Graphs from REW

[PENG]

In the REW plots below, blue is Audyssey with minidsp 2XHD, it got rid of the dips at 50 and 70 Hz.

Mr. TheWarrior seems to be a helpful person, but I disagree with what he said or implied (not saying he's all wrong, just seemed extreme to me) about Audyssey, iirc he told me Audyssey would mess up the phase of my speakers/subs, without first viewing any of my REW plots. When I did post one, he did not follow up so he might have changed his mind a little but anyway that's distant memory that I shouldn't even have remembered.

I have a high regard for Dr. F. Toole but I am also have high regards for other well experts who may or may not have shared some of the same views that Dr. F. Toole has in his books. I have yet to order his latest book to learn more, as so far I have been relying material that he has made available online for free.

Minidsp did make slight improvements for me but with Audyssey alone, I was getting +/- 3 dB 18-90 Hz 1/24 smoothing and better in multi channel stereo, so I moved the mini to one of my stereo only system.

I am of the opinion (based on my perceived logic) that using the 2XHD with Audyssey SubEQHT will not likely get me overall better result for multi-channel material, any difference will probably be insignificant. For Stereo application, it most likely will work very well with SubEQHT/XT32, but at the moment I don't have an extra suitably equipped AVR for otherwise a good experiment for me.

 

[TheWarrior]

In the REW plots below, blue is Audyssey with minidsp 2XHD, it got rid of the dips at 50 and 70 Hz.



View attachment 22390

Mr. TheWarrior seems to be a helpful person, but I disagree with what he said or implied (not saying he's all wrong, just seemed extreme to me) about Audyssey, iirc he told me Audyssey would mess up the phase of my speakers/subs, without first viewing any of my REW plots. When I did post one, he did not follow up so he might have changed his mind a little but anyway that's distant memory that I shouldn't even have remembered.

I have a high regard for Dr. F. Toole but I am also have high regards for other well experts who may or may not have shared some of the same views that Dr. F. Toole has in his books. I have yet to order his latest book to learn more, as so far I have been relying material that he has made available online for free.

Minidsp did make slight improvements for me but with Audyssey alone, I was getting +/- 3 dB 18-90 Hz 1/24 smoothing and better in multi channel stereo, so I moved the mini to one of my stereo only system.

I am of the opinion (based on my perceived logic) that using the 2XHD with Audyssey SubEQHT will not likely get me overall better result for multi-channel material, any difference will probably be insignificant. For Stereo application, it most likely will work very well with SubEQHT/XT32, but at the moment I don't have an extra suitably equipped AVR for otherwise a good experiment for me.

Thanks PENG. I know we've discussed Audyssey on multiple occasions. I apologize if I left you hanging, in the past.

With the example above, those are very pretty graphs showing that each EQ solution provided a different result. But it is only one piece of the necessary information to make any useful comments. And that seems to be where my 'telling' rather than 'showing' has gotten me into trouble.

If you agree and understand that when it comes to reproducing bass frequencies in small rooms (defined as rooms having dimensions smaller than the wavelengths being reproduced) then it should be easy to understand why everything matters: The dimensions of your room, the furnishings, the orientation and distance of/from listeners and speakers, the LFE crossover, the acoustic phase of the subs, the DSP inputs, the resolution of the measurement, the performance across a broad range of frequencies - all of it needs to be known to make informed decisions. There would be far too many pages to cite in order to quote Floyd saying 'everything matters'.

I understand you to be an engineer who does not shy away from math. Perhaps after my 'showing' (that is so inappropriate on a male dominated forum) I might be able to sway you in to crunching those numbers to calculate your modes. You have tested solutions that yield your stated opinion. Perhaps curiosity/obsession/audioholism will make you want to try something different after the 'show'. (Just stop already)

If nothing else, hopefully I can crack a joke, on occasion.

Sorry I am painfully slow in showing results, to everyone.

 

[Pogre]

I just want to know, after I have repositioned everything and can't move much around any more, how do I apply the knowledge to be helpful? Not to mention getting measurements with any degree of accuracy would be nearly impossible given my layout (multiple open spaces leading elsewhere, a hallway, notched out parallel surfaces, vaulted ceilings, etc...).

I get that knowing these things would help with my understanding of what's happening in my room, but what would I do differently than I already have? And I can assure you Alex, my system doesn't sound like garbage because of arbitrarily applied filters, adjustments and exaggerated bass bloat like you said you were going to assume. You're making a lot of assumptions when you make statements like that, and it's a wee bit insulting.

 

[Pogre]

Sorry I am painfully slow in showing results, to everyone.

I'm not busting your balls for being a little slow, but haven't you been working on it for like, a year? I don't want to spend a year (or even months) measuring my room and crunching numbers. It's a bit much, don't you think?

 

[ATLAudio]

I just want to know, after I have repositioned everything and can't move much around any more, how do I apply the knowledge to be helpful? Not to mention getting measurements with any degree of accuracy would be nearly impossible given my layout (multiple open spaces leading elsewhere, a hallway, notched out parallel surfaces, vaulted ceilings, etc...).

I get that knowing these things would help with my understanding of what's happening in my room, but what would I do differently than I already have? And I can assure you Alex, my system doesn't sound like garbage because of arbitrarily applied filters, adjustments and exaggerated bass bloat like you said you were going to assume. You're making a lot of assumptions when you make statements like that, and it's a wee bit insulting.

Even when you have an "easier" room like mine, I have considerations where a tape measure and a simple formula can't help. My ceiling has 2 soffit vents, one which takes up the front of room, but not the whole length of it. One wall has a huge set of French Doors, and tons of glass. One wall has no studs, and has a closet full of junk along the whole length, and an exterior wall behind it. There's also virtually nowhere else I can position my subs and speakers and seats to say nothing about measuring all the loft and fluff and wood which make up my furniture.

Bottomline:
Have realistic goals and expectations. I have a front sofa, I want a smooth response across the three seats. I verified this using multiple REW measurements.

REW can show you your room modes and has plenty of tools to do this. Multiple people have posted about this as well. This is indisputable information and doesn't rely on a math and measuring task which is astronomically impossible to duplicate.

Everything does matter, but attempting to calculate anything meaningful of that everything with just a tape measure and high school algebra is absurd. Also it's just the law of averages that dicates that some unknowns will have less of an impact on others.

Again @gene with Toole's peer review wrote an article on doing this. If any crucial information was abridged, contact Toole and ask why he didn't demand its inclusion.

 

[Mitchibo]

Here's where I'm at right now.

View attachment 22388
Gene posted another method I haven't tried yet in my thread that I might tackle this weekend.

View attachment 22387
This is one of many sets of sweeps I did while I was collecting data and calibrating. This one shows me that a higher crossover definitely gives me a better response. I didn't use smoothing on any of my measurements either. I wanted it all. The good, the bad and the ugly.

Is this a composite of subs only or all woofers available?

 

[Pogre]

Is this a composite of subs only or all woofers available?

All woofers available. I had my towers active for the sweeps.

*Edit: My sub only sweeps look a lot better, but since I'll be using my towers to listen it just made sense to include them with my sweeps. I'm very interested in trying out what gene suggested with bi amping my front towers to treat the woofer sections like 2 more subs, then sum them all as one.

 

[William Lemmerhirt]

I just want to know, after I have repositioned everything and can't move much around any more, how do I apply the knowledge to be helpful? Not to mention getting measurements with any degree of accuracy would be nearly impossible given my layout (multiple open spaces leading elsewhere, a hallway, notched out parallel surfaces, vaulted ceilings, etc...).

I get that knowing these things would help with my understanding of what's happening in my room, but what would I do differently than I already have? And I can assure you Alex, my system doesn't sound like garbage because of arbitrarily applied filters, adjustments and exaggerated bass bloat like you said you were going to assume. You're making a lot of assumptions when you make statements like that, and it's a wee bit insulting.

Precisely my question as well. Guess we’ll have to wait and see!
Also, to assert that your room sounds bad because the graph is smooth and seemingly fixed at random is just not fair. REW author john mulcahy is a smart cookie, and REW is not an arbitrary piece of software. If it makes an adjustment, it’s not by accident. I’m not saying it’s perfect, and certainly it’s more granular features get glossed over by many a layman, but it’s pretty damn good. So, like you’ve asked a few times poges, I’d be curious to see what a graph should look like to support great sound, if not one that looks like yours. (Yes I know an spl graph is only part of the picture)

 

[agarwalro]

When dealing with imperfectly cuboidal rooms, you have to fudge things to get L, B H and kick of the modal analysis calculation. Here are a few examples:

Wall niche/cut outs - Ignore them if small, consider it a flat surface. If large, do the analysis twice. Once ignoring the niche depth another time adding the niche depth to that dimension.
Slanted ceilings - Run the math three times using minimum, maximum and half way height.
Openings to passages - Ignore them.
Wall too wall cupboard/ bookshelf - Use solid surface behind the doors or shelving as the boundary.

All the approximations work since their dimensions are small compared to wavelength in bass bands.

The end goal of this exercise is to get all the possible modes and see which frequencies have overlapping modes or bands that have a lot of them clustered. Audio anomalies therein will not be removed by EQ/DSP alone. Something physically needs to change, aka. subwoofer or seat location, multiple subwoofer, bass traps/resonators, etc. Having a flat FR does not make a system calibration complete. It's the minimum.

Ironically, while making the math unappealing to all but masochists, these very deviations (from ideal/cuboidal room) also impeded room induced resonances and break up mode clusters.

 

[Pogre]

When dealing with imperfectly cuboidal rooms, you have to fudge things to get L, B H and kick of the modal analysis calculation. Here are a few examples:

Wall niche/cut outs - Ignore them if small, consider it a flat surface. If large, do the analysis twice. Once ignoring the niche depth another time adding the niche depth to that dimension.
Slanted ceilings - Run the math three times using minimum, maximum and half way height.
Openings to passages - Ignore them.
Wall too wall cupboard/ bookshelf - Use solid surface behind the doors or shelving as the boundary.

All the approximations work since their dimensions are small compared to wavelength in bass bands.

The end goal of this exercise is to get all the possible modes and see which frequencies have overlapping modes or bands that have a lot of them clustered. Audio anomalies therein will not be removed by EQ/DSP alone. Something physically needs to change, aka. subwoofer or seat location, multiple subwoofer, bass traps/resonators, etc. Having a flat FR does not make a system calibration complete. It's the minimum.

Ironically, while making the math unappealing to all but masochists, these very deviations (from ideal/cuboidal room) also impeded room induced resonances and break up mode clusters.

This, at least is explanatory. Thank you. Still no easy task tho. One of the openings from my listening area is to a 10 x 12' room. The opening is an archway that's 62" wide so I can't really ignore it.

After that I have a hallway. The hallway is also definitely getting pressurized. It has a dogleg in it, but I can feel and clearly hear the bass all the way back to the end of it (in the bedrooms too if I open the doors, which I normally don't). If I could close it off I suspect it would have massive effect on how my subwoofers behave. Wouldn't I want to include it?

I get what you're saying about roughing it out a little and ignoring a few things is okay, but I have a lot of uneven parallel surfaces and a dogleg hallway. I don't have a single wall that's straight for its entire run. Big pop outs for archways, a pop out for the pantry with the fridge right beside it (it's a great room layout with a kitchen behind me), big openings, windows, a sliding glass door and more. It's a lot to account for or try to guesstimate. And if I get it wrong...

Or use REW (which still required time and effort, but not nearly as much as taking physical measurements and crunching numbers on the back of an envelope) and get close to the same results. What if my fudging and guesstimating while crunching numbers gives me something completely wrong? I'd be banging my head off the wall!

At least with rew I can see and hear an improvement. There's no doubt about it, and based on the room's behavior to my adjustments I think I know where most of my room modes are. That's where I went with placement and distance settings instead of peq. I was able to improve things with placement, but still there when I take measurements. The "trouble spots" are always in the same bands.

I'd also repeat Bill's question, what should a graph of a good response look like? I did generate waterfalls, spectrum, etc. also.

Phew. I swear I meant to do just a short response! I'm not challenging, I'm asking. I really want to know.

 

[Mitchibo]

This, at least is explanatory. Thank you. Still no easy task tho. One of the openings from my listening area is to a 10 x 12' room. The opening is an archway that's 62" wide so I can't really ignore it.

After that I have a hallway. The hallway is also definitely getting pressurized. It has a dogleg in it, but I can feel and clearly hear the bass all the way back to the end of it (in the bedrooms too if I open the doors, which I normally don't). If I could close it off I suspect it would have massive effect on how my subwoofers behave. Wouldn't I want to include it?

I get what you're saying about roughing it out a little and ignoring a few things is okay, but I have a lot of uneven parallel surfaces and a dogleg hallway. I don't have a single wall that's straight for its entire run. Big pop outs for archways, a pop out for the pantry with the fridge right beside it (it's a great room layout with a kitchen behind me), big openings, windows, a sliding glass door and more. It's a lot to account for or try to guesstimate. And if I get it wrong...

Or use REW (which still required time and effort, but not nearly as much as taking physical measurements and crunching numbers on the back of an envelope) and get close to the same results. What if my fudging and guesstimating while crunching numbers gives me something completely wrong? I'd be banging my head off the wall!

At least with rew I can see and hear an improvement. There's no doubt about it, and based on the room's behavior to my adjustments I think I know where most of my room modes are. That's where I went with placement and distance settings instead of peq. I was able to improve things with placement, but still there when I take measurements. The "trouble spots" are always in the same bands.

I'd also repeat Bill's question, what should a graph of a good response look like? I did generate waterfalls, spectrum, etc. also.

Phew. I swear I meant to do just a short response! I'm not challenging, I'm asking. I really want to know.

It seems like you are still on the right track. Do you have multiple lps you are aiming to satisfy? If not, then a great set of speakers and tune everything to where “ you” sit and turn up the volume. Seriously it sounds like you’ve done all the things available to you. You’ve got your XO working correctly, you may have the best possible placement, after this passive treatment and equipment seems like the only viable solutions left. What say everyone else?

 

[VonMagnum]

Alright, I finally unpacked my UMIK-1, connected it to my old Macbook Pro and took some measurements of my AL-III Dipole Ribbon speakers (see photo below; they're 6 feet tall). After playing with the active crossover Q/EQ/Trim controls to see exactly what effect they had on the room, etc. I took the measurements below in the graph and other than changing the level for the bass to make a "flat" graph (versus the boost I like to listen with),I ended up almost the same as I set by ear for the final graphs. I don't know what a "typical" room response interaction looks like with a variety of speakers so you'll have to tell me if you think it's decent or not considering it's got no EQ other than the active crossover settings that mimics the original passive controls on the back of the AL-III speaker (before the cheaper "S" model came out and got rid of the controls to save on the crossover costs). The total adjustment range is like 3-5dB I think (varies by each control). I smoothed the response by 1/6 octave. The red line is where I normally set the woofers (which cross at 200Hz) and the green line is the "flat" level setting. I hope I set it up correctly...

The mic tripod was sitting on some stacked cushions to get it closer to ear level and pointed straight ahead about 10 degrees above vertical. Compared to my old house, the bass response in particular sounds much better on the wood floor now than concrete (tactile feel) and I think the room itself behaves better in terms of sound. It's not only varied in shape behind me but there's a long hallway that seems to act like a bass trap and evens out the response over much of the range (wasn't that flat sounding at the old house). Using an SPL meter with test tones, I thought it measured pretty darn flat from 80Hz down to 30Hz and the graph seems to confirm that (couple of dips 90-125 and 165-180, but with the boosted bass it actually creates a nice averaged "drop" down to even flat with the ribbons, which is what i thought it sounded like and the graph seems to confirm save some minor bumps.

The green curve is about +/- 5dB averaged room response if I read it correctly. Obviously, the bass adjusted curve goes quite a bit higher below 200Hz. The speakers are rated 27Hz-20kHz +/- 3dB. They're also now 24 years old... (I've had to tighten up the ribbons three times) so might not be quite what they were new, but still seem to sound pretty darn good to my ears. They're also dipoles and Sonic Holography is engaged (which I tried on/off and it actually improved the response smoothness a little bit in some areas so I left it on).

Now the question is whether a Mini-DSP with DIRAC would/could make a significant improvement with the system. Is it worthwhile?

 

[William Lemmerhirt]

First thing is to go to limits tab and set for 45 low and 105 high. Your looking for a 60db window. That will show you more precisely the response. The 1/6 is fine as that’s pretty close to human hearing. The taller window hides a lot. Looks like almost 15db bass boost?

 

[VonMagnum]

First thing is to go to limits tab and set for 45 low and 105 high. Your looking for a 60db window. That will show you more precisely the response. The 1/6 is fine as that’s pretty close to human hearing. The taller window hides a lot. Looks like almost 15db bass boost?

I'm seeing more like 8db above 'flat' bass levels (e.g. 60Hz is at 80dB, boosted it's at 88dB and it pretty much tracks the entire way save the bottom and tops of the curve where it meets the ribbons). It's an analog knob and I didn't test it by ear (used to measure about 6dB above 'flat' with the SPL meter, but then I didn't pink noise flat either, so it might be off a dB or two from what I used to see. I don't listen at 'flat' so it doesn't much matter. It sounds better downstairs at +4 to +6dB as well to my ears.

 

[VonMagnum]

OK, here's a tighter graph for you. "Flat" is within limits of 71dB to 81dB throwing out the final dip almost right at 20kHz itself (+/-5 from 76 midpoint). Boosted is 71 to 89dB range (i.e. bass below 200Hz boosted by 8dB). That also gives a 20Hz response at 75dB (not too bad for dual 10" woofers, although obviously down from 89dB high. Bass is pretty solid at about 25Hz upward, though. They were rated 27Hz-20kHz). All the output from 200Hz up to 20kHz is from the 48" ribbon driver (2-way system). The dipole imaging makes them sound almost holographic in the room too (close your eyes and you'd almost wear someone is actually standing there singing, etc.) Kind of freaked me out when I first got them.

I'm thinking corrected, they could be pretty flat indeed over the range (and boosted have a nice even smooth slope). How much 'audible' difference that would make is the real question.

 

[VonMagnum]

I'll probably get to the home theater response tomorrow or Monday. I've got some SSD drives coming among other things, though so it depends on how smoothly things go. I plan to bump the old Macbook Pro up to a 500GB SSD from its 7200 RPM Hitachi spinner.

 

[PENG]

OK, here's a tighter graph for you. "Flat" is within limits of 71dB to 81dB throwing out the final dip almost right at 20kHz itself (+/-5 from 76 midpoint). Boosted is 71 to 89dB range (i.e. bass below 200Hz boosted by 8dB). That also gives a 20Hz response at 75dB (not too bad for dual 10" woofers, although obviously down from 89dB high. Bass is pretty solid at about 25Hz upward, though. They were rated 27Hz-20kHz). All the output from 200Hz up to 20kHz is from the 48" ribbon driver (2-way system). The dipole imaging makes them sound almost holographic in the room too (close your eyes and you'd almost wear someone is actually standing there singing, etc.) Kind of freaked me out when I first got them.

I'm thinking corrected, they could be pretty flat indeed over the range (and boosted have a nice even smooth slope). How much 'audible' difference that would make is the real question.

View attachment 29502

It looks decent, but I would like to see 1/12 or even 1/24 smoothing. In my experience, no matter how hard I tried the manual method using spl meter, tones, crawling, adjusting phase, PEQs etc., it could not get anywhere closed to a 30 minute session using Audyssey, and presumable Dirac or AARC could do a slightly better job, on paper anyway.

 

[VonMagnum]

I've been reading up more on the different graphs, measurements, theory, etc. at work on some down time and even the makers of MiniDSP don't recommend using less than 1/6 octave smoothing above 250Hz or so due to your ear's inability to hear it. I seem to recall it getting some pretty up/down variation with no smoothing, but then these are dipolars and may excite the room a bit more than typical with a delay because if it.

I'll try some settings and graphs when I get home, however. I'm more interested in decay times for room modes, etc. than overall frequency response at this point. I also want to see at least the mains on my home theater plotted and compare against the Audyssey predictions, no Audyssey, etc.

 

[PENG]

I've been reading up more on the different graphs, measurements, theory, etc. at work on some down time and even the makers of MiniDSP don't recommend using less than 1/6 octave smoothing above 250Hz or so due to your ear's inability to hear it. I seem to recall it getting some pretty up/down variation with no smoothing, but then these are dipolars and may excite the room a bit more than typical with a delay because if it.

I'll try some settings and graphs when I get home, however. I'm more interested in decay times for room modes, etc. than overall frequency response at this point. I also want to see at least the mains on my home theater plotted and compare against the Audyssey predictions, no Audyssey, etc.

Higher resolution is needed for comparison purposes, not that you hear the difference. So if you are not interested in comparing the effects of, say for slightly different toe-in angles (just a example), then 1/6 smoothing is fine.

 

[VonMagnum]

I don't mind posting the graph when I get home regardless. I was looking to find out whether a mini-dsp is worth buying for this system. I need to see if there's enough that it can improve, but I suppose there's no real way to be sure if it will make a substantial audible difference without trying it out. If nothing else, it's confirmed my spl measurements were more or less accurate on bass.

Given it's the main living room, I'm not going to put bass traps in the corners, for instance. Toe-in could be adjusted and distance from the front wall to some limited extent, but not by much with the piano there, etc.