Home Theater Forum and Systems banner

The best Audyssey test... single mic test

35K views 166 replies 17 participants last post by  Wayne A. Pflughaupt 
#1 ·
So far...

a single test... with me at the MLP... with the audyssey mic placed on top my head... (yes, like a stupid little horn coming out the top of my head)...

has produced the best results for me. I find the results to be perfect. It's when you add more than that single test I hate the results.

Of course... this is no sub. I did this for music since in direct mode the SVS Ultras are too much bass. 1/3 smoothing for ease of seeing the "curve" change.
 

Attachments

See less See more
1
#107 · (Edited)
Almost ready to formally evaluate a couple different Dirac calibrations of my room. Both use Wayne's Triangular Mic Pattern as shown below. During the informal evaluation period with just casual listening/viewing, the results are magnificent! But I'm getting ahead of myself. At this point I'm unsure of whether the performance leap is solely due to the mic pattern, or if it's also attributable to having moved my mains back to their SS&I sweet spots. It's been tough not to just forge ahead with glowing feedback, so I've lived with the system for a week in its current configuration and state of calibration to become acclimated. Finally, I hope to take the "winner" of this comparison and create a third calibration with tighter spacing. I anticipate the difference (if any) between the first two trials will be minimal; yet I may need to run my previous default cal of 1 @LP, then 4hi/4lo to re-establish a familiar reference and verify the improvement isn't just due to relocating speakers.

Again, these comparisons are necessarily subjective, but I've tried to level the playing field as much as possible by using the same customized house curve across all trials, and by using the same playback volume for each demo track. No, they weren't level-matched. Aside from not having the requisite gear, the comparison timetable is much too long to worry about levels (we're talking more than a week)! I'll be giving my general impressions of Dirac calibration results for each mic pattern, and specific examples taken from my evaluation notes. None of this is carved in stone--I'm open to suggestion. There may be some objection but as it stands, I don't plan to smooth the dip in the crossover region at this time, because I've finished the 1st round listening sessions. Post-Dirac smoothing of FR through the crossover region would give the next trial(s) an unfair subjective advantage. IMO, the dip shouldn't interfere with evaluating SS&I. Please let me know if you disagree.

Wayne's Triangular Mic Pattern (Dirac Trial #3)
Text Line Font Triangle Slope



Lou's Modified Triangle (Dirac Trials#4)
Text Line Font Triangle Slope




Corrected FR (Dirac Trial #3 vs #4)
Text Green White Blue Line




Waterfall: Sub+Mains (Dirac Trial #4 Before Dirac)
Text Plot Slope Electric blue Pattern




Waterfall: Sub+Mains (Dirac Trial #4 After Dirac)
Green Leaf Grass Ecoregion Plant
 
#108 ·
With this conversation about EQ I happened to see this video on another site and although it is an hour and change long I found it interesting and thought it might be somewhat relevant to the conversation. It is obviously his opinion but I felt that he touched upon some very good points.

 
#109 · (Edited)
Just setting the stage :D for my future comparisons of triangular mic pattern...

My General Impressions of the Triangular Mic Pattern:
  • 2ch - Expanded soundstage in all three dimensions
  • 2ch - Increased image focus
  • 2ch - Absence of midrange glare or lower treble hardness
  • 5.1ch - Major improvements in spaciousness

Reference Mic Pattern (unless otherwise specified):
  • 1 location @LP
  • 4 locations in rectangular pattern below seat back
  • 4 locations in same rectangular pattern above seat back

Reference Material (Eagles - "Hotel California" from Hell Freezes Over):
  • 00:15-00:16 ... Someone moans "yeah" at chest level in R channel very deep down in mix (almost inaudible)
  • 00:17-00:30 ... Lead guitar with increasingly fast picking should not move or grow/shrink
  • 00:23-00:33 ... Crowd Noise #1 should have 3 distinct whistles and 2 female shouts from different locations in arena
  • 00:32-00:50 ... Percussion toward front wall, convincingly-sized drum kit, wood block struck in R Ch should resonate toward back left of hall
  • 00:35:00:40 ... Crowd Noise #2 should have 3-4 distinct voices of varying intensities among other chatter.

Edit:
I'd like to explain my choice of reference material, because it's not a selection one would think to be at the top of anyone's list. It's not particularly dynamic, doesn't have an expansive soundstage, and is of generally mediocre quality. But that's exactly the point. When I voice my system for SS&I, using source material of stellar SQ only results in those types of recordings sounding their best. Poor-to-mediocre recordings suffer. Conversely, using source material of poor quality severely compromises the really good stuff. Some of it boils down to the type of music you listen to most. The rest of it boils down to listening preferences in order of importance: detail, spaciousness, etc. I want most of my collection to sound good so I aim toward warm and comforting, yet still detailed and revealing. I'm extremely familiar with the first minute or two of this track, and have used it for demo work so often I find I cannot listen to casually any longer, as it forces me into analytical mode. In any case, I wanted to use the same main piece for evaluating the Triangle Method as I did to tune the system. There were others involved as well, but this was my go-to reference.
 
#110 ·
Just setting the stage :D for my future comparisons of triangular mic pattern...

My General Impressions of the Triangular Mic Pattern:
  • 2ch - Expanded soundstage in all three dimensions
  • 2ch - Increased image focus
  • 2ch - Absence of midrange glare or lower treble hardness
  • 5.1ch - Major improvements in spaciousness

Reference Mic Pattern (unless otherwise specified):
  • 1 location @LP
  • 4 locations in rectangular pattern below seat back
  • 4 locations in same rectangular pattern above seat back

Reference Material (Eagles - "Hotel California" from Hell Freezes Over):
  • 00:15-00:16 ... Someone moans "yeah" at chest level in R channel very deep down in mix (almost inaudible)
  • 00:17-00:30 ... Lead guitar with increasingly fast picking should not move or grow/shrink
  • 00:23-00:33 ... Crowd Noise #1 should have 3 distinct whistles and 2 female shouts from different locations in arena
  • 00:32-00:50 ... Percussion toward front wall, convincingly-sized drum kit, wood block struck in R Ch should resonate toward back left of hall
  • 00:35:00:40 ... Crowd Noise #2 should have 3-4 distinct voices of varying intensities among other chatter.
Thanks I'll try it out.
 
#111 ·
Adding calibrations for Trial 1 & 2 mic patterns led to some interesting impressions I just had to share! And just to recap:
The Trial 1 pattern is 1st location @LP, then 4 random locations below the seatback, and 4 random locations above it.
The Trial 2 pattern is triangular with 1 location @LP, then 4 along hypotenuse, and 4 above/below/left/right of LP.
The Trial 3 pattern is triangular with 3 locations @LP, then 6 along hypotenuse.
The Trial 4 pattern is triangular with 3 location @LP, then 2 along hypotenuse, and 2 to the left and right of LP.

Without more detailed and rigorous study it's hard to draw solid conclusions. But it seems that more mic positions along the hypotenuse enhance imaging at the expense of soundstage. The difference is not night and day, but it's there. Images with very specific size and place within the stage, morph into slightly larger blobs of sound when more mic locations are located off the hypotenuse' axis.

This discovery of a relationship between hypotenuse patterns and SS&I quality developed into a preference. Granted it's a preference over a slight difference, but it's enough of a difference to me. I prefer the dedicated hypotenuse locations for music listening, but for movies I resort to calibrations with randomly added locations. The additional randomizations have another positive effect: they result in tight cohesion of sound transitions between front and surround speaker pairs. That came as a surprise to me, as I didn't even know what was wrong until I heard it reproduced correctly! The purely randomized pattern (Trial #1) had left a hole-in-the-middle effect that Dirac was able to fix using the both the purely hypotenuse pattern (Trial #3), and the randomized hypotenuse pattern (Trials #2 & #4).

A couple more observations:
  1. I found the purely hypotenuse pattern (Trial 3) to be more sensitive to head movement than the randomized hypotenuse patterns (Trials 2 & 4). For instance, sitting up straight to raise one's head above the seatback should drastically change SQ. When sitting up in this manner, the SQ improved for Trials 1 and 2 (higher number of random positions), but degraded for Trials 3 & 4.
  2. I also discovered a strange phenomenon in Trials 2 & 4 related to spaciousness in movie soundtracks. Scenes with high reverb such as in an empty alley or church could sound surreal and even disorienting with my head against the seatback. The effect grew worse with increasing proximity. No such phenomenon was encountered with purely hypotenuse patterns.
 
#120 ·
The little kid in me like to play with his new TONKA in his backyard ! ( I am sure thta I am not alone )
I watched the video suggested by Tcarcio. Specially that part where he talk about 2 mics on a dummy head.
If I remember well, Talley tried to EQ (Audyssey ?) with the mic on his head (forgive me if it is not the case and I misunderstood your posts about it).
Anyway, it gives me the following idea that I will try soon hopefully: 3 measures at LCP + 3 along the hypothenuse then, I replace my trepod by yours truly himself sitting at the MLP and 1 more measurement with the mic at my left ear (after putting ear plus of course) + 1 more at the right ear.
Stay tuned for the (subjective) results. !
 
#121 ·
Wow, talk about covering a lot of ground. Look out, audio geeks at work / play!

Setting up some new speakers for review and decided to move my sub to dead center between the mains, so I am re-running calibrations for nanoAVR-DL for my 5.1 setup and also for the "review speaker" 5.1 setup. Also calibrating Dirac Live full version (upcoming review) on my media server for my 5.1 setup. Might get all that done tonight, so hope to have some data to throw into the mix.

Did one nanoAVR-DL calibration that came out sounding weird, but I'm pretty sure what the problem was, will re-do and report on it later.

One thing I have done, starting with the nanoAVR-DL review awhile back, for the LP measurement(s), to eliminate chair back reflections from that oh-so-critical first measurement, I have a 1 x 8 board about 4 ft long standing up about where my spinal column would be situated but spaced forward 6 inches with a blanket wadded behind it, slanted back slightly, the mic hangs straight down JUST touching the board at the LP height and distance from speakers. The resulting measurement is very flat and very clean of chair reflections. It is like saying, "Dirac, don't worry about my chair reflections on this first measurement, just take care of the room, please." The board is removed for the non-LP measurements, which are all significantly further from the chair back. The one time I did an A-B trial with this technique for first 3 measurements vs no board, I could easily tell the imaging was significantly tighter from the trial WITH board. It was a difference between very good vs very VERY good. But it is a bit of trouble for the gain, so might not be worth it to you, just mentioning it since there is so much experimentation going on. I know Flavio will want to have me flayed, drawn, quartered, and skinned alive for suggesting it, but there you go.

Real mixed feelings about taking measurements while seated, as I reported in the sticky on mic patterns. I got so much variation it seemed pretty fruitless, but ya'll might come up with a solution to that. The board was my closest solution, simulating mic next to ear (board = skull minus all the messy ear flesh), and centered & facing front of room and all, kind of a mono cyclopse Van Gogh approach (MCVG method???).
 
#122 ·
I have done several Dirac Live calibrations over the last couple of days and will add the following observations to the discussion.

  • I did not try the right triangle pattern for Dirac Live, but I did one pattern that was smaller, the furthest points from LPC being 1.5 feet away and the closest being about 6 inches away. The resulting sound stage tended to come apart at the seams with very little shifting of the head position.
  • The tip of the mic hanging against a hard surface such as a board, the MCVG (Mono Cyclops Van Gogh) method:rolleyesno:, really seems to give Dirac Live a kick start in the direction of giving best possible soundstage and imaging (SSI). Here is a photo of my setup for doing that. I have done this a number of times, and it has been the foundation for some very satisfying calibrations. One nice thing is that you can do it once, save the project at that point, and recall it to add other mic pattern variations to as desired.
  • In reference to the preceding point, I ran one test with the first three measurements taken at the same initial point with the tip of the mic touching the board plus the other 5 random, and another with only the first one measurement taken that way and the other 8 random, random also meaning with no attempt to repeat any kind of pattern between the two tests. Upon comparing those two calibrations, it was extremely difficult to hear any difference between the two. Only with one track was I able to discern a very slight SS&I edge in favor of the 3 initial + 5 random (3+5) approach. It was an extremely subtle difference, and on a different day I might call it the other way.
  • All my experience indicates that a more broadly spaced, randomized calibrate calibration pattern is better for Dirac Live - just like they say to do. The closest point or points might be a foot away from LPC and the farthest might be as far as three feet away for a single-seat two-channel setup. That might put a measurement point much closer to one of the front mains than the other, which seems counterintuitive, but the Dirac Live algorithm seems to invite this kind of rich variety of data to work with and knows what part of that data is useful and what part is not.

Blanket rolled up behind board to space it from the back of the chair and to absorb/prevent resonances in the board.
Skin Joint Leg Footwear Hand


Board over blanket, hanging mic with tip of mic just touching the board at the target point, Listening Position Center (LPC). Plumb string behind mic aids position repeatability (non-WAF-approved permanent mark on ceiling above not shown).
Wood Plywood Hardwood Wood stain


Close-up.
Black Black-and-white Monochrome Monochrome photography Tree


String removed for taking of measurement.
Wood Plywood Hardwood Floor
 
#125 ·
Also, out of curiosity, I recently ran a calibration with all mic positions except those with the board pointing at the ceiling vs. a calibration with all mic positions hanging, pointing at the floor/chair. The UMIK-1 has enough of a sensitivity drop at HF at 180° that it made quite a difference. Mics pointed at floor picks up less HF info, Dirac compensation made the filter set brighter by probably 3 to 4 dB at HF. With a small-diameter mic - Beyerdynamic MM1 - the difference is very small.
 
#126 ·
Wayne what are your suggestions when using XT32 in this situation...

Four seats across the room, the centre of the room is between the two inner of the four seats.
My seat is on of those two inner seats.
I take the 1st measurement, one only (call it MLP) between the two inner seats of the four, over arm rest. This gets all the distances and delays correct.
I have always then clustered four measurements within one foot around my seat, one of the two inner seats.
Then the rest within the other four seats....

Is it better to cluster those four measurements round MY seat for imaging or to actually take them dead centre of the room, where the 1st one is done?

http://www.hometheatershack.com/forums/attachment.php?attachmentid=97377&stc=1&d=1437260445
 

Attachments

#127 ·
Beautiful room, Murray.

I would keep the first measurement and all "close clustering" of mic positions together in one group, always. Splitting them only weakens the desired effect.
  • If your room is larger, as yours appears to be, or if your speakers are horn-loaded tweeters, like Klipsch or the PSA MTM-210, keep that cluster over the center armrest, as you will experience little to no shift in the two-channel center image sitting in either of the two front-center seats.
  • If the room is smaller, putting you closer to your mains, and with other tweeter types - OIW, you are noticing a larger shift of two-channel center image sitting in one of the two front-center seats - then try clustering on the center of one of those two seats. It will shift the two-channel center image to straight ahead of that seat. The soundstage will be slightly "smushed" to one side, but probably not enough to even notice it.
  • If you only do movies there, keep the mic position clustering over the center armrest, the center channel speaker will keep center images properly placed and the overall balance will stay centered right for the room.
 
#128 ·
Thanks for that Wayne.
My room is 4M wide by 6M deep, the middle row (MLP) is the middle of the three rows of seating.

I will re do my cluster in the centre of the room over the arm rest, measure and see if there is a difference.

TOP IMAGE
This is what I get with XT32 on and the cluster over just one of the middle seats. Left, right and sub, crossovers all 80Hz

LOWER IMAGE
This is what I get with XT32 ON and if I change just the L&R to 120Hz and the centre channel to 80Hz, a very different result.
Both measurements are taken from the same spot in one of the centre seats. Do you think its better to keep the crossovers set this way considering the result is much smoother?

http://www.hometheatershack.com/forums/attachment.php?attachmentid=97393&stc=1&d=1437276658

http://www.hometheatershack.com/forums/attachment.php?attachmentid=97401&stc=1&d=1437276791
 

Attachments

#140 ·
haven't gotten my Dirac Full license yet but I tried out the SE. Pleased with the results on SS&I out of the box with a single mic position. Although the bass seemed to be lacking a bit.

I have so much work still to do... need to hang all the treatment I've built and build more.
 
#141 ·
RapalloAV: Interesting. Of course the first measurement looks the best. The second isn't bad. Wouldn't it be great to get that first curve at the actual LP!!

I suggest you try this: Take the first five measurements over the seat where you want the sound to be the best. The first measurement right where the center of your head would be, not over the armrest, but centered over that seat, and the next four measurements clustered around that first measurement, then the final measurements where you normally would.

But pay attention also to the way it sounds. How precise is the image clarity? How well defined is the soundstage? How does that compare after the suggested setup compared to the way it sounds now? Good measurements are important, but I will take an imperfect FR curve with great soundstage and imaging any day.

That is my suggestion. But if you are happy with the way it sounds now, what you are getting really looks pretty good.
 
#142 ·
Wayne, the first cluster of four mic positions were done from the seat I sit in, not over the arm rest. These are the measurements you have just looked at.


Ive just finished doing another 1st cluster set over the arm rest in the centre of the room which Im about to post. Then you can tell me which set you regard as looking the best....

Sorry I just re read your post. I now see you want me to try the 1st measurement from the MLP (centre seat) not over the arm rest. I will try that and report back.
 
#146 ·
Whats funny is I don't even look at the FR response. I run the test and listen to it. My ears cannot hear a graph.

I think looking at the graph leads to too much errors really when the reality is you don't hear it.
 
#147 ·
There is a lot of truth in that, Talley. I think it is good to listen before looking at measurements for that very reason. Then there are times I will see something in the measurement that I missed and will pay closer attention to it and it might even sway my judgment a bit. Other times I will see the measurement and think, Eh, it sounds just fine to me.
 
#151 ·
These were the measurements I did last night with using one of the middle seats as the 1st cluster of mic positions rather than over the middle arm rest. This is really the true MLP.

To me the measurements look similar to when I took them over the arm rest. Imaging sounds exactly the same to me, no improvement, for some unknown reason the highs seem too bright to my ears, why I wonder???

Again the green is the left speaker, dotted right. EQ on.
1st pix over the middle seat MLP
2an pix over the end seat
http://www.hometheatershack.com/forums/attachment.php?attachmentid=97625&stc=1&d=1437428387


http://www.hometheatershack.com/forums/attachment.php?attachmentid=97633&stc=1&d=1437428387
 

Attachments

#155 ·
Ok Wayne Ive just put all the four subs facing forward back to they used to be. After listening to all my test discs I just couldn't get any bass from the subs when facing the side walls. They just really lacked any impact or punch!!!!! I kept turning them up to try and get back what I used to have, I was up to +6db and still I had WAY less than when the subs face forward and set on zero db. What a disappointment. Why would they be so bad this way???

I turned them to the front and re did XT32 and bingo, all the bass was big again with power and punch and only on zero. This might be where I now have to live with things apart from ripping the whole place to pieces or buying a new house and rebuilding a new HT without any mistakes....:D

Oh well, thanks for all your help and input, you have been great!

If there is anything else you can think of please yell out! :T
 
#156 ·
Indeed, there was not the improvement I had hoped for why turning the subs. In fact, the narrower notches turned into a broad depression, not an improvement at all. I agree with your decision to leave them in their original position.

Trying to do anything further would probably involve having individual time delay controls for each of the four subwoofers, more detailed measurement capability with impulse response and precise time delay measurement possible, all with quite a learning curve. And even then, with very limited ability to place the subs optimally, there is still no certainty in achieving a big improvement. I wish I could say otherwise.

You have a wonderful room and excellent sound, and it might be that you have reached the point of diminishing returns for your efforts that makes this a perfect place to call it good.
 
#157 ·
Thank you so much Wayne, you have been an excellent help to me and your words are always encouraging. I wish I could go further and even try all the time delays you hinted upon, but it looks like that is way beyond my knowledge. I wish there was someone in NZ that had the knowledge to help me, but we very much lack people here with knowledge like you.

I have a couple of "other" small tests Im going to try tonight and will get back to you if anything is promising....
 
#158 ·
Aaaah ! There are a lot of passionate people here. I had forgotten my own teen age passion for music 2.0 system for so long (I once had the best B&W speakers in town).
Now that I came back to my first loves (but 5.1 instead of 2.0) , I must keep in my mind this french say that we have "Le mieux est l'ennemi du bien" Free translation: the best is the ennemy of better...
 
#159 · (Edited)
Wayne you might like to see the improvements I have since you last heard from me. I have been adding extra bass traps to my under stage area where the subs are. Every day I have added more and more 4" dense rockwool behind the subs on the front wall and the sides. Behind the subs its full to the thickness on 16", the two sides have 12". I have also fitted another trap behind the back row of seats to the thickness of 4", this area has already a number of bass traps but its now full, there is no more room.

As every day Ive added rockwool Ive taken readings, each day Ive seen the peeks in the bass come down further and further, its been amazing. Each day when I listen its obvious the improvement. This has given me the incentive each day to go out and buy more rockwool and add as much to the boundary corners I can. Ive pretty much finished, there is one last two corners in behind the screen only on the side corner walls, there are bass traps on the main front wall with membrane over. I might do this last bit to 4" thick.

Take a look and see what you think.

1st pix shows the rockwool stuffing started
2nd pix the measurements I took on Sunday after two days of stuffing. L&R + subs EQ on
3rd pix Tuesday after two more days of stuffing. L&R + subs with EQ ON. (top dotted EQ OFF)
4th pix Sunday Room An L&R + subs 12 ppo smoothing EQ ON.
5th pix Tuesday Room An L&R + subs 12 ppo smoothing GREEN EQ OFF - BLUE YELLOW EQ ON (you can see here the bass is smoother than sundays)
6th pix Tuesday Room An L&R + subs 3 ppo smoothing GREEN EQ OFF - BLUE YELLOW EQ ON.

Overall this does show how much bass trapping reduces peeks.

http://www.hometheatershack.com/forums/attachment.php?attachmentid=98233&stc=1&d=1438062584
http://www.hometheatershack.com/forums/attachment.php?attachmentid=98241&stc=1&d=1438062584
http://www.hometheatershack.com/forums/attachment.php?attachmentid=98249&stc=1&d=1438062584
http://www.hometheatershack.com/forums/attachment.php?attachmentid=98257&stc=1&d=1438062584
http://www.hometheatershack.com/forums/attachment.php?attachmentid=98265&stc=1&d=1438062584
http://www.hometheatershack.com/forums/attachment.php?attachmentid=98273&stc=1&d=1438062584
 

Attachments

This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top