Home Theater Forum and Systems banner

1 - 10 of 10 Posts

·
Registered
Joined
·
47 Posts
Discussion Starter #1
It would be so nice if REW could some day support ASIO.

I realize that it gets a little tricky with Java and would be a feature that is platform specific.

Here is the ASIO SDK you just need to register to download it http://www.steinberg.net/en/company/3rd_party_developer.html

I wanted to measure behavior of EQ ing Meridian does with REW and it was a real kludge to get it to work.

I had to use two USB sound cards with D to A + A to D in both directions, Yuck.

I'm not sure if all the A to D and D to A conversions are causing some weirdness over 10KHz.

The stuff below 250Hz is room correction filters being applied.

The stuff in the middle are Meridian Trifield Center Channel EQ Curves and I wanted to know what they did.

Please disregard the not standard vertical scale. This was very hard to control partly because of the Kludge setup and getting everything all in a reasonable dynamic range. Nothing appears to have clipped.


 

·
Registered
Joined
·
438 Posts
That is quite the crazy config you got going on there.

I too think ASIO would be nice, but I also think that's asking a lot of John. In fact, I don't think any of the other computer based measuring methods use ASIO....the really nice pieces of gear (like TEF, EASE, or LMS, etc...) all have their own external hardware that provides all the test tone creation and capture. Then all the computer sees is digital data (not audio data). At least I think that's how they work. The downside is that requires hardware and I don't think that's in the mindset of REW either.

I think making REW open source might be interesting, but that's also a lot of work for John too and I secretly believe he enjoys the coding...
 

·
Premium Member
Joined
·
6,711 Posts
Looking at those curves I'd say you have "C weighted SPL meter" checked, uncheck that and do "Apply Windows" from the IR Windows dialog for each of the measurements to level things out a bit.

ASIO is on the feature list, although there is a limited amount of Java support for that and using JNI does render the builds platform specific, but I'll get to it eventually.
 

·
Registered
Joined
·
2 Posts
Hello,

Look at jsresources org there's something on ASIO & Java if FAQ, the link is:
faq_misc.xxxx#asio replace xxxx with html can't post link here as newbie.

Haven't studied if it uses JNI or what.
 

·
Registered
Joined
·
47 Posts
Discussion Starter #6 (Edited)
Looking at those curves I'd say you have "C weighted SPL meter" checked, uncheck that and do "Apply Windows" from the IR Windows dialog for each of the measurements to level things out a bit.

ASIO is on the feature list, although there is a limited amount of Java support for that and using JNI does render the builds platform specific, but I'll get to it eventually.
Thank you I'll check the C-Weight option. I normally use an ECM8000 Mic so I may have that off already but it's possible it's on as I just did a fresh install of REW and may have forgot. Keep in mind these EQ curves from Meridian do take into consideration how Humans hear and may have a C-Weight like shape. Thinking about it now, I'll bet I have C-Weight off but the ECM8000 cal file still on.

That's awesome you have ASIO at least on your Radar. I understand the awkwardness of dealing with it in Java. I believe ASIO is a straight DLL interface no COM or anything so it might not be too bad.
 

·
Registered
Joined
·
47 Posts
Discussion Starter #7
One other question regarding this test I'm doing.

My goal is to find the best EQ setting (4 to choose from) for my Center to best match my mains. It would be nice to remove all room, speaker and Mic effects and have that be my correction curve applied to everything.

So I would do a sweep measurement with the Mic and Main left speaker, and some how set that as the correction curve and if I measure again it should be flat. Then switch to the center and measure all 4 EQ's and see which is the flatest.

My measurements I did above was just a curiousity of what the EQ's were trying to do.
 

·
Premium Member
Joined
·
6,711 Posts
It looks like the trifield settings are more aimed at changing how the centre sounds than matching it to anything in particular, the variations in the responses in your plots are unlikely to tie up with the measured response of any speaker and certainly not the "Left" measurement in your plot.
 

·
Registered
Joined
·
47 Posts
Discussion Starter #9
It looks like the trifield settings are more aimed at changing how the centre sounds than matching it to anything in particular, the variations in the responses in your plots are unlikely to tie up with the measured response of any speaker and certainly not the "Left" measurement in your plot.
Agreed.
 

·
Registered
Joined
·
47 Posts
Discussion Starter #10 (Edited)
Ok, I got it perfect now, very interesting results.

You can see for reference that the "Left Direct" is perfectly flat from 20-20Hz. That means my recording setup is not doing anything weird at any frequency. I think the ECM8000 calibration (which was on) was messing things up in my last set of plots.

I did use the analog outputs and was able to simplify down to one USB sound card. I could not get the EMU to work without ASIO but the TASCAM worked perfect with or without ASIO. Except without ASIO I have to work in Analog.

Rember there is No Mic, No Room Effects, No Speaker Effects here. This is what Trfield Center EQ does to your music.

Keep in mind that this is not capturing everything Trifield does. This does not capture how Trifield plucks out uncorrelated and correlated information.

I removed MRC (room correction) and all speakers and added 5 large analog speakers.

When I first did the test it looked exactly like I had small speakers rolled off at 80hz. I checked and rechecked. Then I remembered the Speaker Size thingy. The volume was quite high on the G86 (96) to get all the levels in the correct dynamic range. I have to match levels on a feedback channel that REW uses for realtime calibration (standard setup with these room calibration tools like REW and others do this). So then I set all speakers to MAX to shut that off.

Then I did Trifield Left to compare and you can see there is something being done to the mains (I was wrong about that). But oddly enough it slowly boosts frequency as you go up in frequency. Then I shutoff the center using Speaker Layout and measured Trifield Left again (expecting that with no center that it might be flat). It was not and has a huge high freq roll off.

I also checked if Center EQ effects mains and it does not (didn't bother to show plot).

For fun I did the rears too (only with Trifield Rear EQ=flat) and well it's flat.

The Center EQ with Flat, EQ1, EQ2, EQ3 should be very accurate now for the entire 20-20Khz range.

The top 3 traces are all Left Channel (Trifield in 2 conditions and Direct)

The middle 4 Traces are the Center Channel with the 4 EQS (including Flat).

The Bottom Trace is the Rear.

All the relative levels of these should be accurate. Absolute levels are not.

Nothing is being smoothed here. Just 4 Sweeps Averaged.

You'll also notice that above 10Khz the 4 EQS on Center don't differ by that much. I think the center always has a gentle roll off almost all the way through which is what the flat EQ is showing.

 
1 - 10 of 10 Posts
Top