or Connect
AVS › AVS Forum › Audio › Audio theory, Setup and Chat › Simplified REW Setup and Use (USB Mic & HDMI Connection) Including Measurement Techniques and How To Interpret Graphs
New Posts  All Forums:Forum Nav:

Simplified REW Setup and Use (USB Mic & HDMI Connection) Including Measurement Techniques and How... - Page 224

post #6691 of 9584
and lastly the raw graph comparison is pretty different as well. The only thing thing changed is the crossover from from 80hz to 150hz. Sub trim is set at what Audyssey set it at, Audyssey is off, mini dsp is entirely bypassed...



eek.gif

How can 50-60 hz go from a 3db variance to a 13db variance? The mic I bet is within 5 inches of where it was the other day. The only room change is I moved a flatscreen 23 inch computer monitor and a very small lamp. I guess the ottoman can be out another few inches too?

Also the right speaker I changed to a different speaker wire

Ok, I just realized PGM 2 was still implemented on the subversives. But it still looks different and the 50-60 hz area baffles me. surely a 5inch variance can't make a 10 db difference right?

Edited by jlpowell84 - 11/11/13 at 7:00pm

Gear mentioned in this thread:

post #6692 of 9584
When I am comparing two graphs is it preferred that I overlay or two graphs like I have here?
post #6693 of 9584
Quote:
Originally Posted by jlpowell84 View Post

When I am comparing two graphs is it preferred that I overlay or two graphs like I have here?

I suppose that is a personal preference, but I find it easier when they are plotted on the same graph in contrasting colors.
post #6694 of 9584
Quote:
Originally Posted by jlpowell84 View Post

When I am comparing two graphs is it preferred that I overlay or two graphs like I have here?
Overlay makes comparisons of FR graphs easier to see, and change the limits on your graph so that it shows 5db steps as opposed to 10db steps.


Max
post #6695 of 9584
Quote:
Originally Posted by djbluemax1 View Post

Overlay makes comparisons of FR graphs easier to see, and change the limits on your graph so that it shows 5db steps as opposed to 10db steps.


Max

+1
post #6696 of 9584
How do I change to 5b steps?

I have some new graphs i want to upload but I need to figure out 5db increments...
post #6697 of 9584
Got it...sigh...sigh...hmm...to big of a vertical db selection...cool.gif
post #6698 of 9584
I think I may have developed some significant decay improvements. I wasn't quite expecting this but I guess when using the mini dsp to cut it REALLY helps present Audyssey with a much better and easier job to do. In fact unless I am missing something here this could be a significant reason to a growing number of users of the mini dsp and Audyssey partnership. I remember a couple of you posting pre and post Audyssey waterfall and spectrograms. You're EQ lines were much better but at the sake of bloated decay times.

Big disclaimer here. I have always used one of the PEQ options called "peak". Now I selected an option this time called "subwoofer EQ". It had an different and desirable effect on the EQ line. For example if I had a -8db cut selected with a Q of 3.5 and selected "Sub EQ" and then "Peak" with the same -8db gain selection and the same 3.5 Q setting it made entirely different looking EQ lines. The peak option had a more singular type effect and the Sub EQ a broader effect if that makes sense. But it obviously and unexpectedly had a effect on decay. If anyone is interested I would be willing to post direct graph comparisons of the frequency response and waterfall/spectrogram with just the peak and sub EQ selections changed to get a DIRECT comparison and what it does.

Here is the baby...

This is my waterfall after the other day when just using the peak cuts. An important note is both times I had PGM 2 on the submersives selected, and a low shelf filter engaged. The biggest difference is the before/worse graphs I had an 80hz crossover and the better graphs have an 150hz crossover.


This is the waterfall after using the "Sub EQ"



Here are the Spectrograms.
Before with peak EQ cuts


And after with Sub EQ cuts


And here are a couple waterfalls before the low shelf filter was engaged. I think I will end up with a shelf filter that is about a 6db up at 30hz over 100hz rather than the 10db up I have now. I thought the decay times were pretty dang impressive! I could way off base with everything but it seems legit to me. These two waterfalls are the development of the house curve. The one you seen above is the final.



We are talking basically 200ms almost to 30hz!
I don't even have a room treatment built yet...
Edited by jlpowell84 - 11/12/13 at 12:14am
post #6699 of 9584
Quote:
Originally Posted by djbluemax1 View Post

The animation does a great job of illustrating the iceberg analogy. The small differences near the bottom of the graph (aside from just exposing more of the bottom of the iceberg appear to simply be the slight differences that occur between measurements, even with identical mic placements.

...or it's noise.
Quote:
Originally Posted by djbluemax1 View Post

Now you mentioned 2 things:
A) Important = where the ringing starts relative to the first slice and how fast it decays
B) Not important, the bottom of the graph

For A) The ringing in the ~50-55Hz region begin fairly high up and continues all the way down and I assume that means that it's something that should be looked into. In comparison, the ~30Hz region starts extending out pretty low down on the graph. Does that mean it's nothing to be concerned about? Likewise with the ~100Hz region. The mountain begins to stick out pretty low (~25db) down from the top of the graph (the first slice?). Granted, in the 100Hz region the time to decay in the noise floor is already pretty short, but if, for example it still extended out to 450ms at the bottom of the graph, would it still be insignificant?


Max

Those "shoulders" sticking out after and below the first slice (t=0) are either remaining ringing in an already damped room or noise. That high up it's probably NOT noise.

Now the questions is, how much low frequency ringing is tolerable? In my opinion it should be as low as possible. Very hard to do with passive absorption.
Listen to successive short sine bursts to get an idea how much the room changes the sound coming from the speaker. You can download such signals from here: https://dl.dropboxusercontent.com/u/21936387/audio/test_signals/LF_TEST.zip
First listen to them with headphones on to get an idea how they should sound then listen over your speakers.
Edited by markus767 - 11/12/13 at 1:56am
post #6700 of 9584
Quote:
Originally Posted by jlpowell84 View Post

I think I may have developed some significant decay improvements.

You haven't.

The top graphs have peaks @ 100dB and the bottom graphs only peak around 87dB. Adjust your bottom graphs to reveal an extra 13dB and the waterfalls will look all but identical!

On another note, if to -35dB is all that's interesting in waterfalls, mine looks decidedly better smile.gif



Still wouldn't mind working on the 60-70hz region though.
Edited by Audionut11 - 11/12/13 at 1:46am
post #6701 of 9584
Quote:
Originally Posted by Audionut11 View Post
 
Quote:
Originally Posted by jlpowell84 View Post

I think I may have developed some significant decay improvements.

You haven't.

The top graphs have peaks @ 100dB and the bottom graphs only peak around 87dB. Adjust your bottom graphs to reveal an extra 13dB and the waterfalls will look all but identical!

 

Rather than looking at the waterfalls, I presume the spectrogram differences is what he was purporting.  If the graphs posted relate to each waterfall (i.e. top waterfall goes with top spectrogram and so on), then the Sub EQ cuts appear to have had a nice impact on his bass decay above 30Hz:

 

Compare Peak EQ cuts:

 

 

With Sub EQ cuts:

 

 

Appears to be a noticeable improvement imho.

 

@jlpowell84:  I've been meaning to ask you this for a while but would you mind using the "large" preview option when posting graphs?  It makes it a lot easier to review your plots.

post #6702 of 9584
Considering all the waterfalls graphs are scaled differently, I don't think it's safe to assume anything.
The peak EQ spectro has larger areas of peak red imho, this could mean that the top scale was incorrect which will result in 2 different noise floors.
The only way to be sure is if jlpowell84 posts the mdats.

@jlpowell84
Which plugin are you using for minidsp? I don't see any subEQ filters in the minidsp, nor in the EQ tool in REW.
post #6703 of 9584
I do select large. Perhaps a Mac setting? Keith do you know?

I do large setting and hit submit and this is what we get. I can't tell from memory but yours are bigger right Keith?
Edited by jlpowell84 - 11/12/13 at 8:15am
post #6704 of 9584
Quote:
Originally Posted by Audionut11 View Post

Considering all the waterfalls graphs are scaled differently, I don't think it's safe to assume anything.
The peak EQ spectro has larger areas of peak red imho, this could mean that the top scale was incorrect which will result in 2 different noise floors.
The only way to be sure is if jlpowell84 posts the mdats.

@jlpowell84
Which plugin are you using for minidsp? I don't see any subEQ filters in the minidsp, nor in the EQ tool in REW.

I understand they are scaled a little differently. I hope people don't throw a tissy fit and just not reply because of that. That would be quite ridiculous. Even scaled differently look at the decay time! The MS is quite different right?

I use the Open DRC AN and I assure you it's there.
post #6705 of 9584
Quote:
Originally Posted by Audionut11 View Post

Considering all the waterfalls graphs are scaled differently, I don't think it's safe to assume anything.
The peak EQ spectro has larger areas of peak red imho, this could mean that the top scale was incorrect which will result in 2 different noise floors.
The only way to be sure is if jlpowell84 posts the mdats.

@jlpowell84
Which plugin are you using for minidsp? I don't see any subEQ filters in the minidsp, nor in the EQ tool in REW.

I will post identical information if you can't see that the decay times are quite different from a slightly different graph value. It may be a couple days as I am away from home.
post #6706 of 9584
I can see differences, sure. But just taking the top 2 waterfalls into account, you have 1 scaled with a peak amplitude around 100dB with a noise floor of 60dB, and the other with virtually the same peak amplitude but a noise floor of 50dB, both with different time ranges.

Surely you can understand that it would be a helva lot easier to see meaningful differences if they were scaled the same. And I personally wouldn't call it ridiculous if people don't reply to your graphs, when they have to take the time to allow for the differences in scales when making determinations, simply because you didn't take care in ensuring they we scaled identically. Right?
Or maybe I am the only one with the problem wink.gif

MS?
Edited by Audionut11 - 11/12/13 at 7:58am
post #6707 of 9584
Milliseconds
post #6708 of 9584
Quote:
Originally Posted by Audionut11 View Post

I can see differences, sure. But just taking the top 2 waterfalls into account, you have 1 scaled with a peak amplitude around 100dB with a noise floor of 60dB, and the other with virtually the same peak amplitude but a noise floor of 50dB, both with different time ranges.

Surely you can understand that it would be a helva lot easier to see meaningful differences if they were scaled the same.

MS?

I had 600 ms because the decay time was so bad. I had to shrink the noise floor to get 5 db vertical increments. I could have made the same but it was late and I felt the point I was trying to convey was easily read enough without recreating another snapshot of old mdats
post #6709 of 9584
I personally like to scale to the measurement that has the least desirable results, so that it is very easy to compare visually what effect any changes have made. It rules out guesswork!

And while you might be able to easily determine what is happening, and someone with sufficient experience might be able to easily determine the outcome, I like to keep things similar in an effort to assist newcomers. That's why the thread has graph rules afaik.
post #6710 of 9584
Quote:
Originally Posted by Audionut11 View Post

I personally like to scale to the measurement that has the least desirable results, so that it is very easy to compare visually what effect any changes have made. It rules out guesswork!

I understand and agree with that approach 100%. It was just a long night is all. Perhaps I should have waited... frown.gif
It seems easy to see the drastic difference in the extension of decay on the millisecond line to me though. I am new and get excited at what looks like improvements or a discovery that makes improvements. Please have mercy on my me smile.gif
post #6711 of 9584
I'm fairly new to acoustical measurements also, having long only had my ears to make judgements.
That's probably why I'm harping on like a broken record about it wink.gif

It's the different noise floors that are throwing me. The different frequency scales aren't helping either, just to throw salt into the wound tongue.gif I'll be quiet now!
post #6712 of 9584
I have no need or desire to be right or have the last word. Just look at what that attitude has done to the Audyssey thread. I made my graphs wrong. If you regular seasoned REW thread guys don't want to comment for the same reason perhaps give a +1 and I will accept no comments until I fix my graph values. No comments on several hours of work is very discouraging smile.gif
post #6713 of 9584
Quote:
Originally Posted by jlpowell84 View Post
 
I had 600 ms because the decay time was so bad. I had to shrink the noise floor to get 5 db vertical increments. I could have made the same but it was late and I felt the point I was trying to convey was easily read enough without recreating another snapshot of old mdats

 

I'm guessing your dB increment issue is related to using a Mac.  I just checked my version and no matter what I set the limits to in the waterfall, it's always 5dB increments.  Perhaps this is a setting somewhere else in REW?

post #6714 of 9584
Quote:
Originally Posted by jlpowell84 View Post

I have no need or desire to be right or have the last word. Just look at what that attitude has done to the Audyssey thread. I made my graphs wrong. If you regular seasoned REW thread guys don't want to comment for the same reason perhaps give a +1 and I will accept no comments until I fix my graph values. No comments on several hours of work is very discouraging smile.gif

 

Try not to be so defensive.  I think Markus illustrated fairly well yesterday how misrepresentation of information can very quickly lead to misinterpretation.  I'm fairly certain at least a few of us were convinced that higher SPL measurements resulted in "different" waterfall representations.  Audionut11 makes a good point when trying to compare two plots.  They really need to be the same scale with the same noise floor to make comparison (and ultimately feedback to you) worthwhile.  You might be better served to run your REW using the Windows emulation mode if that's an option as it seems more than one issue has arisen with your plots that could be related (i.e. preview size and scale increments).

post #6715 of 9584
Quote:
Originally Posted by jkasanic View Post

I'm guessing your dB increment issue is related to using a Mac.  I just checked my version and no matter what I set the limits to in the waterfall, it's always 5dB increments.  Perhaps this is a setting somewhere else in REW?

I don't know. I did increase the top end awhile back to get a higher measurement. Remind me, are the basic graph values we go by in the start up guide right?

Nonetheless you seen the spec you circled and the differences. Speaker and sub placement have NOT CHANGED. Just the filter option from peak to sub eq. It makes me wonder if the sub eq is taking decay into consideration. It does say it disables from 7 to 6 filters available when sub eq is selected. That is my best guess. I will see what the mini dsp thread yeilds. Cool though smile.gif
post #6716 of 9584
Quote:
Originally Posted by jlpowell84 View Post
 
I don't know. I did increase the top end awhile back to get a higher measurement. Remind me, are the basic graph values we go by in the start up guide right?

 

The 3.1 version of the guide says this:

 

 

I think this is a good starting point.  You might find that you can decrease the top limit depending on your peaks but as long as you present the info consistently, I don't think anyone will quibble over it.

post #6717 of 9584
Quote:
Originally Posted by jkasanic View Post

Try not to be so defensive.  I think Markus illustrated fairly well yesterday how misrepresentation of information can very quickly lead to misinterpretation.  I'm fairly certain at least a few of us were convinced that higher SPL measurements resulted in "different" waterfall representations.  Audionut11 makes a good point when trying to compare two plots.  They really need to be the same scale with the same noise floor to make comparison (and ultimately feedback to you) worthwhile.  You might be better served to run your REW using the Windows emulation mode if that's an option as it seems more than one issue has arisen with your plots that could be related (i.e. preview size and scale increments).

That's a tough one. I can't see myself buying windows just for that reason. Although having both on my Mac would certainly have benefits. If it is just window display size and db increments that us hard for me to justify. Hurry what's another reason I would need windows?
post #6718 of 9584
Quote:
Originally Posted by jkasanic View Post

The 3.1 version of the guide says this:




I think this is a good starting point.  You might find that you can decrease the top limit depending on your peaks but as long as you present the info consistently, I don't think anyone will quibble over it.

Ok. I do remember I was at 115 top and had to hit 59 bottom was the changing point from the vertical axis being 10 to 5. So if I brought down to 110 that would mean 54 minimum for the bottom. I will confirm this all soon.
post #6719 of 9584
Quote:
Originally Posted by jlpowell84 View Post

Remind me, are the basic graph values we go by in the start up guide right?

Yeah they should be.

edit: Already answered above.

Quote:
Originally Posted by jlpowell84 View Post

Nonetheless you seen the spec you circled and the differences. Speaker and sub placement have NOT CHANGED. Just the filter option from peak to sub eq. It makes me wonder if the sub eq is taking decay into consideration. It does say it disables from 7 to 6 filters available when sub eq is selected. That is my best guess. I will see what the mini dsp thread yeilds. Cool though smile.gif

I've had a quick look at your model and still can't find the subEQ option (not saying it isn't there, it clearly is), I'd just like to know what it is and how it works smile.gif
Could it be switching to a FIR filter and hence helping to reduce ringing?
post #6720 of 9584

I've been watching the BF ads and there is some awesome deals on HDMI laptops with a copy of Windows preinstalled.  I may pick one up just as my dedicated measurement device (since I currently don't have HDMI on my laptop).

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Audio theory, Setup and Chat

Gear mentioned in this thread:

AVS › AVS Forum › Audio › Audio theory, Setup and Chat › Simplified REW Setup and Use (USB Mic & HDMI Connection) Including Measurement Techniques and How To Interpret Graphs