Simplified REW Setup and Use (USB Mic & HDMI Connection) Including Measurement Techniques and How... - Page 18

By custom do you mean for the various mic positions (0, 90, 180 etc.) or is there something that makes the UMM-6 calibration "better" than the OM mic?

Each UMM-6 is individually calibrated, with the three calibration files you mention. AFAIK, the OM mic does not have these.

Edit: the calibration files are actually 0, 45, and 90 degrees.

^^
I sent my OM back to parts express to be calibrated to 5Hz...I would think the calibration would be at 0 degrees.

Is this only true for the v1 mics that have not been sent back for calibration? AFAIK, the v2 mics are individually calibrated for at least one position (presumably 0 as laugsbach indicated above).

Thanks for correcting me on the mic positions...I knew something was wrong when I typed it!

Wow, just found this thread; looks like I have a lot of reading to do!

Anyway, I was doing a sub crawl using the mic and took some REW measurements today. However, that was before I found this thread. So my measurements don't fit the standard (eg: waterfall graphs are only up to 75dB, and only measured 10-120Hz).

But I'll post them anyway. I understand of course that these graphs may not tell me anything because I didn't measure it properly, but with a newborn, the ability to do measurements (especially up to 100db) would be very infrequent!






Some other info:
- Graphs are with Audyssey OFF
- The noise floor is indeed ~40dB in that room, so I didn't make the waterfall plots go below 40dB
- I did the sweeps with all speakers and subwoofer combined (not like what was recommended, at least for the frequency response in this thread)
- Crossover to the sub is 80Hz
- One sub, SVS PB-12 Plus

Questions:
1) Which of the three would be the best placement for the sub? (with the caveat that I didn't generate the graphs completely properly)

2) How do you generate an ETC graph? Not sure if it was posted earlier, as I've only had time to read the first 5 pages of the thread, but any links to a post containing the info would be appreciated. Or the info itself...

3) The SVS PB-12 Plus has parametric EQ--is it better to try to nullify peaks/troughs in the frequency response curves first using parametric EQ before Audyssey, or try to do it after Audyssey on what could not be corrected adequately?

4) Does decay matter much for the frequencies below 20Hz, considering they are mostly inaudible? I would imagine that "muddying" of the frequencies which are felt rather than heard don't really matter.

5) How do you make REW play a sweep at a higher dB level to average around 100Hz across the frequency response to generate a proper waterfall graph?
Edited by beezar - 1/21/13 at 8:23pm

I don't have an opinion on which is the best sub placement based on your graphs. The answer is always more subs.

How to generate an ETC graph is the last section of the Guide.

If you are going to use the EQ in the sub, I believe it is recommended to do this before running Audyssey.

The answer to your question regarding decay below 20Hz has not been discussed yet.

To generate a test at 100dB, you simply need to calibrate the REW mic at that level. Read my previous post, along with the warning that there could be risk to your equipment. Proceed with caution.

Wow, a new MAC version. Finally. Thanks.

Does anyone else notice that it still takes about 5 minutes to load? I used to have that problem with CyberDuck (until I stopped using it). The new version seems as bad as 5.0.

But it did get me to play again with my old info. There's no way in hell I'm getting to 100dB, but here it goes with the floor at a reasonable 40dB:

 

 

I believe some of those targets you quoted are from the white paper I co-authored with Jeff Hedback. There is a ton of pretty detailed discussion about each measurement and relevant targets. You guys should read it, if you haven't already. There's a whole bunch of targets defined in there for things like frequency response, time decay, ETC and so on.

Maybe people missed it (or ignored it) earlier on. To whet your appetite here's the summary page:



And here's the decay target graph:



Please note those targets were for two channel. Am happy to brainstorm with you nice people what the modifications for HT might look like. At a high level they would involve bringing in the bass response frequency response targets quite a bit, decay, ETC, mid/high frequency response and likely T60 targets would remain as is.

Outstanding! Thanks for the link, downloading it on my phone as I type this. It'll give me some reading material till I get back to my laptop.

BTW, as might be surmised, I was unaware of your white paper. Those targets were simply from many discussions with lots of folks including advice from Ethan Winer.

To the other readers of the thread, I also forgot to mention that there is a range for those decay times. Obviously, the higher the initial signal, the longer it will take to decay to ~40db. Those decay goals are for measurements at about 75-85db, which are reasonable levels for the average person.

I don't know many folks measuring at 100db. The ONLY reason I measured ONCE at 105db was as the last step in measuring at 75db, 85db, 95db, 100db and finally 105db, to see if my Mains could hit 105db at the MLP without obvious compression/distortion. I compared each louder trace with the previous lower SPL trace and they all tracked the same. If they had deviated at the higher SPL, I would have stopped.

If anyone wishes to try running a sweep at 100db, DO NOT GO STRAIGHT TO THAT SPL. Run a sweep at 75db at the MLP, then without moving the mic, run one at 85db. If they don't look identical aside from a 10db difference, STOP. Your system can't play higher without significant distortion. Unless you're sure your system is capable, I would increase the level by 5db steps (because it takes approximately 10x the power for a 10db increase) from 85, to 90, then 95, then finally 100db, ALWAYS checking each consecutive graph against the previous lower SPL sweeps. Anytime the louder sweep doesn't look identical to the lower SPL sweeps, STOP.

Hearing protection must be used if you wish to run a full range sweep at 90db or higher. At 95db, the higher octaves were piercing. I see why those sonic 'anti teen loitering' devices are so effective.

My preference though, is to run a sweep at ~80db with a noise floor around 40db or lower and optimize the decay for under 300ms to 40db, which apparently is in sync with Nyal's recommendation. Maybe some of the folks I've spoken to have read the white paper. Well, I'm off to read it. Thanks again for the link Nyal!




Max

 

I may give it a go. I won’t cancel my UMM-6 though. If I get along with REW, as I expect to, I will sell the OM kit so will obviously need the other mic as well.

 

I assume that it is OK to download the updated REW and install it from a flash drive?  My Windows laptop has never been connected to the net and I want to keep it that way so that I don't have to run all that system-crippling anti-virus stuff (lack of need for such being another reason I swapped to the Mac). If I can update from a downloaded file, that will be great.

 

Do you run your laptop on battery power when measuring or on the mains supply?  I always use a fully-charged battery.

Seeing how the new beta supports OSX, why don't you just load it on your Mac? Then you can sell OM AND the Windows lappy?


Max

P.S. If you really like your Windows lappy, then yes, you can DL the file to a flash drive (or an Android smartphone with micro-USB to USB connector?), unless the flash drive needs a different format for Mac vs Windows? Never used Macs. If both lappies can read the drive, then yes, you can do it that way.
Edited by djbluemax1 - 1/22/13 at 4:02am

 

Something wrong there Jerry for sure. That room would not only measure like crap, it would sound like crap too and we all know that is not the case in your room. I wonder what is wrong though?

 

WRT to running 100dB tests, I had already decided NFW!  I will be happy with 85dB. If this level above the noise floor is good enough for Audyssey, it's good enough for me - I have never had the chirps take a second run due to ambient noise, so I am taking it as OK. In my room the noise floor is about 40dB. That means my test sweeps will only be 45dB above the noise floor?  Is that a dealbreaker?  If it is, then I think I am out. Like you, I'm not keen on watching my drivers fire all their little mechanical innards out into the room as they die in a blaze of noisy glory...

 

Incidentally, OM seems to do OK without ear-bleeding levels...

 

Ah - I have the version 1 OmniMic... before measuring the first time you have to type a number (something like 30VR40) into the appropriate box in the software. I don't think it uses a calibration *file* as such (unless that action creates one). I will do a search for .omm files before I try using the OM mic, as posted above.

 

Now that is a nice summary!  Easy to read and to understand. I will try the White Paper but if it is too math-based and technical it will go over my head probably - but those easy to understand conclusions/targets are terrific and incredibly helpful. Thanks!

 

Fabulous information, as always, Max - especially the section I have bolded. Thanks! This thread is turning into a goldmine of useful information! :) 

 

Thanks Max. I decided to use Windows for REW because most of the other guys do, so I figured it would be easier to get advice and support if we were on the same platform. Chances are, once I am confident in using REW, I will switch to the OS-X version.  I have to keep the lappy anyway because I need it for Audyssey Pro. Up to now it has only ever run OmniMic and Audyssey Pro. Now it also runs REW - it is the most reliable Windows machine in the world! LOL. 

 

Thanks for the info on the updating procedure too.

Sub'd.

Version 1 OM does not use a calibration file or an .omm file. You could send it back to PE and they will calibrate the OM and give you an .omm file to input into REW.

 

Thanks for sharing this advice, Max.  I wish I had followed it myself.  I won't be making any measurements at 100dB ever again, and running a sweep at 80-85dB sounds like a good guideline to me.

So Larry, when are you going to come by and school me on the in's and out's of REW?

Bill

Ah right - thanks. I'll be selling the OM kit once I get up and running with REW.

FWIW and I do this professionally, I almost never measure above 85dB SPL (as you would measure with a pink noise signal playing and a C weighted SPL meter). Only exception would be doing a 95dB sweep to identify resonances in the room from things like light fittings, HVAC covers, equipment racks etc. As long as you have 30dB above the noise floor that's good enough IMO.

No math

Thanks Max, I hope it is useful information. I am always interested in finding areas within it that could be improved, new targets added, etc. Jeff and I have discussed a v2 and I would like to do one that adds some targets for multi-channel / HT reproduction.

High level sweeps can be useful to identify compression though.... Like a sub running out of steam down low or a tweeter up top.

That info is not too terribly useful for setup, but good info to have IMO. It can help determine max playback levels as once compression starts distortion normally increases.

 

Thanks Nyall - that is very reassuring. There is just no way I would want to run tests at 105dB!  No way at all. If 85dB is good enough for you, it is good enough for me :)

 

I downloaded it and transferred it to my iPad for reading later. Quick glance - looked very accessible, even for us virgins. Thanks for it!

 

That would be terrific if you could find time to do that!

 

Yes, I was intrigued by Max's method of testing his speakers right up to Reference. Doing it in 5dB steps as he suggests seems like a safe way to do it.