-with-mains-multi-sub-optimizer.2103074/post-61779385
@@jeffmery Thanks for the reply. I’ll watch that other video you linked and then start tinkering. Thanks again!
@@run4b33r Sure thing. Let me know what method you end up using and how it turns out for you!
I think REW's EQ is great for a SINGLE seat only. But if there's multiple seats, wouldn't MSO do a better job overall than REW? Correct me if I'm wrong.
I agree 100% and mention this explicitly in a few of my others videos.
If all you care about is a single MLP, then REW will provide excellent results with much less time and effort compared to MSO. There's a place for both tools depending on your specific room and goals.
Thanks a lot for this video. I have been fiddling with REW for a while and your videos explain the functions and their meanings really well. Just added a sub (only one). My woofers go down to ~20hz. I found that crossing them at 60-80hz and letting the sub handle all the lows makes the system sounds worse. If I EQ the sub alone (@80hz) and leave the woofers to play all the way down as well the system sounds much better. Did you try to play subs and speakers together i.e. overlapping the same frequencies in the low end? I use the miniDSP 4*10HD and with 5 EQ filters on the inputs and another 5 on each output it is hard to really reach flat in my room (3m*11m). It seems that to do that you need something with many more filters like Dirac/audiolense.
There's a lot to unpack here and a lot of different things to consider. With more than one listening position, it's going to be nearly impossible to get an even response in more than one seat with a single sub. It might be possible to get an even response at one position with one sub, but that heavily depends on the room and the positions of the listener and sub in relation to the modes in the room.
Sounding "better" or "worse" is too subjective. We really need before and after measurements to quantify what we're hearing. I thought my new subs sounded worse than my old one. Only after taking before/after measurements, did I realize that I was used to poor quality bass. What I was hearing was significantly better (more accurate, with more detail and much better decay), but I needed to get used to the sound of high quality bass.
I don't mean this response to come off as "you don't know what you're hearing". It's absolutely that *I* have no idea what you're hearing, or used to hearing, so it's hard to offer specific advice to improve things.
@@jeffmery WOW, thanks a lot for this elaborate response. Indeed it also measured worse. The sub has a roll of at ~26hz. So combining the mains and sub letting the sub reinforce the mains over the same frequencies works better. Another thing is that in my case the miniDSP is everything, preamp, DAC and RC. Unfortunately it only allows up to 9msec delay so aligning the sub and mains is also a challenge.
@@ofergover2978 I used a miniDSP 2x4HD in a very similar way in a desktop setup. They're versatile devices :-).
if it measures worse with the sub, then it's likely to be either phase cancellation between the mains and sub. This could be fixed by delay, but it seems like there may not be enough of that in your hardware. You could potentially use the alignment tool to see what would be required. Then you'd know if the 9ms was enough or not.
Towards the end of the video, you were adding a house curve using the input side (3 db, 6 db, etc). Then you went to EQ them again on the output side (and zero-ing the input side), and you could save it in Config 1, 2, 3...
Why do you have to EQ them again on the output side? Can't you just keep what you did on the input side? (Config 1 = flat, Config 2 = +3db, Config 3 = +6db, etc).
I'd have to go back and watch the video. I know I jump around quite a bit so I honestly don't recall what I was doing at the time. You absolutely run things as you describe with different house curves on different configs. I did this myself when I owned a Yamaha AVR to build house curves for different content.
I had the same thought after watching. I can't run DIRAC on my AVR for a while, so this would be a perfect solution until I can. I imagine having a proper high-pass in place would remove any risk of overdriving the subs compared to doing everything in PEQs
Why didn't you use REW's alignment tool to time align your subs? When I get my second sub, I will be doing that and allowing my Onkyo's Dirac Live to EQ my subs as one. Right now, Dirac Live does an excellent job of EQing my single sub and adding a house curve.
On not using the alignment tool here ... I didn't want to teach multiple things in one video. It's already nearly an hour long and teaching the alignment tool would have added at least another 20-30 mins.
I have a separate video on using the alignment tool here: ua-cam.com/video/8bwpLfbLiZ4/v-deo.html
I think Dirac Live does a pretty darn good job. My previous processor had Dirac Live and the multi-sub DLBC license and results were great. My only major gripe is that the implementations vary from manufacturer to manufacturer. That can make helping folks somewhat difficult since the capabilities of the AVR / Processor are so different.
To get the most out of it, I'd suggest you verify the results with REW after uploading to your AVR. It's hard to compare a single point measurement against a multi-point measurement. However, in my experience, there can be differences large enough to warrant some investigation. Pay close attention to the crossover region between the sub(s) and non-sub speakers. Good luck!
@@jeffmery Thanks. I wonder if it would help to use the MiniDSP to separately implement a flat target curve EQ to each sub (measured at the MLP), then allow Dirac to EQ this new virtual sub.
@@azdonw I think that is a good approach. When I had a Dirac processor, I would use MSO to align and EQ my subs to flat. I'd then present them to the processor and Dirac as a single sub. I used Dirac to implement my house curves and align the subs with the non-sub speakers.
@@azdonw When using a miniDSP and something like REW or MSO, I will EQ to flat in the miniDSP. I'll then implement my house curve within my room correction software (Dirac in this case). This keeps my sub optimization separate from my room correction. Makes it easy to change things around and then return to a known good state if something goes sideways.
Besides using minidsp Iam using AU labs and works good. What else is out there?
Anything that can be used to 1) split a single input signal to multiple outputs 2) has a delay setting per output and 3) has a user programmable PEQ on each output. That's about it. There are a ton of devices that have this capability. REW can export EQ into formats supported by a lot of them. But even if there's not a direct export, PEQ parameters can be manually entered per-channel to get the same results.
When you run the EQ process, you can select "Generic" as the device. Then, in the EQ filters window, you can limit the number of PEQ filters to the same number supported by your device.
@@jeffmery What other softwares are out there like au labs for Mac?
@@avia4281 I use SoundSource published by RogueAmoeba. PEQs can be entered manually or via a file import using the "Headphone EQ" function. Ignore the name on that. It's just EQ and they have a doc detailing the file format.
Full disclosure - i only use this with a 2.1 system on my desktop setup. I've never tried, and probably never will, hooking it into the theater processor.
I'm getting 2 chirps on timing reference, beginning and end, is this correct?
Yes. It was changed to have a second chirp in a more recent version. Totally expected.
@@jeffmery
Thanks Jeff.... Before I sweep my 2 subs am I turning audyssey off, are there any other avr parameters prior please...
@@user-id4rl5qn1d Audyssey off. make sure the miniDSP is fully cleared out as well. Other than that should be good.
@@jeffmery Thanks for advice, been reading up on best way to time align subs to mains using rew, what's the simplest way please, seems to be multiple options.. if crossover is 40hz, run a sine wave and watch level on spl meter until it maxes out??
@@user-id4rl5qn1d Use the "Alignment Tool". ua-cam.com/video/8bwpLfbLiZ4/v-deo.html
Ive followed these steps multiple times and my actual is never even close to the predicted!
99% of the time it's a problem with the measurements or implementing the filters. Make sure you are using an acoustic timing reference for all measurements. That is a hard requirement. Make sure the gains and delays are cleared from the subs before measuring.
This is a great reference for figuring out the issue: www.andyc.diy-audio-engineering.org/mso/html/reference-manual/faq.html
@@jeffmery Acoustic Reference seemed to be the missing link. Thanks Jeff!
@@patrickmiller4987 Awesome! Easy to miss for sure 🙂