Wow. Seeing the effects of CrossO etc (I think?) on FR! (1 Viewer)

[Mary M S]

I have run my first inept attempts at charting FR. How cool is it to see what you hear!
But not so easy to interpret the results, from my miniscule knowledge base.

Some questions occurred. I ran the first charts noticing a mention made in posts to leave all DSP modes off. To do this I assumed if I ran the test in Stereo, through my pre this avoids the DSP modes. I also ran other graphs various ways: One: Main amp to speakers OFF, hoping this gives me just the sub output, as it is selfpowered.
Ran more in 6-channel bypass and so forth. Although I was not squeaky -clean-careful jotting down my numbers (and was interrupted by people coming and going so I had to start and stop repeatedly. I see enough of a base-line-trend to know I was not far off in taking down and converting the numbers fairly correctly.

Some things jumped out at me, right away. Regarding two graphs run in Stereo mode.
When a total newbie and setting up my pre, I saw the rule of thumb by THX etc for cross-overs to be set at 80hz. However, going only by my own fallible ears, 80 did not sound pleasing in my room so I set my Frnt/Cent/Surrounds at 60/60/100. To me, this sounded better.

When looking at the graph of FR in Stereo mode which should give me my FL/FR and sub. First run I get this.
60/60/100
EDIT: SEE LINK TO GRAPH POST # 9

So I look at the line graph of this and think I have some null? Valleys and peaks, I wonder if switching my crossover will smooth the line. I run it again stereo mode at 80hz for the fronts. Getting wider swings in my peaks and valleys.
EDIT: SEE LINK TO GRAPH POST #9
It appears to me as if I did get a smoother response down the line when my crossover was set at 60 for the FL/FR. When at 80 I get a steep dip bottoming near 71hz. I also see I have some things going on above 100hz, and not sure what to make of that.

By coming back in here yesterday and doing some search functions using different words, I keep running into little bits of information, to add to my whole. I asked what dB I should run the tests at and was told keep it in the 80’s and you should be fine. I erred on the side of caution, (since I have never done this before) and set my meter to read at 70dB. (the level these numbers are based to). I now see doing some more reading that 70db was most likely too low. I was worried about transient spikes knowing the RS meter is not sensitive at certain freq., so I kept the vol. way down. I see from my reading I should bump that level and rerun the tests, but wondered if what I am seeing on these two graphs is what I think it is, crossover related differences?

It seems I have read that rooms often “hump” around 50hz. I see sign of this in these two charts with the hump again much smoother when crossover for mains is at 60. However, when running these tests in other modes, 6-chan analog, PLII etc, the hump with its high centered at 50 has a larger rise .

Any instructions on if I headed the right way. (how I ran the graphs) and what I need to look for.
Looking at the smoother response graph (crossover at 60) I need to be concerned with Freq at 56 and whatever is going on in the 100 to 160 range. Correct?

Any thoughts instruction is greatly appreciated, but will completely understand if those who know, are too bored to tears diagnosing other (call us clueless ) posters trying to make sense of their own graphs!

 

[John S]

I need to run a similar tests at my primo listenign position. I did extensive testing when I first brought my current system on-line, but did them at 1 meter 1 channel driven, because I just wanted to test the equipment.

I'd expect my living room (No dedicated theater til next year), to be really poor, even though the end resulting performance I have now is pretty impressive.

 

[Lee Carbray]

Try playing around with phase settings and see if that helps the dip as ~56Hz.

Edit-That is with the crossover at 60Hz.

 

[Kevin C Brown]

Mary- Those numbers are ... too many numbers for me to digest.

A suggesion: go ahead and graph them, do a screen shot to a jpg file (then you can edit the size of the screen shot just down to the graph in MS Photo Editor), and then open an account on www.villagephotos.com then we can all see the graphs.

Kind of like this:

http://img.villagephotos.com/p/2004-...freqStryke.jpg

Or you can embed it too:



I similarly use 60/80/100 Hz for the crossovers for fronts/center/surrounds&rears. If you look at my graph, I have a dip at 50 Hz which I first though was a phase problem with the fronts and the sub. But no matter what I did with phase, that dip didn't go away. Must just be some node I'm sitting in...

 

[JamesDB]

Hi Mary,

Were you planning on using some sort of active EQing or just room treatment and poistioning changes?

J

 

[Mary M S]

uh excuse me Kevin [digs toe in dirt and looks up :b ]
do a screen shot to a jpg file (then you can edit the size of the screen shot just down to the graph in MS Photo Editor) or embed it

[puts hands on hips embarssed to admit it] now just what the hell does that all mean. which would be simplest for the computer challenged? (I’ve never posted images before) Step by tiny step? or maybe a tutorial you could point me at somewhere???


Were you planning on using some sort of active EQing or just room treatment and poistioning changes

No positioning changes are possible, other than moving mains by inches and pointing surrounds. The sub only has one possible location and is corner placed behind display. Every few months I get serious and quit just playing at HTF and try to really learn something new. My husband just called me a Nerd this morning , when he saw me studying the graphs I had printed out last night (wish you could see them Kevin ) (but I did not know how to plug them into my post above), he went on querying me what kind of an EQ device was I thinking about. (I smell him making a present list)!

Right now I’m primarily just trying to learn to plot correctly, See the effects of crossovers (and thank you Lee! now Phase) to see what I have been listening to and how with minor tools at hand I can improve it. (One asset I have left unused at neutral, is the QEQ(?) on the SVS sub still available. Left it alone at setup of sub, for lack of understanding it.
Since for me, one thing leads to another I might end by looking for some sort of EQ device but not till I learn, much much more.

I apologize for my tabs; first post. I don’t know what happened to them. They were in perfect rows in Word, till I pasted in HTF.

 

[Kevin C Brown]

Mary- OK, this is what I've found to work:

Set up Excel so that you have your freq response graph looking the way that you want it. Make it as big on the screen as you can. Now hit the keys Ctrl-PrintScreen, then separately Shift-PrintScreen each a few times. (I don't know why, but I don't know which set works, and sometimes it doesn't work the first time anyway.) So now, you have the screen in the buffer.

Fire up Microsoft's Photo Editor. Sometimes under the "Microsoft Office Tools" folder in the Start menu. Now go to the Edit menu, and "paste as new image" should be highlighted, click on that. Now, you should have a copy of the screen pasted into Photo Editor. Now you can use the dashed rectangle to the left of the magnifying glass to "click and drag" around just the portion of the graph that you want the picture of. Now you can "cut" that portion, or "copy" it. (For some reason, "cutting" it sometimes doesn't work, but "copying" always does.) Now go back to the edit menu, and again paste as new image. You should now have just the graph in Photo Editor. Go to the File menu, and "Save as", and now you should have a JPG file of your freq response!

Try that far and see what happens. The next step would be setting up an account for yourself on www.villagephotos.com (or any other image hosting site), upload the file, and copying the url into HTF. See how far you get and check back.

 

[JamesDB]

Hi Mary,

I decided to plot your graphs on excel out just to take a quick look. I wish that you could have posted data checking at each integer Hz 20-120 since its hard to tell what some of these dips may be. It could be a room mode which you might be able to ascertain or perhaps even confirm with one of the simple online calculators. Or it could be destructive interference between the 3 speakers near the cross over wavelength (or maybe both).

If its the latter perhaps moving the speaker locations slightly may help in correcting this or at least move the dip to somewhere more manageable. Otherwise, if its a mode, though not as easy to correct a dip completely you can probably get a bit of an increase by EQ (3-6db without much amp stress, better than nothing).

Its interesting to me that you are also getting an increase at the 20Hz frequency. I think many of us would love this to happen so naturally. Perhaps a parametric EQ could help you at least smooth (cut maybe 5db)out the 80-100 range. A bit of a bump in the 50Hz range may not be so bad since to some it sounds a bit better for movies (though I can't stand it for music). When I first got my parametric EQ I tried it correcting ruler flat and it somehow sounded too dry and sterile. I like a bit of a natural house curve with a few small "errors" in response.

I also had tried moving my sub to different locations, but I would first focus an making sure you have good integration with your mains before worrying about EQ. Even though 80Hz and under is not that directional I can still hear it if lets say the sub is beside me and the speakers are up front.

Also I had heard, but maybe you need to confirm this as I know more about pro audio than home, that running different crossover frequencies for your 5 speakers can lead to time/energy problems not correctable by EQ. Chu Gai would probably know more about this.

J

 

[Mary M S]

I am exhausted. Kevin you are The Man. An argument with SBC over getting my free home page up [4-hours, gave up]. Realizing I had no photo editor loaded from Office,- finally finding “Paint” to convert file to jpg [2-hours]
And on and on. But TA-DA!!!!

80 Vs 60 X-Over Graph

Please I hope this works.

 

[Mary M S]

different crossover frequencies for your 5 speakers can lead to time/energy problems not correctable by EQ

Good point James, I am vagely aware that might be an issue (diffrent x-o to time align) When I left my system the last months at 60/60/100 it was because (pre-FR chart) it sounded better to my ear than 80/80/80.

Lastly when I went to 60/60 on FL/FR and Center but then at that time decided to bump the surrounds to X-Over @ 100. It was because the surrounds were irrating me. They are little Def Tech Pro Cinema 100's. They just did not appear to handle a broader freq range very well. (to my ear) Particulary not at the volumes we sometimes run this system at.

ps. James how very nice of you to plot them yourself so that you could see them!

 

[Kevin C Brown]

Mary- Awesome! Kind of neat that you have a rising response at 16 Hz. Yup, Paint works too.

See, your curve is an example where I don't think a BFD is going to do much good. I.e., some dips, but not too many overt peaks to mash down. That's why I am not using mine at the moment either.

The 60 Hz curve does look better than the 80 Hz curve. You have optimized the phase between the mains and sub? (Does your sub have a continuous phase knob, or a 0/180 deg switch?)

Nice!

 

[Mary M S]

The 60 Hz curve does look better than the 80 Hz curve. You have optimized the phase between the mains and sub? (Does your sub have a continuous phase knob, or a 0/180 deg switch?)

I think continuous, but will have to check. I had more help when setting up sub (than normal) because a new Subw is exciting. My husband stayed behind the display since he considers my ear more sensitive and rotated phase till I heard the strongest bass at seating. I did not use the SPL meter, only ear, since all the guys were in a hurry to throw some DVD's at the SVS.

I'm still sorting terms like house curve, which I had BACKWARDS (from reading the term, I thought rising towards the higher freq) till I finally spotted posters graphs and realized the rise should be gradual in direction towards the lower freq.

If I had the tool of equalization (do not for now), I guess per results of the 60 X-over graph, would try (if possible) smooth the dips at 18,56 and 111, bringing down the 80-100 range and 125. Right now I will cont. to see what X-O' s phase etc do.

Just realized that the QEQ on the SV is only one-band (I had thought it was two) and I'm not sure what freq. range it choice of application lies in, if any of my current peaks listed above do fall within its zone, will try equalization at one problem area.. (time to get out all my manuals)

You have to realize that this is like a whole new world to me (to see the FR graphs). I have set things up by ear, other than speaker calibration, by reading recommendations in the past, and felt by ear that I heard differences between some choices as when I reverted to a 60hz X-over. But now I can see whats going on, which (for some reason)am very excited, (another tool available)

For now just looking at these charts (sub/mains). I will run the tests, at a higher dB (to see what impact that might have in FR Or if the line peaks and nulls locations, track the same - at the higher dB. I'm curious to see if charting at a higher dB this does change the response as regards which freq, the current peaks/nulls lay at.
I will play with phase, and if there are any recommendations for using the one band available on the sub, - will see if that effects.

After recovered from getting this far, (not the least of the labor involved just posting my graph ) intend to plot more. When having a long slow weekend wamt to change every parameter, available, to see what (if any) differences. Trying the surround crossovers at 60/60/80 vs 100 etc, in 6-speaker response, on so forth.

Sorry to go on, while I'm sure this is nothing for all of you, (graphing and interpreting FR) it feels like a watershed for me, ever so slowly adding to my better understanding sound!

ps: I forgot ports! if memory serves (the sub is behind a display) I currently have one port plugged tuned to 16. Could be fun to see what diffrence porting might have on my room r.

 

[Kevin C Brown]

Good stuff! This is *one* way of doing it. A lot of people have come to the conclusion that a different ("best") way, is to vary the phase knob for the flattest freq response through the crossover. That takes more time with plots and stuff though.

One thing to perhaps keep in mind, is manually plotting points is a real drag and can take a lot of time. How would you like to do what you are doing now, but in 5 seconds? (Makes perfecting phase *a lot* easier.) All you need is a PC within cable distance of your system, and ...

www.etfacoustic.com

It is a complicated piece of software, but ... the demo is free, and the help files are really quite good.