room response...

[DrWho]

Here's my room response, was wondering what y'all thought about it...

The system tested consists of custom rebuilt Marantz 940 speakers powered by a denon 1803 in a room with dimensions of 12x20x8 (WxDxH) feet. I used the ratshack spl meter (and a table of correction values) at my listening position 15 feet away in the center. The back wall pseudo opens up into the kitchen (there's a counter with cabinets above it, and 4ft of wall in the RR)

I originally set the volume so that 1k was 90db, then started at 12Hz and worked my way up and 1k measured to be 93db the second time! twas rather frustrating.

What do your guys' response graphs look like?

[DrWho]

I've also borrowed an excel file from fabulusfrankie and expanded it and revised it a bit to automatically construct the graphs and apply the correction values for the ratshack meter. however, my isp won't let me host it, and the forums won't let me upload it. Is there anyone capable and willing to host it? Or if you'd just like it for yourself, drop me an email.

-Mike

who@oldfarts.dk

[m00n]

What did you use for your test tones? How big is the excel file? You can always submit it in your post. You can post other files than just images.

[hurdy_gurdyman]

Here's graphs of my Heresy's in room response before and after tweeking, and my EV open baffles.

http://www.geocities.com/the_hurdy_gurdyman/Moreupdates.html

I have a program that makes the proper graph for speaker response, but I have to draw the response the hard way.

I'd be interested in having a copy of yours. Thanks.

the_hurdy_gurdyman@yahoo.com

Now, if someone could just explain to me how to post pictures directly in the posts it would make things a lot easier.

Dave

[DrWho]

I alread tried uploading the file to the forums, but they didn't like the file extension. It is 38.5kb in size (40kb on disc).

The test tones I used were posted by someone else on the forum and I just used those. That other person used the NCH Tone Generator (which is free, if you reinstall it each time the demo runs out). I also have the program, but was too lazy to use it when someone had already done the work for me. I then burned the .wav files onto an audio CD-R and played it back with my CD Player.

To get the image that you guys see there, I took a screen shot of the graph it drew and then applied some photoshop magic to make it a smaller file. If anyone knows how to export a graph to an image, or even print to an image in excel, let me know!

One thing I was wondering though...is it normal to have dips and peaks constantly varying by 3db both directions? I know a flat line is ideal, but what is considered "acceptable" for a "normal" room?

[WMcD]

Dear Dr.,

Yes, room response is typically very ragged. If you see data from a speaker test which fits in a 6 dB window it was done in anechoic conditions or gated, or smoothed a lot.

I use an LMS system which is pretty much an industry standard. Raw room response looks like a seismograph when the big one hits.

The first impression I got was that my speakers were junk or the measurement system was haywire. Not so.

On the other hand, I took a home made K-400 outdoors, put it on a ladder, and used a gate function. (This means the system listens to a very short burst and turns off before a bounce off the ground comes in.) It fit plus or minus 3 dB except for one data point. But in a room without gate-ing it looks very ragged.

You can hear the room effect if you feed a constant tone to a speaker and walk around the room. At some locations the sound drops out to almost silence. Where you stand makes a lot of difference. It is an ear opening experience which is difficult to believe unless you've done it.

The theory is that simple constructive interference only results in a 3 dB peak. This is where the reflection is in phase with the source. On the other hand, in destructive interference, where the two are 180 degrees out of phase, the result can be near total silence. Hence dips tend to be much deeper than peaks.

Of course, in a room, it is not that simple. Particularly at low frequencies, you can pick up about 10 dB. Corner placement helps this.

The theory may explain your graph. The dips above 1000 Hz may be simple interference. The peak below 50 Hz may be room gain in the bass.

As far as graphics, I don't work with Excel much. However, when I'm in a bind, I hit the print screen button on the keyboard. The image on the sceen gets sent to the clipboard or some register. Then I go to another program. MS paint is one. Then on the edit menu (paste) or Control V, the image gets inserted. Then it can be saved and converted to jpg format. Bitmaped images are too big.

This works well with old DOS programs. Run them in a window. Graphics can be exported this way.

I have not tried it but you should be able to export a spreadsheet as html. Fool with it a bit. Others on the bbs know more than I.

Best,

Gil

[hurdy_gurdyman]

If you are using sine waves like the NCH generator, you will get false readings because of standing waves building up. You need warble tones, which are too short in duration to build up standing waves. The Stereophile test disc two has a nice set of warble tones. The dips and peaks you are now getting will be a lot smaller with the warble tones, and will give you a more accurate idea of what your speakers are doing.

Dave

[Manuel Delaflor]

The response is a lot "jumpier" than mine, using a pair of Heresy's with a Tempest sub I have a fairly "flat" response from about 21Hz to 16Khz, only a nasty peak at 50Hz with I guess is a room mode but positioning and tunning the phase/volume of the sub couldn't resolve it.

Maybe this is because I choose the position of everything in the room to allow that (waf factor is away from my listening room!) and spent several days achieving it.

_____________

Edit:

I tried to upload a couple of graphs, no luck I will attempt it later.

[DrWho]

in regards to the peaks below 50Hz...that's partially due to another reason: the cabinets in which my speakers sit on (i really need to get some pics one of these days)...anyways, these speakers weigh about 90 lbs. each and are sitting on top of my cabinets about 1 foot from the cieling. the cabinets start to resonate at frequencies below 50Hz (mostly at 30Hz). It's a really cool effect and is amazing for HT, but does get a bit annoying with music sometimes. because my CD player is also in this cabinet, i've had to shock mount it because the vibrations cause it to skip...it also vibrates the floor too, because the cabinets themeselves are heavy too <> (it's fun to scare the dog with this, hehe)

The theory is that simple constructive interference only results in a 3 dB peak. This is where the reflection is in phase with the source. On the other hand, in destructive interference, where the two are 180 degrees out of phase, the result can be near total silence. Hence dips tend to be much deeper than peaks.

now this I find interesting...it makes sense and makes no sense at the same time

and regarding warble tones, what is the physical concept behind how that works? i would think that once a compression wave is created, that it would still move throughout the room just like any other wave...and wouldn't you need an RTA to measure the response? also, would that truly be a room response or isn't that more of a pseudo speaker response? what does the waveform of a warble wave even look like? hehe, sorry for so many questions...i've never heard of warble tones before.

I did wanna point out that I don't want to confuse room response with speaker response...i would think a room resonse is more representative of the sound one is hearing? my speakers are supposed to have a response +-3dB from 20Hz-20kHz, but that's obviously not true in my room

And when taking room nodes into consideration, how concerned should we get? in every system i've listened to, higher frequencies go from loud to off with even an inch of movement...lower frequencies are like 5ft. I would like to treat my room some, but these room nodes lead me to believe a larger different shaped room would have a more realistic impact.

[hurdy_gurdyman]

In regards to warble tones and sine waves. In real music you almost never get a continous tone for any duration. Music is constantly changing in frequency and intensity. It's difficult for standing waves to build up this way. A standing wave is a continous tone and, like feedback, can build when left un-checked. This doesn't happen with music. So, if you use sine waves to measure your room response, you are exiting the standing waves to the maximum in ways that music never would. Thus, your measurments show an absolute worse case that your speakers can sound instead of what they would normally sound with music.

Warble tones flunctuate up and down on a centered tone. There is not enough time for any frquency to build up. This gives a response much more like real music would. The resulting response graph you make will much closer relate to what you are actually hearing. All in all, with warble tones, you get a much more accurate look at what your speakers are really doing in your room with actuall music instead of artificial test tones. It works! The only thing better would be to invest in a real time frquency spectrum analyzer set up using pink noise (or is it white noise?). That's out of my price reach.

Dave

[DrWho]

ok, so where might one find warble tones to use then?

keep in mind that i'm looking for a free aproach here (so far, the only audio purchase i've ever made was the center channel and reciever and speaker repair kit...$500 total i think).

btw, what is wrong with tweaking the room so that even test tones lay out a flat response! cuz if that's the worse it could be, then i might be able to settle for that

I do know that I have an RTS laying around somewhere, but i misplaced the silly thing! it was a real nice one with a mic and everything all integrated into one box. at the time i got it (yes, for free) i had no idea what it did or how to use it so i stuffed it away somewhere!

also, i would still like to see other response graphs for those that have them, just for a reference. i look at my reponse and can't help but think there is a very simple pattern behind the peaks and dips that perhaps a simple solution might be in order. each speaker is a 4-way which causes me to wonder if there's not phase issues happening

and for dave...which response graph is your final "good" one? I was looking at them and was under the impression that the first graph was better? maybe i just read them wrong, i dunno.

[hurdy_gurdyman]

----------------

On 12/8/2003 6:00:32 PM DrWho wrote:

ok, so where might one find warble tones to use then?

keep in mind that i'm looking for a free aproach here (so far, the only audio purchase i've ever made was the center channel and reciever and speaker repair kit...$500 total i think).

btw, what is wrong with tweaking the room so that even test tones lay out a flat response! cuz if that's the worse it could be, then i might be able to settle for that

I do know that I have an RTS laying around somewhere, but i misplaced the silly thing! it was a real nice one with a mic and everything all integrated into one box. at the time i got it (yes, for free) i had no idea what it did or how to use it so i stuffed it away somewhere!

also, i would still like to see other response graphs for those that have them, just for a reference. i look at my reponse and can't help but think there is a very simple pattern behind the peaks and dips that perhaps a simple solution might be in order. each speaker is a 4-way which causes me to wonder if there's not phase issues happening

and for dave...which response graph is your final "good" one? I was looking at them and was under the impression that the first graph was better? maybe i just read them wrong, i dunno.

----------------

I had a friend copy the warble tones for me from a Stereophile Test Disc. You could say that was free. I don't know of any free downloads of warble tones.

As for my graphs, the first one was the stock Heresy. It was sitting on the floor and had a nasty dip at 400Hz because of floor cancelation. Also, I have changed the way I measure since the first graph comparison, so it isn't really a valid comparison. I used to have the meter firmly mounted on a tripod pointing between the two speakers. Now I rotate the meter and take the highest readings. Before the final speaker tweeking, using the new measuring technique, the peak at 63 kHz was up around 4 dB higher then it was after the tweeking. I built a notch filter and brought the peak down to acceptable range. This helped a lot in getting rid of a nasty sibilent sound that was annoying me. For some reason the P-trap didn't make any real difference, so I came up with my own solution by trying a few different things.

I think I'll remove the first Heresy graph, as well as do an update on the EV graph in the near future so it's not so misleading. I've just been lazy lately.

Dave