Small yet Mighty. The tale of the 12's

[SME]

Whatever you end up with, it's probably a good idea to compare the sound you get to no EQ and no Audyssey with the sub level set by ear.  For a bass boost can you use a Linkwitz Transform filter?  If you know the approximate Fb and Qtc of your enclosure, then you can use LT to extend the cut-off frequency a bit lower and maybe lower the effective Qtc.  I suspect this will sound a bit better than a high Q shelf at 30 Hz.

[EndersShadow]

I have no idea if the MiniDSP scan do a Linkwitz transform.....  I will take a look.

It’s a 5.5 cf box (so maybe 5.0 after bracing and drivers).... with 4 of the JBLs and I have modeled it in WinISD.....

 

I know boosting that much may not be a good idea, so I’m still experimenting to see how I can extend the bottom without sacrificing the top end too much...

 

I think for now I’m going to focus on dialing in the Mini and just turn off XT32 until I’m happy with the results..

[maxmercy]

Sometimes XT32 tries to do things it shouldn't, but overall, it is a pretty decent product.

JSS

[EndersShadow]

31 minutes ago, maxmercy said:

Sometimes XT32 tries to do things it shouldn't, but overall, it is a pretty decent product.

JSS

I’m pretty happy with XT32.... just tired of redoing it once a day lol.

 

Once I’m happy with the sub sweeps with the filters then I will redo XT32...

[EndersShadow]

Latest REW sweep... like this a LOT more...

 

 

 

[Ricci]

Do you have the raw measurement with no filters for comparison? 

[EndersShadow]

1 hour ago, Ricci said:

Do you have the raw measurement with no filters for comparison? 

Yes.  I took a TON of measurements once I got the distances locked in earlier last week.  I took measurements at the following locations: Armrest (far right side), middle of the outer cushion, between the middle cushion and outer cushion, the center of the middle cushion, and lastly between the middle cushion and the center of the sectional.  I measured (but didnt use in the average) the dead center of the sectional, but threw it out of the mix because its literally less than a foot away from one of the subs so the response will be crappy no matter what I do there.

The image below will help better explain what I mean in case its not clear.  The ones with X in them were taken but NOT used in the averages.  

*Note* the Sonotubes (up front) are NOT yet in play.

I took 3 groupings of those measurements at different distances in the room.  One close to the back, one in the middle of the cushion, and one in front of the seats. I then averaged all those measurements and used that average to generate the REW filters I've used (the fact the picture says REW 9 filters indicates it my 9th iteration* of filters I'm trying lol).

Once I install a new set of filters in the Mini I then redo XT32, (measuring the subs as a single combined unit) and run the same set of measurements again to get an UPDATED average measurement across that same grouping (show in the picture in green).

Then I added boosts until the response looked decent and ran those exact same locations AGAIN to get my LAST average (shown in the picture in tan).

 

*Each time I redo the setup I learn a bit more.  This time I decided to NOT cut down the top end nearly as much as I was and simply boost the low end to even it up.

@Ricci If you want me to I can zip them and send them to you to look at.  I'm not going to say I fully understand REW and what its telling me (outside the fact I have some long decay in the waterfall graphs and I think distortion is much higher than I'd like *as a result of my boost obviously*).
 

[EndersShadow]

So I played Godzilla King of the Monsters tonight....

 

The bass SCARED me with how hard it was hitting so I made an executive decision.

 

I went from this

 

 

To this

 

 

Was still stupid loud with lots of impact, but not red lighting the Crowns as bad.  Also by reducing the boost I was likely drastically reducing the distortion as these drivers aren’t really meant to dig this low...in addition to reducing driver stress....

 

I will do some REW later so I can see how much I lost below 20hz

 

This is what the Original settings provided (the tan line)

 

*Note* This was at a volume of 50 on the AVR.  Not sure what SPL that equates to....  I was listening to Godzilla at 70.... so I’m sure it was significantly higher SPL....

[EndersShadow]

Got a few minutes to redo my REW w the reduced PEQ filters.  I measured at 50 volume and 70 for giggles.

Lots more rolloff below 20hz compared to the boosted one but for now I’m ok with that.

Once the sonotubes come in perhaps room gain will help.

Original v Reduced PEQ

50 v 70 on the volume.  On this the rolloff below 20hz seems much more pronounced.  Not sure why this is.  I can tell you that things shake all over the house with the 70 volume lol...

And lastly here is a THD plot from my 70 volume measurements.  I think @Ricci asked for something like this at one point in this thread. Hopefully it shows what you want Ricci, if not, let me know and I can repost.

[SME]

15 hours ago, EndersShadow said:

Got a few minutes to redo my REW w the reduced PEQ filters.  I measured at 50 volume and 70 for giggles.

[...]

50 v 70 on the volume.  On this the rolloff below 20hz seems much more pronounced.  Not sure why this is.  I can tell you that things shake all over the house with the 70 volume

It looks like the amp and/or driver is completely out of gas below 16 Hz or so with the sweep at volume 70.  What volume do you normally listen at and what level (dBFS) was the sweep at?  You don't seem to be getting much room gain at all and may want to dial back the bottom-end boost a fair bit more to clean up the sound and avoid stressing the driver.

[EndersShadow]

44 minutes ago, SME said:

It looks like the amp and/or driver is completely out of gas below 16 Hz or so with the sweep at volume 70.  What volume do you normally listen at and what level (dBFS) was the sweep at?  You don't seem to be getting much room gain at all and may want to dial back the bottom-end boost a fair bit more to clean up the sound and avoid stressing the driver.

Normal listening is about 50 on the volume knob (so the lower sweep).

 

70 on then volume was a “Home alone with no kids/wife watching a good movie loud”.

 

You are correct about the room gain, not getting much, but with a massive open floor plan with only one corner that’s expected.

 

I did figure the amps were out of gas which I kinda expected with 2 bridged Crown XLS2000s on a single 20amp dedicated circuit.  And I’m adding two more amps to it so we will see how much worse it gets with those added.

 

I don’t know if I can dial the boost back without making the response worse (rolloff starting in the early 30hz range), so we will see...

 

I do know that even with my CRAZY boost earlier as long as the volume was kept around 50 it was safe for use, and had low distortion, same with the reduced PEQ at 50 on the volume knob.  I will try to post some of those distortion graphs for comparison later, but my computer is old and wont run more than one program at the same time and I've got to get my kido up for school :).

 

Got a second, ok here we go.

 

Distortion on the 50 volume sweep with reduced PEQ

SPL Difference between the 50 and 70 volume

 

And for giggles here is the distortion on my massively boosted PEQ sweeps (also at 50 on volume)

 

Here is what that boosted PEQ looks like in the MiniDSP

 

So looking at that as long as I keep the volume in check a bit more I can run with larger boosts, which was what I was planning on.  

I'm buying a MiniDSP wireless adapter so I can load in and adjust things on the fly easier, which means I can dial up the boost when doing my normal thing, and then dial it back when its just me and I can go NUTS.....

[SME]

Let me ask again, a bit different:

what kind of REW sweep is this?  Which channels and what level in dBFS?  If this is, say, "-10 dBFS" on left and right channels only, that's gonna be a lot lower in level than what a movie can do at the same MV.

[EndersShadow]

1 hour ago, SME said:

Let me ask again, a bit different:

what kind of REW sweep is this?  Which channels and what level in dBFS?  If this is, say, "-10 dBFS" on left and right channels only, that's gonna be a lot lower in level than what a movie can do at the same MV.

Lets see if I can answer this is any sort of smart way lol. I think I know what your getting at, but I'm going to give you all the data points on everything anyway just to cover my bases.

 

Here are the settings in my Denon X4100

• Audio/Volume: Scale - 0-98* | Limit: 80
  • *My options for the scale are -79.5db -18db OR 0-98 (which is what I have mine at)
• Speakers/Crossovers: Front: 80hz
• Speakers/Bass: Subwoofer Mode: LFE | LFP for LFE: 80hz
• Speakers/Level: Subwoofer: -2.0
• Audio/Subwoofer Level Adjust: +6.5db

 Here are the REW settings:

 

 

Each REW sweeps was run with those settings (for good or bad), the ONLY diff was the volume change from 50 (out of 98) to 70 (out of 98)

[SME]

I see you're sweeping on "6: Display Audio Output 1 4".  Is that just a single channel?  Is it LFE?  Or something else like center?

[EndersShadow]

3 hours ago, SME said:

I see you're sweeping on "6: Display Audio Output 1 4".  Is that just a single channel?  Is it LFE?  Or something else like center?

Channel 4 is the subwoofer channel.  Here is the default channel mapping per MiniDSP

Here is a typical mapping of HDMI channel numbers:

• Channel 1: Front Left
• Channel 2: Front Right
• Channel 3: Center
• Channel 4: Subwoofer
• Channel 5: Surround Left
• Channel 6: Surround Right
• Channel 7: Surround Back Left
• Channel 8: Surround Back Right

https://www.minidsp.com/applications/acoustic-measurements/206-umik-1-asio4all#channels

[SME]

Thanks!  I'm using Linux, and the 7.1 channels are ordered differently, even with HDMI output.

So basically you are running sweeps on the LFE channel (not really "SUB" in this case) at "-12 dBFS", which is in terms of RMS level of the sine wave vs. peak output.  That's actually approx. -9 dBFS in terms of how much headroom there is before the sine wave would clip.  This is definitely quite a bit less than what you may encounter in an actual movie track.  A movie track includes bass managed content from L+R+C and surrounds, which adds up to 10 dB vs. what LFE alone can do.  Typical big effects in a heavy hitter may be "-5 dBFS" in terms of all the channels.  A handful of movie effects go almost completely full-scale.

So what does this mean?  It means max output demanded when viewing a movie is likely to be quite a bit higher than your sweep level.  Like "-5 dBFS" wrt all 7.1 channels playing is comparable to a sweep level that's 14 dB higher than you are doing.  For ULF, it's not quite that bad because movies don't often use the whole bandwidth just for ULF, but I would say that based on your measurements and info, you're probably running pretty close to the limits with a heavy hitting film at "MV 50".  If you ever want to play at "MV 70" you should get rid of all boost and maybe turn the sub down overall.  But then why bother?

Something else to look at is to make certain your gain staging is optimized between AVR/processor, MiniDSP, and sub amps.  It's possible that your "MV 70" sweeps are clipping the signal before it even reaches the amp.

[EndersShadow]

6 hours ago, SME said:

?Something else to look at is to make certain your gain staging is optimized between AVR/processor, MiniDSP, and sub amps.  It's possible that your "MV 70" sweeps are clipping the signal before it even reaches the amp.

Yes I need to go back into the MiniDSP and sweep at 70 and look at the INPUT levels and output levels to make sure I’m not clipping.

im fairly certain I’m not clipping the output (as I’d be red lighting the lights on the amp and I’m not), but I could very much be clipping the input.

 

I think I tested that much earlier and determined I wasn’t both by looking at sweeps and also by monitoring it while playing real world content like the EOT intro, EndersGame clips and a few others know to contain lots of UKF/Bass content....

 

But I’m betting I did that at 50 not 70 as I had a hard limit set at 60 (so kids can’t crank it on accident), and increased that to 80 very recently.

 

You are also right as what I really want to do is figure out where my limit is on this configuration so I can be under or at it.  If I remove the boosts I will lose out a massive portion of the low end I have now.  I’ve attached a sweep before XT32 and any REW filters are installed (of the subs just as they sit.  You will notice they run out of gas fast stock without the low shelf filter and the other boost is just there so they’re isn’t a dip when that low shelf filter starts.

*The darker line is the response.  The green line is the EQd response and you can ignore the red circle, it was added to show the first section I was EQing lol...

I’m not there yet but working towards it lol.  I’ve got some buddy’s in Chicago that I’m going to bribe/beg/pay to come down once all my subs are done and help me dial them in but since I am still not totally setup I’m flying solo right now.

 

[SME]

10 hours ago, EndersShadow said:

Yes I need to go back into the MiniDSP and sweep at 70 and look at the INPUT levels and output levels to make sure I’m not clipping.

im fairly certain I’m not clipping the output (as I’d be red lighting the lights on the amp and I’m not), but I could very much be clipping the input.

Each device in the chain can clip on either the input or the output.  You'll likely only see clip lights on the amp when clipping on the amp output.  You definitely could be clipping the MiniDSP output without seeing amp clip amp lights.

[Ricci]

SME's post is spot on. It takes careful investigation to figure out where clipping is occurring. MiniDSP's are not known for having robust voltage headroom on either the input or output. EQ complicates matters. 

Many receivers will clip the sub output internally under a WCS signal and well below maximum output voltage. The amp is the easy part of the gain stage. Most have clip indicators and as long as the gain control is up a few clicks from the minimum the input won't clip before the output. 

[EndersShadow]

I agree with you both.  That said I got the balanced 2 x 4 which I think has 4 VRMS in and out (still not a ton but better than the stock 2x4).

 

 

 

I know the amp lights on my Crown XLS2000s aren’t hitting red so there’s that but I do plan to spend some time doing some real world content viewing (tomorrow while wife and oldest are at a JoJo Siwa concert lol) while looking at the input and output of the MiniDSP.

 

The red lights on the Crown are indicators of the input signal correct? So if they aren’t red lighting 24/7 and normally are barely 1/2 way at least the AMP isn’t clipping (but clipping could still be occurring at the MiniDSP input or AVR output) correct? The knobs on the Crowns are simply the gain and they are 4-5 clicks below max.

 

I lack the skills and tools to verify if clipping is happening inside the AVR itself lol...

[EndersShadow]

Ok.  Ran threw some demo clips* and I’m at the ragged edge on the MinDSP input at 70 (which is above my normal listening volume of 50).  

 

Im not clipping except in rare really loud circumstances.

 

Output seems fine as is.

 

Just to be a bit safe I dialed down the subwoofer adjust from +6.5 to +5.0 so that should help white a bit.

 

For now I think my gain structure is maximized outside potentially adding a high shelf filter that kicks in around 15hz to help reduce amp and driver strain since it seems they are close to out of gas there....

 

*Clips were:

Intro to Godzilla King of the Monsters - Large footsteps

Captain America Winter Soldier - HeliCarier launch

Olympus has Falledn - C130 attack (when the monument falls I triggered the biggest red light event on my amp so far)

The Polar Express - Train Arrival

[SME]

Did they improve the level monitoring feature in the GUI?  On my old MiniDSP 2x4 unbalanced and OpenDRC-AN, IIRC, the displayed levels were short-term averages.  (IIRC on the 2x4 it wasn't even RMS average.)  This would tend to miss momentary high output transients in movies that might clip the output.  If it's indicating short-term average then "riding the ragged edge" means you might actually be clipping a lot.

Here's one experiment you can try.  Play a movie clip and note the highest level that pops up.  Maybe play that part a couple times to make sure you get the same number.  Then, add 3 dB headroom to the MiniDSP output by decreasing the output gain -3 dB and increasing the amp gain +3 dB.  This should give the same average volume, but you'll have 3 dB more headroom.  Now play the movie clip again and see if the highest level numbers you get are -3 dB compared to before or if they are higher.  If they are higher, then you were probably clipping and may still be clipping.

Sorry to make this a chore.  It makes a big difference for sound quality to not clip or distort transients all the time.  Depending on circumstances, this may not be nearly as obvious as with speakers because the sub output drops a lot in the high frequencies, meaning the clipping won't sound as harsh.  IMO gain structure problems and internal clipping are also a very common problem in HT systems if they get turned up.

[EndersShadow]

4 hours ago, SME said:

Did they improve the level monitoring feature in the GUI?  On my old MiniDSP 2x4 unbalanced and OpenDRC-AN, IIRC, the displayed levels were short-term averages.  (IIRC on the 2x4 it wasn't even RMS average.)  This would tend to miss momentary high output transients in movies that might clip the output.  If it's indicating short-term average then "riding the ragged edge" means you might actually be clipping a lot.

 

No idea if they improved the monitoring features, just know they are there and can help you judge if your clipping the input. 

What I meant by that was when playing movies only the parts meant to be REALLY loud approached or hit the -0db point on the input graph.  I didnt hit it on the output graph at all, although I noticed when I got close I was only at the third of 5 lights on the Crowns.  In general on the MiniDSP input graph I was well below 0 with most stuff (like hovering around -10, -12).  I'm fine if it hits 0db once in a while, but it wasnt steady in like the -2, -3, etc except again for the times that it should be.

I also didnt repost last night, but I rewatched the first 45 minutes of Godzilla and it did the same thing.  Most content was in the green/yellow area of the graph, with occasional super loud sections (like godzilla's fire breath, or footsteps) approaching between -3 or 0 on the graph.

I also threw on some music (Lindsey Stirling Artemis, Tool Fear Inoculum and the signal was coasting through all of that with ease as well (though music content doesnt normally dig below say 25hz lol)

Now I was likely clipping the first time I watched the movie even after reducing the boost (in the MiniDSP) because I had the subwoofer level at +6.5db on the Denon.  I adjusted that setting in the Denon down 1.5 to just +5.0db prior to doing any testing/listening last night.  That means that when watching the movie last night all parts that were approaching clipping (at the new setting of +5.0db) on the input graph likely were well past it initially at the old setting of +6.5.

 

4 hours ago, SME said:

Here's one experiment you can try.  Play a movie clip and note the highest level that pops up.  Maybe play that part a couple times to make sure you get the same number.  Then, add 3 dB headroom to the MiniDSP output by decreasing the output gain -3 dB and increasing the amp gain +3 dB.  This should give the same average volume, but you'll have 3 dB more headroom.  Now play the movie clip again and see if the highest level numbers you get are -3 dB compared to before or if they are higher.  If they are higher, then you were probably clipping and may still be clipping.

Sorry to make this a chore.  It makes a big difference for sound quality to not clip or distort transients all the time.  Depending on circumstances, this may not be nearly as obvious as with speakers because the sub output drops a lot in the high frequencies, meaning the clipping won't sound as harsh.  IMO gain structure problems and internal clipping are also a very common problem in HT systems if they get turned up.

While it may be a chore, you are correct, the payout is worth the effort.  One of the reason I am moving to 12 12's is to have a lot of cone area to hopefully allow me to play loud without having to stress.move the drivers tons (except for special occasions).  I also am trying to be very conscious about not clipping because I want as clean a sound as I can get (understanding I'm already asking more from these drivers below 30hz than they'd really like to do).

I am also not one to say I'm an expert at gain structure, but I want to make sure I get it right, again because I want to get the best sound I can out of what I have.  If I had my druthers I'd build 2 really large ported LTT's and be done, but thats not a viable option for me.

Two questions (sorry if they are remedial ones):

• When you say "decrease the output gain -3db"..... where? In the MiniDSP input gain? output gain?
• How would I know how much to bump up the volume on the amps to increase 3db? Run a sweep at the beginning, reduce the input, increase gain until the sweep with the changes matches the original one?
  • I also dont know if I have 3db left on the amp to give lol, I think I'm maybe 4-5 clicks below max on my Crowns full volume, which is likely NOT going to be a 3db increase

[SME]

18 hours ago, EndersShadow said:

No idea if they improved the monitoring features, just know they are there and can help you judge if your clipping the input.

Oh?  I wonder how it detects that?  I guess I'd have to read the manual or something to better understand what it's doing.  Maybe.

18 hours ago, EndersShadow said:

What I meant by that was when playing movies only the parts meant to be REALLY loud approached or hit the -0db point on the input graph.  I didnt hit it on the output graph at all, although I noticed when I got close I was only at the third of 5 lights on the Crowns.  In general on the MiniDSP input graph I was well below 0 with most stuff (like hovering around -10, -12).  I'm fine if it hits 0db once in a while, but it wasnt steady in like the -2, -3, etc except again for the times that it should be.

Like I said, that's only a short-term average, then you could be clipping a lot more than you realize.  The peaks could be quite a bit higher and be clipping even though you don't see it go over "0" much.  To avoid this complication, you need to use an artificial signal like a sine wave.  You may want to do this with the speakers/sub(s) disconnected, if your amp is OK running like that.  Then you can test using high level sine waves without breaking anything.  Presumably REW lets you run a "worst-case scenario" sine wave by setting hte level to near full-scale (-3.02 dB in REW) and sending it to *all channels*.  Start with the master volume low and watch the MiniDSP input volume.  Slowly increase it and watch to see if it tracks the volume changes dB by dB.  If the level shown in the MiniDSP does not increase as much as it should, then you know you are clipping, either in the Denon output or MiniDSP input.

If your Denon is like mine, then MV "70" is the same as "-10".  If you are using bass management but have Audyssey turned off, then you need to keep the sub trim at -2 dB or below to ensure nothing in the receiver clips on a worst-case scenario (WCS) event at MV "70".  With Audyssey turned on, you need more headroom, and it's tough to say how much.  IIRC Audyssey can apply boost as much as +10 dB, and that's not counting the extra headroom needed for other aspects of the EQ.

If you are clipping on the MiniDSP side, IIRC you can set a jumper to attenuate the signal coming in so that it doesn't clip.  My Denon puts about a bit over 4 Vrms, which may be too hot for the MiniDSP without the jumper set.  Lastly, you want to optimize the MiniDSP output gain vs. the AMP gain/volume.  Here you have two options.  One is to ensure that a WCS signal will pass cleanly through the MiniDSP, repeating the process with the WCS signal and the gradually MV, monitoring the output level.  Again, it should increase in step with the MV.  If it doesn't it's starting to clip.  This is tricky because the WCS may be hard to predict with a lot of EQ in use.  This method may also lead to enough noise as to be annoying.  Another method is to determine when the amp starts clipping and then dial-in some extra headroom.  I do this, giving myself ~+3 dB.  This requires knowing whether the amp is clipping/limiting or not.  This method is simpler and minimizes noise, but it's still possible to clip the MiniDSP output if you are running beyond the headroom gap between the amp and the MiniDSP.  For me that's not a problem.

I've oversimplified the process a bit because the amp may not have very fine gradations in volume.  So you may need to set the amp to the notch that is higher than required and then compensate by adjusting the MiniDSP and/or Denon output(s) down to maintain the right level balance right with the mains.

Hopefully this makes sense?  Another option which works to test the Denon and MiniDSP input and outputs for clipping but not the amp itself is to turn the amp down to its lowest setting and connect the sub(s) to it.  Then when you adjust the master volume up with the WCS signal playing you should hear a sudden increase in distortion after you hit a clipping point somewhere.

19 hours ago, EndersShadow said:

 

• When you say "decrease the output gain -3db"..... where? In the MiniDSP input gain? output gain?
• How would I know how much to bump up the volume on the amps to increase 3db? Run a sweep at the beginning, reduce the input, increase gain until the sweep with the changes matches the original one?
  • I also dont know if I have 3db left on the amp to give lol, I think I'm maybe 4-5 clicks below max on my Crowns full volume, which is likely NOT going to be a 3db increase

Question 1: I essentially answered above.  You need to adjust output gain to optimize match with the amp and input gain to adjust match to the Denon.

Question 2: That seems like that would work if you don't know how much you gain with each "tick".  I'd guess that you get at least 1 dB per click on the Crown.

[maxmercy]

What you really need is an oscilloscope.  You can do it with a soundcard and a voltage divider and some apps are out there for it.

To view a signal realtime:

https://makezine.com/projects/sound-card-oscilloscope/

Or, you can use REW itself and use the Oscilloscope window and look at the captured trace and look for flat tops.  Measure from each input/output to ensure no clipping with a -0dBFS sweep level.  Just be sure not to fry the soundcard input by checking levels carefully or using a voltage divider.

I was thinking of making a true WCS test disc, but REW is pretty good at doing almost everything the disc would do.  With REW and EqualizerAPO  (https://sourceforge.net/projects/equalizerapo/) you can simulate a WCS, and see where it clips in the signal chain (watching to make sure you do not clip any inputs), and then KNOW you will be free from clipping no matter what signal content is fed into the system. 

EDIT:  Apparently REW beta now has a LIVE oscilloscope feature tracking the measurement and reference channels in real time.  THAT is a very handy tool.  If you guys have PayPal, donate a few bucks to REW.  It is truly a terrific tool that many have gotten better audio experiences from.

JSS