Simplified REW Setup and Use (USB Mic & HDMI Connection) Including Measurement Techniques and How... - Page 59

I'm becoming sold on that...once I get the center channel resolved I'll have some graphs to finalize. Although now that I have a better mic with the UMM-6 and the 90 degrees cal file, I may double check the 'level matching' within the sub and non-sub sections of my powered fronts+center, and if need be, do a fresh Aud Pro run without my existing Target Curve edits.

I'm not so sure about that Stuart. Once you have an ASIO4ALL 8 channel setup within REW your AVR's display should show "Multi CH IN", and then in REW you can select which channel to measure. Please confirm this. And then if you try to engage PL II nothing will happen since the AVR is already defaulted to "Multi CH IN". BTW, PL II should only work with a 2 CH stereo input, since it's a "listening mode" made to matrix a 2 CH stereo signal into "fake" 5.1. Nonetheless, IMHO it really does a pretty good job in that "fake" department.

 

In the particular analysis I was conducting (for Markus, to address an issue originally brought up by IgorZep in the 818 thread), I specifically omitted the subs.  I suspect the sub measurements would demonstrate the same effect that Audyssey is having.  Perhaps a homework assignment for Stuart?  

I'll add it to the list...so how do I do Pre Out with HDMI? Or is this strictly an Aux In method?

 

The procedure is described starting on page 67 of the Guide.  

So I see: I didn't realize your example was HDMI until I reread it just now. One more inch down the rabbit hole for me. But a worthwhile one.

Hi sdrucker, I have a Denon and it display Multi Ch In and it does what Kevin posted. What I notice is when REW opens it would switch to Dolby PLII (in about 2 second or so) but switches back to Multi Ch In when you press measure button.

Pic shown with Audyssey "on" (red dot)

Got it, thanks! The lesson is that for HDMI, leave the AVR on Multi Channel In and don't switch to stereo for any REW measurements. I took the Guide too literally.

Off to play....

Just testing dropbox...

https://www.dropbox.com/s/zdrbx0t4r9087ot/left%2Bcen%2Bright%2Bsub%20position%20B.mdat?m

Jerry please take a look these measurements. I measured left+cen+right+sub. Then averaged. Not sure this is correct, though. When I test left speaker as SMALL, am I not also testing the sub? IOW is it correct to measure the sub separately?

Also, if I want to make measurements at another spot, do open a new file, or continue with same file?

Edit: nevermind, found my answer looking at other posts.
Edited by Pres2play - 3/3/13 at 11:24am

How did you get three curves onto one graph? Did you just measure at different volume levels?

Audyssey did an amazing job there.

 

To improve readability, I simply added an offset to the first and third measurements, +10 to the first, and -10 to the second.  Open the Controls icon, select which measurement you want to apply the offset to, and enter the offset value.

Learned something new today, time to relax and enjoy the day off.
thanx

Is this normal to have a higher dB reading with Audyssey turned on?

pic with LCR with Sub and Audyssey on

pic with LCR with Sub on and Audyssey off

Something seems peculiar with your measurements with Audyssey engaged.  The lower frequencies are getting a serious boost, which will likely cause the bass to sound emphasized, even boomy.  Did you by any chance have DEQ turned on during the measurements, or did you manually increase the sub trim levels after the calibration?  Audyssey should be improving the bass response, not making it worse.

 

And the deep dip in the left channel at 60-70Hz with Audyssey on seems strange as well.

Got it. Thanks.

Jerry, ah yes DEQ was turned on with Audyssey. When compare graph with Audyssey on, should I turn off DEQ?

I didn't increase the sub trim levels after the calibration (left it as is -5). It does sound kind of boomy and don't know why about the deep dip.

Here is a new graph, with Audyssey and DEQ on, then Audyssey on and DEQ off, and last no Audyssey no DEQ.

It seem more flat with Audyssey and DEQ off .

 

You should measure without DEQ while you are troubleshooting your room and FR. DEQ is very predictable in what it does, so there's no purpose served ny measuring with it on anyway.

 

What crossovers are you running for LCR?  80Hz?  If so, you have a problem with the splice - you can influence that by using the Sub Distance Tweak.

 

By coincidence, I recently made a PDF document describing the Sub Distance Tweak in detail. You can download it from this FAQ answer:

 

d)1.   Post-calibration suggestions (optional).

Well, always wanting to run before I can walk, or REW before I can walk in this case, I have been looking at this:

 

 

 

Now this looks fascinating. I have had a play around with it and have got the general idea, but that is all. What interests me is that I have recently been moving towards the idea of using Parametric EQ to attempt to solve some niggles that Audyssey Pro can't overcome on its own. What is sensationally interesting in the EQ section of REW is the way you can use it to create filters to combat problems and then to upload those filters automatically to something like the Behringer FBQ2496. All more or less automatically. 

 

Problem is, I have reached the limit of my current understanding, despite having read the relevant section of the REW Help a couple of times. The Help is very good at explaining the 'how' but not so good at explaining the 'why'. I shall persevere.

 

Keith, have you posted your latest graphs?  I would be interested in seeing what you consider as "niggles".  And are you sure that PEQ will resolve them?  

 

Can we see a L+R+Subs, and a Waterfall?

 

As Keith says, never engage DEQ during measurements.  To the left of the dip, the corrected response looks better than the uncorrected response.  Also agreeing with Keith, the big dip looks to be in the neighborhood of the crossover.  Have you read the instructions for the sub tweak?  Do you have more than one sub?

 

if you were to smooth out that dip, the Audyssey response curve would look reasonably good, and better than the in-corrected response curve.

Did the Auto Setup early this morning and now that you mention the crossover, i didn't change it to 80Hz. Was left alone as 60Hz fronts, 100Hz center and 60Hz surround. Will read you Sub Distance Tweak, thanks.


Running with one sub in the right corner. Most likely didn't drink enough coffee before running setup . Will perform some adjustment and redo REW later.

OK, guys, after enough "delay", we have charts!

Blueish & Purpleish lines are post-Audyssey Pro (XT32, DEQ is off); others are with no Audyssey engaged. All measurements are from MLP, i.e. the center of our sofa, slightly toward the front edge of the seat, at ear level:

Left/right mains (Mythos ST) and subs (two ULS-15), with 1/6 smoothing after measurement:



Center (CS-8080HD) and subs, with 1/6th smoothing after measurement:



Subs only, no smoothing:



Currently the L/R are crossed at 80 Hz, and the center at 100 Hz.

Somehow the FR is turning out better than I'd have expected after the Audyssey run, although I do want to see if I can improve the distance tweak on the main/sub splice, and maybe hammer out the 300 Hz and 700 Hz peaks for the L/R and center. And it seems there's a fair amount of correction going on through at least the 300 Hz area of my mains, as well as some minor adjustment for the mid-range comp and at the higher freqs (the higher end above 5 kHz is something I addded in the Curve Editor to provide more "air" on some instruments, BTW).

Also, what I find interesting is that these charts look closer to what I had at 1/6th smoothing on my Pro graphs than what I used to get from OmniMic. Given that the UMM-6 mic is individually calibrated and there's a formal 90 degree cal file, I'm even more sold that REW is the most appropriate tool to conduct room assessments than from AJ's charts yesterday. Comparing REW with the UMM mic to Audyssey Pro's individually calibrated mic measurements strikes me as much more apples to apples than OM.

However, that 70 Hzish needle sharp null for the subs bugs me, and may be a symptom of the deeper issues beyond the (mostly) deceptively reasonable FR charts. Or maybe I'm taking Jason too seriously today.....

Why? Here's a waterfall, based on L/R and subs, run at 1/48th smoothing (default), 600 ms delay, and (for now) a 40 to 90 db range:



Lots of stuff decaying to 40 db (or worse), and resonance city? And my measured noise floor is closer to 55 db than 40, which means that this is even "worse".

By the way, this is in a carpeted living room (23 x 17 x 8.5),with an open floor plan to the dining room in a generally L-shaped space, and I don't have any current formal room treatment.

Thoughts?
Edited by sdrucker - 3/3/13 at 11:41am

Keith, if you don't mind, can you post the above chart without the EQ chart in front, so I can see how the individual measurements including averges should look.

 

I haven't posted any graphs for a week or so Jerry and I will tell you why. I had a helluva week last week and came close to throwing in the towel. It started with the new treatments which arrived on Tuesday and were duly installed where my magic string dictated they would do best. After installation, just for the hell of it, I ran the usual set of REW measurements and they were noticeably worse than before. I put this down to not having yet run Audyssey, so I did, and guess what - they were worse still!

 

Here, for example, is my worst channel of the lot - the centre channel, before adding the new treatments (the RF and LF + subs are measuring much better):

 

 

 

 

And here it is after adding the new treatments and after running Audyssey (overlaid on the above):

 

 

 

WTF?? What has happened around the XO region? I have no explanation for this at all. The graph is actually better with the treatments added and without re-running Audyssey! But it is still worse than the first one above.

 

So I have spent an entire week troubleshooting, First thing I did was remove all the new treatments and then load the Audyssey cal that I have been using happily since last November. That took me back to square one. I then added the new treatments one at a time, measuring each time as I went. Not a deal of difference between them (the latest treatments are 'icing the cake' and I wasn't expecting dramatic differences from them - the big differences had already been earned with the original treatments some time back).

 

Next I made a new Audyssey Pro calibration, without the new treatments and it was not as good as the original above. But I then referred again to my November notes and saw that I had used a different mic pattern in November - 7 positions clustered around the MLP and two around the other seat. I didn’t have the strength to do yet another Pro calibration (you know how long it takes!) so I reinstalled the treatments and measured again, with the Audyssey cal just made (with all 9 mic positions around the MLP).

 

The next day I redid the Pro cal yet again, using the mic pattern from November, and, surprisingly, the result was better - not as good as the original but a lot better than the disaster above. By now I was half crazy. 

 

I then spent one or two days tweaking things like the sub distance, experimenting with Pgm1 and Pgm 2 of my Submersives, different crossovers and so on. Finally, I managed to get an improvement, although still not as good as the original at the top of the this post.

 

Here is the best I could get, overlaid on the original:

 

 

 

It's not bad - and much better than the disastrous one - but still not as good as the original. And what explains that boost to the bass?  I have NFI! I have triple checked everything - DEQ is off, the subs are in the same Pgm mode etc. (Please note that that dip centred on about 450 Hz is a 'feature' of my room and cannot be removed with anything, other than repositioning subs/speakers/MLP. which is impossible (believe me, I have tried). 

 

Yesterday some semblance of sanity returned as I realised that the system sounds amazingly good. Imaging is palpable. Dialogue is so good the actors could be in my room with me - I can hear them breathing and sometimes moving their tongue inside their mouth. Bass is extended and powerful but without being overbearing. Ringing is minimal and the best I can do with my room and placement limitations. It sounds fabulous. So I am wondering how OCD I have become. I am looking at things on graphs that apparently cannot be heard in real life. Remember Bill Fitzmaurices 'Graphitis Nervosa' and his remark about barely being able to hear the initial impulse at 20Hz, let alone the reverberations of it? Look at the difference between the two plots above - the blue plot has kept a smile on my face since last November. The purple plot is a few dB different at 1/6th smoothing, which I am told more or less represents the way we hear things. Can I hear any difference?  No. Can I see a difference? Yes.

 

Having said all that, I am not the kind of person who gives up easily. Which is why I have been reading up on Parametric EQ and checking out the amazing EQ tab of REW - just go and have a play with it and see it flatten the plots in real time as it creates correction filters. Use the Behringer which permits for 20 filter adjustments - it is nothing short of amazing. I am even starting to wonder if, assuming I can master it, Parametric EQ might give me a better result even than (hushed tones) Audyssey?  We have our rooms treated, so a lot of Audyssey's magic work in the time domain is not really needed for us...

 

With a Behringer FBQ2496 I could create, with REW's magic, 20 different filters, adjusting them for response, loudness and Q. Even my ignorant playing with the EQ tab reveals the potential here.

 

So, I am still not satisfied that I cannot do better. I still have unanswered questions (like why is there a bass boost in the latest graphs, why did Audyssey apparently make things worse not better?). I will continue over the next few days to ponder it and to see if I can improve it, or at least get it back to November's status quo. But meanwhile I am trying to remember what my hobby is - it is movies - and reflecting on the amazing aural experience <oo-her> that I can get in this small, awkward little room of mine.

 

Any ideas are, of course, more than welcome!

@Stuart,

 

I think the improvement provided by the Audyssey calibration is quite good, especially on the center channel.  The dip in the sub response at 70Hz is very narrow, but seems to be contributing to the dip seen in the left and right measurements.  I don't recall--have you already tried to smooth the crossover region by tweaking the sub distances?

 

The ringing in the waterfall graphs would undoubtedly improved with treatments.  Will treatments pass the WAF?  Of course, the real test is how you think the bass sounds.

 

Nice job with the REW measurements!  

@ Keith,

 

First of all, you are not alone.  Getting absolutely consistent results from one session to another is the exception, not the norm.  There are so many variables that consistency is a challenge.  I have described before my process of using a tape measure, marking measurement points with adhesive dots, and using a plumb bob contraption hanging from the bottom of the mic to ensure my measuring points are consistent.  Many have ridiculed this procedure as the height of OCD, but it has provided a degree of consistency.

 

Having said all that, it is unlikely given my experience that the addition of treatments (GIK 244's?) would cause things to be worse.  So, I would look to something else, perhaps a slight variability in your measurment technique, etc. that is causing the graphs to be different.  I am assuming that you re-checked the distance tweak, because treatments could affect the phase relationships and require a different distance adjustment.

 

And finally, if it sounds good now, maybe the original measurements were flawed, and the current measurements are accurate.  In other words, the treatments really have made things better, and you should trust your ears.  If you start down the PEQ trail, we are going to have to schedule an intervention and pack you off to rehab! 

 

There's no way that needle sharp dip is going to be audible IMO.

 

If your noise floor is closer to 55dB, then the graph is better than you are currently seeing - redraw it with a bottom limit of the actual moise floor - 55dB - and see. Once the ringing has entered the noise floor, you're not going to be hearing it are you?  And remember what Bill Fitxmaurice said ;)

 

HST, there is way too much ringing for way too much of the spectrum there - treatments are your next port of call Stuart!  I hope your wife is easy-going like mine :)

Thanks, Jerry! I thought those FR graphs looked good overall, at 1/6th smoothing (I actually used to run 1/12 with OM). And yes, this is after the distance tweak with the mains that I did when I used OM. I may see if I can improve on it with REW. But overall, what Audyssey Pro+XT32 have accomplished is nice.

As for treatments, I can swing them if they're in corners or directly behind the subs, and in a color fabric that fits our room (maroon or red, maybe black?).

If I were interested, what's a good next step? Contacting GIK with my room info and the graphs?

@ Keith: I'm still getting used to waterfall interpretation. I'll redraw later tonight and update my post.
Edited by sdrucker - 3/3/13 at 1:30pm