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 290

post #8671 of 9577
Quote:
Originally Posted by Kelvin1965S View Post

Just to close this off, I finally got round to having a listen to the Queen DVD again after re running Audyssey. While it seems a bit better than previously, I think I just prefer this disc with Audyssey switched OFF: It just seems to add an edge to the sound that I don't like, though it could be due to this being an older disc and only DTS rather than DTS MA.

What it does mean is that I will try to further improve my manual Pre Audyssey subwoofer/crossover results when I can measure with only one speaker at a time, so I'll be able to have a good bass response without using Audyssey, but without the 'edge' added to music. I think I still like using Audyssey for surround (movie) content because it helps to blend the speakers together and at least when I drop my screen it dampens the higher end a little (not unlike Re-EQ) so perhaps I should just do my Audyssey set up with the screen up and then lower it to act as a treble cut. smile.gif

 

It's worth remembering that Audyssey was never designed to work with music really. Music has no standardised content criteria so it's all a bit of a kludge. Various combinations of DEQ on and off, with various RLOs have all been tried, with varying degrees of success for music. Quite a few members report preferring Audyssey off altogether for music and use it only for movies. I can’t comment as I don't listen to music as such in my HT room, favouring a separate, ancient, 2-ch system in another room for music.

Gear mentioned in this thread:

post #8672 of 9577
Quote:
Originally Posted by Brian Fineberg View Post




a more typical audyssey curve...notice the dropoff in the HF

 

Brian, it is not a good idea to measure full range plots with the R and L speakers working together. They interfere with each other in a way that is not representative of real content, so R+L measurements aren’t very useful.  If movies are your main concern, I suggest measuring the Centre channel + Sub(s) to give you a better idea of what is going on. The centre channel absolutely dominates for movies, carrying as much as 80% of all the content, so it is vital to get this channel optimised. Obviously if you can also optimise L and R then that is great, but not if it is at the expense of the centre channel.

 

The current graph (above) doesn't reveal what is actually happening in your room because of the interaction between L and R speakers when played together using a single monophonic sweep.

post #8673 of 9577
Quote:
Originally Posted by kbarnes701 View Post

Brian, it is not a good idea to measure full range plots with the R and L speakers working together. They interfere with each other in a way that is not representative of real content, so R+L measurements aren’t very useful.  If movies are your main concern, I suggest measuring the Centre channel + Sub(s) to give you a better idea of what is going on. The centre channel absolutely dominates for movies, carrying as much as 80% of all the content, so it is vital to get this channel optimised. Obviously if you can also optimise L and R then that is great, but not if it is at the expense of the centre channel.

The current graph (above) doesn't reveal what is actually happening in your room because of the interaction between L and R speakers when played together using a single monophonic sweep.

good to know....I guess things have changed since I last read the FAQ. now I gotta pull out the measuring gear again haha (at least for full range) I was indeed measuring C + SW's for 10-300Hz:
post #8674 of 9577
Quote:
Originally Posted by Brian Fineberg View Post

good to know....I guess things have changed since I last read the FAQ. now I gotta pull out the measuring gear again haha (at least for full range) I was indeed measuring C + SW's for 10-300Hz:

The Guide still includes R+L+Subs in the Measurement Guidelines (page 73) - so perhaps a revision is called for.

Measurement Guidelines
Measurement Speaker(s)
Sub(s)
On/Off
Audyssey
On/Off AVR Sound Mode Measurement Range
Graph
Smoothing

1 Left Front Off Off Stereo 15Hz-20kHz 1/6
2 Left Front Off On Stereo 15Hz-20kHz 1/6
3 Right Front Off Off Stereo 15Hz-20kHz 1/6
4 Right Front Off On Stereo 15Hz-20kHz 1/6
5 Center Off Off PLII Cinema * 15Hz-20kHz 1/6
6 Center Off On PLII Cinema * 15Hz-20kHz 1/6
7 Sub1 + sub2 On Off Stereo 15Hz-300Hz None
8 Sub1 + sub2 On On Stereo 15Hz-300Hz None
9 Left+Right+Sub(s) On Off Stereo 15Hz-300Hz None
10 Left+Right+Sub(s) On On Stereo 15Hz-300Hz None
11 Center + Sub(s) On Off PLII Cinema * 15Hz-300Hz None
12 Center + Sub(s) On On PLII Cinema * 15Hz-300Hz None
post #8675 of 9577
Quote:
Originally Posted by ggsantafe View Post

The Guide still includes R+L+Subs in the Measurement Guidelines (page 73) - so perhaps a revision is called for.

Measurement Guidelines
Measurement Speaker(s)
Sub(s)
On/Off
Audyssey
On/Off AVR Sound Mode Measurement Range
Graph
...
9 Left+Right+Sub(s) On Off Stereo 15Hz-300Hz None
10 Left+Right+Sub(s) On On Stereo 15Hz-300Hz None

L+R+Subs is still legitimate IMO for an analysis focusing on the < 300 Hz region, if your focus is accessing the overall bass response - which tends to be omnidirectional - rather than 'full-range' performance of the speakers (15 Hz-20 kHz). Many of us routinely ran L/R joint plots to try to summarize the 'big picture' before we were shown that the phase cancellation and spurious HF interaction from the plots rendered them invalid analytically. I want to say the change in collective philosophy dated back to last summer or shortly later, at least as a consensus.
Edited by sdrucker - 2/10/14 at 8:18am
post #8676 of 9577
Quote:
Originally Posted by djbluemax1 View Post

What Markus is pointing out is that your Audyssey graph is showing a rising frequency response from 300Hz onwards. Something is wrong because you're showing a more than 10db increase from 300Hz to 20kHz. None of us have EVER seen Audyssey do that.

The standard Audyssey curve tries to make the FR pretty much flat from the lowest frequency your system can play flat to, up to 10kHz, and the response curves downward from 10kHz to 20kHz. There's also usually a small dip at 2kHz called the MRC or MidRange Compensation dip.

The other Audyssey curve is the Flat curve, which depending on the implementation, is pretty much as it says, flat. Some implementations still include the MRC, some don't.

We need to figure out why your Audyssey graph looks like that.


Max

Ahh...I see now what Markus is talking about. I didn't even notice that was a problem. I new to this whole graph reading thing.

Thanks for making it clear to me.
post #8677 of 9577
Quote:
Originally Posted by markus767 View Post

Place it exactly where you've placed the Audyssey mic for measurement 1.
I'd reset the AVR and rerun Audyssey. This time do only the mandatory 3 Audyssey measurements and keep the Audyssey mic at measurement location 1. Then measure that point with REW.

Ok I'll give that try, will have to wait till I can get the house to myself though.
post #8678 of 9577
Quote:
Originally Posted by Brian Fineberg View Post
 
Quote:
Originally Posted by kbarnes701 View Post

Brian, it is not a good idea to measure full range plots with the R and L speakers working together. They interfere with each other in a way that is not representative of real content, so R+L measurements aren’t very useful.  If movies are your main concern, I suggest measuring the Centre channel + Sub(s) to give you a better idea of what is going on. The centre channel absolutely dominates for movies, carrying as much as 80% of all the content, so it is vital to get this channel optimised. Obviously if you can also optimise L and R then that is great, but not if it is at the expense of the centre channel.

The current graph (above) doesn't reveal what is actually happening in your room because of the interaction between L and R speakers when played together using a single monophonic sweep.

good to know....I guess things have changed since I last read the FAQ. now I gotta pull out the measuring gear again haha (at least for full range) I was indeed measuring C + SW's for 10-300Hz:
 

 

As Stuart has already said, LR+Subs is legitimate for examining the 15-300Hz range for bass - just not for full range, so it is still worthwhile doing a L+R+Subs measurement for that purpose. But again, if you are optimising for movies, attach more importance to the C+Subs measurement than the L+R+Subs measurement, if any compromise is needed. I find that when I get the best C+Subs measurement I can, it causes the L+R+Subs to be slightly less good (I am thinking of sub delay settings) and vice-versa, so I always choose the best setting for delay for the C+Subs combination. IDK how this would translate to a music environment as the latter is of no concern to me.

post #8679 of 9577
Quote:
Originally Posted by kbarnes701 View Post

The individual plots are nicely similar, but something is wrong. Audyssey can't create that rising frequency response, so something else is causing it. You don't have that rising response with Audyssey off, and Audyssey can’t create it on its own, so something else is. What? is the question.

Can you describe your Audyssey mic placements and general technique?

I have a 3 seat couch.....

#1 audyssey mic placement is at head position in the center seat of the couch
#2 is head position left seat,
#3 head position right seat
#4 head position 2' in front of #1
#5 head position 2' in front of #2
#6 head position 2' in front of #3
#7 head position between 1 and 2
#8 head position between 2 and 3

REW mic placement is position #1 in audyssey.

When I run audyssey I unplug the fridge, and make sure nothing else in the house is running that could alter the measurement process. I place the mic in position one and let audyssey do it's chirps, then in the same order as above I move the mic, and let audyssey do it's thing.

One thing that I'm thinking is when I ran audyssey I used a camera tripod. I just recently got the boom mic and haven't reran audyssey with it. I used the boom mic for REW though. Maybe this could have a effect?

Anyways this is probably getting OT here, I'll take this over the auddyssey thread.

Thanks
post #8680 of 9577
Quote:
Originally Posted by kbarnes701 View Post

As Stuart has already said, LR+Subs is legitimate for examining the 15-300Hz range for bass - just not for full range...

Quote:
Originally Posted by sdrucker View Post

L+R+Subs is still legitimate IMO for an analysis focusing on the < 300 Hz region, if your focus is accessing the overall bass response - which tends to be omnidirectional - rather than 'full-range' performance of the speakers (15 Hz-20 kHz). Many of us routinely ran L/R joint plots to try to summarize the 'big picture' before we were shown that the phase cancellation and spurious HF interaction from the plots rendered them invalid analytically. I want to say the change in collective philosophy dated back to last summer or shortly later, at least as a consensus.

Quote:
Originally Posted by markus767 View Post

The high frequency drop is probably the result of interference effects caused by measuring more than one speaker at the same time.

Quote:
Originally Posted by ggsantafe View Post

The Guide still includes R+L+Subs in the Measurement Guidelines (page 73) - so perhaps a revision is called for.

L+R is a legitimate measure for full range IF you get the mic EXACTLY equidistant from each speaker. If you miss it by even 1/4", you will get what Marcus described and what some are seeing in the upper range of their full range graphs.

Admittedly, this is hard with REW where you don't have real time readings. So therefore, I don't think the guide needs to omit this. Perhaps what it does need is to emphasize this point.

If you are NOT measuring at the LP whereby the LP is between the speakers, or a secondary seating position not located between the speakers, then all bets are off and you should not try L+R for full range at all.
post #8681 of 9577
Quote:
Originally Posted by kbarnes701 View Post

As Stuart has already said, LR+Subs is legitimate for examining the 15-300Hz range for bass - just not for full range, so it is still worthwhile doing a L+R+Subs measurement for that purpose. But again, if you are optimising for movies, attach more importance to the C+Subs measurement than the L+R+Subs measurement, if any compromise is needed. I find that when I get the best C+Subs measurement I can, it causes the L+R+Subs to be slightly less good (I am thinking of sub delay settings) and vice-versa, so I always choose the best setting for delay for the C+Subs combination. IDK how this would translate to a music environment as the latter is of no concern to me.

I haven't done it yet, but if I wind up connecting my pre-outs for my mains to the MiniDSP 10x10 HD, I can even simulate a L/C/R+Subs run without having to use Multchannel Stereo or some odd mode. I think that should work by having a single channel input, and then using the Matrix Routing editor to route the output to all three mains. In my case, with powered DefTech fronts, that might provide the best picture of what the summated bass response looks like, particularly for multichannel music. I can then optimize my sub distance to have the smoothest response for the entire < 300 hz region.
post #8682 of 9577
Quote:
Originally Posted by jim19611961 View Post




L+R is a legitimate measure for full range IF you get the mic EXACTLY equidistant from each speaker. If you miss it by even 1/4", you will get what Marcus described and what some are seeing in the upper range of their full range graphs.

Admittedly, this is hard with REW where you don't have real time readings. So therefore, I don't think the guide needs to omit this. Perhaps what it does need is to emphasize this point.

If you are NOT measuring at the LP whereby the LP is between the speakers, or a secondary seating position not located between the speakers, then all bets are off and you should not try L+R for full range at all.

And who among us (other than possibly AustinJerry, if he's reading this) is OCD enough to get the mic within 0.25 inch of being exactly equidistant? By my count, my mains are about 0.1 feet from one another on the Audyssey distance settings (within 0.08 meter on my Sherwood Trinnov settings) so I'm "pretty close". I will add that when I did L/R vs. separate speaker runs, my HF response above the mid-range correction looked almost identical for both mains, and other than the cumulative increase in db due to a joint plot, the response was spot on at 1/6th smoothing. However, in deference to the majority opinion on the thread, I only look at joint plots for my own amusement and don't regard them as offering any analytic insight anymore.
post #8683 of 9577
I just want to say that after reading a chunk of this thread, I feel confident enough to buy one of these:
http://cross-spectrum.com/measurement/calibrated_umik.html

Time to see how bad my room response is. I really want to thank you guys for all your hard work and thanks to everyone that comes into this thread and helps us new guys dial everything in. Hope to have it ordered in a week or two.
post #8684 of 9577
Quote:
Originally Posted by sdrucker View Post

And who among us (other than possibly AustinJerry, if he's reading this) is OCD enough to get the mic within 0.25 inch of being exactly equidistant? By my count, my mains are about 0.1 feet from one another on the Audyssey distance settings (within 0.08 meter on my Sherwood Trinnov settings) so I'm "pretty close". I will add that when I did L/R vs. separate speaker runs, my HF response above the mid-range correction looked almost identical for both mains, and other than the cumulative increase in db due to a joint plot, the response was spot on at 1/6th smoothing. However, in deference to the majority opinion on the thread, I only look at joint plots for my own amusement and don't regard them as offering any analytic insight anymore.

Firstly, some folks here have measurement software that measures in real time. In this case, its pretty easy.

Secondly, I didn't want the guide to make it sound like such a measure cant be done under any circumstances. It can.

For those that don't want to bother with L+R measurement criteria (getting the mic in the right place), they don't have to. I am not saying everyone should do this measurement and should fret with the mic to the point of OCD. Certainly this would not be the first circumstance where being a bit anal about something is necessary to achieve a proper result however smile.gif But I think it should be conveyed what must be done to get that correct result.
post #8685 of 9577
Quote:
Originally Posted by sdrucker View Post

And who among us (other than possibly AustinJerry, if he's reading this) is OCD enough to get the mic within 0.25 inch of being exactly equidistant?

It's not as bad as all that. This post says that acourate has a mic positioning tool built in. So one could always just run that first.

[edit]I've been assuming that acourate is open source but it's not. So it's not really possible to run it first as I suggested without shelling out
post #8686 of 9577
Quote:
Originally Posted by artur9 View Post

It's not as bad as all that. This post says that acourate has a mic positioning tool built in. So one could always just run that first.

[edit]I've been assuming that acourate is open source but it's not. So it's not really possible to run it first as I suggested without shelling out

Maybe I'm not rigorous enough, but 300 Euros (non-EU price) to get the software just to do this test doesn't strike me as cost-effective. Unless there's other ancillary benefits that supercede what we already get with our automated REQ or REW, it's not worth the USD to me. And it doesn't sound like it would work directly with our UMM-6 or UMIK-1 mics, taking us back to phantom power supply.

I'm assuming that I'd need to use the Pro Acourate software, not the OpenDRC version (which uses the 75 Euro AcourateDRC, but means an investment in a different MiniDSP platform than I already have).
post #8687 of 9577
Quote:
Originally Posted by sdrucker View Post

Maybe I'm not rigorous enough, but 300 Euros (non-EU price) to get the software just to do this test doesn't strike me as cost-effective. Unless there's other ancillary benefits that supercede what we already get with our automated REQ or REW, it's not worth the USD to me. And it doesn't sound like it would work directly with our UMM-6 or UMIK-1 mics, taking us back to phantom power supply.

I'm assuming that I'd need to use the Pro Acourate software, not the OpenDRC version (which uses the 75 Euro AcourateDRC, but means an investment in a different MiniDSP platform than I already have).

Another way, at least in theory (haven't tried it), is to run a 10K-15K sine wave in both channels, and looking at it using the RTA function in the free REW version. In this manner, you are getting real time results and should be a able to tweak the mic fairly easily. In this scenario, you adjust the mic position to where the gain seems to be the highest. Then run REW with a normal sweep to confirm.
post #8688 of 9577
Quote:
Originally Posted by jim19611961 View Post

Another way, at least in theory (haven't tried it), is to run a 10K-15K sine wave in both channels, and looking at it using the RTA function in the free REW version. In this manner, you are getting real time results and should be a able to tweak the mic fairly easily. In this scenario, you adjust the mic position to where the gain seems to be the highest. Then run REW with a normal sweep to confirm.

Seems to me that this is a good way to confirm that the MLP is where you think it is. At least, relative to where your L/R speakers are.
post #8689 of 9577
^^

I thought we did all that stuff based on geometry?

Jeff
post #8690 of 9577
Quote:
Originally Posted by pepar View Post

^^

I thought we did all that stuff based on geometry?

Jeff

Can you get the geometry (by measurement alone) within < 1/4" ? I cant.
post #8691 of 9577
Quote:
Originally Posted by jim19611961 View Post

Can you get the geometry (by measurement alone) within < 1/4" ? I cant.

Why is that important? With two ears on a "movable" head, I don't see the criticality of that level of precision. Do you?

Jeff
post #8692 of 9577
Which beer bottle is the correct position to place the microphone?

Whoops, I forgot the rules here. I don't have a USB microphone so we can't talk about it!wink.gif


post #8693 of 9577
Quote:
Originally Posted by pepar View Post

Why is that important? With two ears on a "movable" head, I don't see the criticality of that level of precision. Do you?

Jeff

Its not important for the ears, its important for the mic.
post #8694 of 9577
Quote:
Originally Posted by jim19611961 View Post

Its not important for the ears, its important for the mic.

And I thought all the measuring and analyzing was to ultimately benefit our ears ... which have nowhere near +/- 1/4" locational precision. Whatever, I guess.

Jeff
post #8695 of 9577
Quote:
Originally Posted by pepar View Post

^^

I thought we did all that stuff based on geometry?

Jeff

Other things aside (e.g. whichever beer has better hearing) my speakers are very sensitive to toe-in. I was thinking of using that technique to see if I could improve my imaging.
post #8696 of 9577
This has been covered before. In this very thread in fact. I will do it ONE more time.

When waves from two different sources are trying to be measured from a single source (mic), if they are not perfectly in phase (timing, which equals distance in this case), they will interact. If the difference in phase is close to 180 degrees, then they will all but cancel each other causing the measurement to be way off (useless).

This shows up at high frequencies because their wavelengths are much shorter than bass frequencies. At 15K, the wavelength is less than 1". Being a 1/4" closer to one speaker and 1/4" further from the other would cause a difference of 1/2" (half the wavelength).

This topic has nothing to do with how we hear, where your ears actually are when listening or anal dribble. This has everything to do with acoustic physics and measuring with a single point source (the mic).
post #8697 of 9577
Quote:
Originally Posted by jim19611961 View Post

This has been covered before. In this very thread. I will do it ONE more time.

When waves from two different sources are trying to be measured from a single source (mic), if they are not perfectly in phase (timing, which equals distance in this case) will interact. If the difference in phase is close to 180 degrees, then they will all but cancel each other causing the measurement to be way off (useless).

This shows up at high frequencies because their wavelengths are much shorter than bass frequencies. At 15K, the wavelength is less than 1". Being a 1/4" closer to one speaker than the other would cause a difference of 1/2" (half the wavelength).

This topic has nothing to do with how we hear, where your ears actually are when listening or anal dribble. This has everything to do with acoustic physics and measuring with a single point source (the mic).

Jim, following up on this point, at what frequency level would say that the 1/4" difference in speaker/mic distance creates that level of wavelength distortion? IOW I might care less at 15 kHz than, say 5 kHz, but YMMV.
post #8698 of 9577
Quote:
Originally Posted by artur9 View Post

Other things aside (e.g. whichever beer has better hearing) my speakers are very sensitive to toe-in. I was thinking of using that technique to see if I could improve my imaging.

If you aim all spkrs, at least LCR, at the main LP, then that position is on-axis and imaging optimal. And if your surrounding seating positions are compact enough, they all fall within the spkrs' dispersion pattern.
post #8699 of 9577
Quote:
Originally Posted by sdrucker View Post

Jim, following up on this point, at what frequency level would say that the 1/4" difference in speaker/mic distance creates that level of wavelength distortion? IOW I might care less at 15 kHz than, say 5 kHz, but YMMV.

I am using the 1/4" criteria rather broadly here. At 15K, you probably need to be within 1/8" or less.

The thing is, if your on target at 15K (in phase), you will in phase at ALL the lower frequencies as well (given their wavelengths are longer).

At 5K, you have a little more room for error, sure.
post #8700 of 9577
Quote:
Originally Posted by jim19611961 View Post

This has been covered before. In this very thread in fact. I will do it ONE more time.

When waves from two different sources are trying to be measured from a single source (mic), if they are not perfectly in phase (timing, which equals distance in this case), they will interact. If the difference in phase is close to 180 degrees, then they will all but cancel each other causing the measurement to be way off (useless).

This shows up at high frequencies because their wavelengths are much shorter than bass frequencies. At 15K, the wavelength is less than 1". Being a 1/4" closer to one speaker and 1/4" further from the other would cause a difference of 1/2" (half the wavelength).

This topic has nothing to do with how we hear, where your ears actually are when listening or anal dribble. This has everything to do with acoustic physics and measuring with a single point source (the mic).

Regardless of where the mic is placed, there will always be frequencies that cancel, and isn't this the point of only measuring one main speaker at a time ... so there is no combing?

BTW, the single speaker under test is the "point source."
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