Jump to content



Sign up for a free account to remove the pop-up ads

Signing up for an account is fast and free. As a member you can join in the conversation, enter contests and remove the pop-up ads that guests get. Click here to create your free account.


Photo

BFD EQ with ETF - sub-mains integration


  • You cannot start a new topic
  • Please log in to reply
33 replies to this topic

#1 of 34 OFFLINE   Ranga

Ranga

    Stunt Coordinator



  • 109 posts
  • Join Date: Nov 30 2000

Posted April 15 2002 - 05:00 PM

Just when I thought that I had finished my EQing experiment, I upgraded my SVS driver and had to start all over again. I managed to get a good house curve (except for a null at 32-35Hz). In fact, with the old driver, my nulls could be boosted for a nice house curve. However, with the new driver, boosting it more than 3db does not seem to help. My second part of the exercise was to measure the overall frequency response of both the mains and the sub. Since there was no way of using the ETF software (as one of my crossovers is controlled by the DVD player), I used the Stryke Bass CD to plot the frequency response for two crossovers - 80Hz - set on the pre-amp fed by an external DAC. 100Hz - set on the DVD player (speakers set to SMALL) With either of these crossovers, my frequency response measurements are all over the place. The nice house curve is gone!. How do I get both the mains (L+R) and the sub EQed using ETF with the 80Hz crossover? I don't think I can use ETF to measure the frequency response with the 100 Hz crossover unless I burn a test CD. Suggestions?

#2 of 34 OFFLINE   Chris PC

Chris PC

    Producer



  • 3,994 posts
  • Join Date: May 12 2001

Posted April 15 2002 - 06:16 PM

Burn a test CD. Sorry. I had to say it cause you said "unless I burn a test CD"



I used an amazing program called Sound Forge 5 by Sonic Foundry and it is just amazing what you can do with that. you can make a CD with tones from 1 hz to 200 hz in .1 hz increments. Of course, you'd have to merge some tracks since each CD is limited to 99 tracks, but anyways Posted Image i say burn the CD and use the ETF program.
Going from projector to flatscreen for a while.... :P

#3 of 34 OFFLINE   Rick Radford

Rick Radford

    Supporting Actor



  • 650 posts
  • Join Date: May 12 2001

Posted April 16 2002 - 02:43 AM

Iirc, ETF has to use MLS signals and cannot use a test tone CD. I don't know why.. but that's what I recall the author of the program telling me. Ah yes.. here's the quote: "You must use the ETF test signal CD with the ETF program. It is much better for this type of measurement than any test CD, in addition to this, a standard test CD will not work" Note: I have not tried a test tone CD to verify this.
--RR

#4 of 34 OFFLINE   BruceD

BruceD

    Screenwriter



  • 1,223 posts
  • Join Date: Apr 12 1999

Posted April 16 2002 - 08:24 AM

Ranga, Can I assume your PC with the ETF software only has a half-duplex soundcard? If it has a full-duplex soundcard, you don't need to use a CD at all (the test signals are generated by the program itself). If you must run in half-duplex mode than you must download the test files for a CD-R from the ETF website.

#5 of 34 OFFLINE   Chris PC

Chris PC

    Producer



  • 3,994 posts
  • Join Date: May 12 2001

Posted April 16 2002 - 08:25 AM

Bummer. Oh well. Do you have the ETF test signal CD to use with the program? Go that route if you can Posted Image
Going from projector to flatscreen for a while.... :P

#6 of 34 OFFLINE   Ranga

Ranga

    Stunt Coordinator



  • 109 posts
  • Join Date: Nov 30 2000

Posted April 16 2002 - 10:22 AM

BruceD My laptop has a full-duplex card and I can use ETF to plot the frequency response of the sub alone without any crossover. In fact, I managed to have a nice house curve for the sub. However, when I connected the output from the PC to either my left or right mains and used the fixed 80Hz crossover on my pre-amp, I lose the house curve. Also, my left and right mains have different in-room frequency response due to their placement (my right speaker has a sidewall while the left does not). I am going to try connecting a Y cable to the PC output going to the CD input on the pre-amp. Will this engage the crossover on the pre-amp? It does engage when I use the CD-input and use stereo direct when playing a CD. For DVDs (and surround processing), I am using the bass mgmt on my DVD player which apparently has a fixed 100Hz crossover. According the Jeff Hipps (Sherwood), the bass mgmt is bypassed when playing CDs (stereo), but I find that NOT to be the case. Assuming that the 100Hz crossover is being used, there is no way currently that I can measure the frequency response using ETF. So, I may have to use the ETF test signals by burning them on to a test CD. I know that both you Bruces (brucek and BruceD) recommend measuring the sub alone, but how then do you fix the mains integration with the sub without re-EQing?

#7 of 34 OFFLINE   BruceD

BruceD

    Screenwriter



  • 1,223 posts
  • Join Date: Apr 12 1999

Posted April 16 2002 - 11:11 AM

I would recommend the following: 1) capture the sub by driving it directly from the PC output port on a single cable, no xovers engaged with the ETF bass frequency impulse (20Hz-200Hz) selected. 2) overlay (on the sub graph) and capture the Left Main Speaker by driving it directly from the PC output port on a single cable, no xovers engaged with the ETF bass frequency impulse (20Hz-200Hz) selected. 3) overlay (on the sub graph) and capture the Right Main Speaker by driving it directly from the PC output port on a single cable, no xovers engaged with the ETF bass frequency impulse (20Hz-200Hz) selected. This gives you the baseline for in-room response of each speaker. You can take measurements by driving each Main speaker individually through the preamp (@80Hz xover) from the PC output to test the effect of the xover on each main speaker's frequency response. Yes, the DVD player seems to offer a problem I don't have a solution for. The object with EQ on the Low-pass bass side (below the xover) is to smooth out the in-room bass response. You primarily do this based on the sub's output, as it probably dominates the bass SPL compared to the mains.

#8 of 34 OFFLINE   Ranga

Ranga

    Stunt Coordinator



  • 109 posts
  • Join Date: Nov 30 2000

Posted April 16 2002 - 11:41 AM

BruceD



Great suggestions.



Just a thought - since I have only two choices of Xovers with my current setup - 80Hz and 100Hz, wouldn't it be easier to EQ by connecting the PC output through a Y cable to the CD input on the pre-amp and activating the 80Hz Xover?



What additional benefit (other than useful info) would I get by measuring the response individually for all speakers (incl the sub)?





[quote]

The object with EQ on the Low-pass bass side (below the xover) is to smooth out the in-room bass response. You primarily do this based on the sub's output, as it probably dominates the bass SPL compared to the mains.

[quote]



From the freq response graphs that I measured (without ETF) using the Stryke CD, it seems to follow a predictable pattern -



Humps at 40-50Hz, 80Hz, and 100-110Hz.

Nulls at 32-36Hz, 63Hz.



The only change is the higher output levels between 22-30Hz when I use the sub.



If I EQ just the sub, the hump at 100-110Hz gets more prominent when measuring the freq response (mains+sub) and is difficult to EQ.



I shall post my graphs after following your suggestions.

#9 of 34 OFFLINE   BruceD

BruceD

    Screenwriter



  • 1,223 posts
  • Join Date: Apr 12 1999

Posted April 16 2002 - 02:52 PM

The reason you want to measure each speaker independantly is because multiple speakers will likely interact with each other and either cause their own peaks and valleys or increase peaks and valleys that are already there. You might try slightly moving the main speakers to get rid of the 100-110Hz peaks, but I'm sure cutting the 40-50Hz peaks with the EQ will also help reduce the 80Hz and 100-110Hz peaks (which are likely to also be harmonics of the 40-50Hz peaks). You should always try to reduce the lowest frequency peaks first, because the peaks at higher frequencies (harmonic multiples of this lower frequency) will also be reduced.

#10 of 34 OFFLINE   Ranga

Ranga

    Stunt Coordinator



  • 109 posts
  • Join Date: Nov 30 2000

Posted April 16 2002 - 03:37 PM

BruceD



[quote]

You might try slightly moving the main speakers to get rid of the 100-110Hz peaks, but I'm sure cutting the 40-50Hz peaks with the EQ will also help reduce the 80Hz and 100-110Hz peaks (which are likely to also be harmonics of the 40-50Hz peaks).

[quote]



Well, EQing the 40-50Hz peak boosted the 100-110Hz peaks. On the other hand, EQing the 100-110Hz boosted my output below 25 Hz.



Also, when I start off with the humps at the lower frequencies, by the time I get to the higher frequencies, I have new humps close to the previously fixed humps. I am not using BANDWIDTH of more than 1/6 octave (10/60).



Anyway, I'll post the new measurements tomorrow for a better reading.



Thanks.

#11 of 34 OFFLINE   Ranga

Ranga

    Stunt Coordinator



  • 109 posts
  • Join Date: Nov 30 2000

Posted April 17 2002 - 10:19 AM

BruceD



Here is my frequency response (corrected SPLs) graph for the following -



sub - black

left speaker - orange

right speaker - red



Posted Image



Although the frequency response for the mains seems to extend beyond the -3db point (55Hz), I see that the slope in ETF starts aroung 40-45Hz. Also, I had to keep the sub's volume quite low to balance the two channels in ETF.



I also tried feeding the PC output from ETF to both the left and right fronts with the 80Hz crossover set, but it appears that the crossover is not active as I do not get any output from the sub. However, when I play music or the Stryke CD, I do get output at the sub.



How do I now go about EQing this setup with ETF? Are you saying that I shoud not worry about the output of the mains in the region above 100Hz and just EQ the the sub's frequency response?

#12 of 34 OFFLINE   Wayne A. Pflughaupt

Wayne A. Pflughaupt

    Producer



  • 5,965 posts
  • Join Date: Aug 05 1999
  • Real Name:Wayne
  • LocationKaty, TX

Posted April 17 2002 - 01:19 PM

Bruce,



Remember that recent thread about full-range EQing where I commented that the L/R mains often exhibit response deviations? Posted Image



I’d sure be interested in seeing Ranga’s curve above 160Hz!



Regards,

Wayne A. Pflughaupt

My Equipment List
“A nice mid-fi system,” according to an audiophile acquaintance.

My Tech / DIY Articles and Reviews

#13 of 34 OFFLINE   Ranga

Ranga

    Stunt Coordinator



  • 109 posts
  • Join Date: Nov 30 2000

Posted April 17 2002 - 02:10 PM

Wayne



I had an earlier graph (No ETF) with measurements upto 1K using the Stereophile Test CD. The X-axis values are different and I did not use a tripod to measure the response.



However, it gives you an idea of the in-room response which looks pretty similar to my previous graph.



Posted Image

#14 of 34 OFFLINE   Manuel Delaflor

Manuel Delaflor

    Supporting Actor



  • 660 posts
  • Join Date: May 25 2001

Posted April 17 2002 - 02:56 PM

Calibrating a sub is indeed a big problem. I have more or less the same results using my laptop with ETF5 and my RS SPL Meter. Big jumps and deeps along the spectrum. The only way I can "correct it" is by putting my sub between the mains, but, of course, I loss extension doing that...
Here was my gear info... Now is on my profile, in case you want to know

#15 of 34 OFFLINE   Ranga

Ranga

    Stunt Coordinator



  • 109 posts
  • Join Date: Nov 30 2000

Posted April 17 2002 - 03:07 PM

I guess the idea is to eliminate the peaks and valleys above 100Hz (the max crossover point) by repositioning the mains and then position the sub in a corner farthest away from openings in your room. After the placement, EQing the sub 100Hz response should result in a good response curve. However, the situation becomes a little complex if your mains are not symmetrically placed - i.e. if you have a side wall near one of the mains and an opening on the other side. In this case, it may be impossible to get placement accurate for the mains. The question then would be - how do you include the mains frequency response when EQing so that you get it flat or have a house curve for the whole system (L+R+sub)?

#16 of 34 OFFLINE   Wayne A. Pflughaupt

Wayne A. Pflughaupt

    Producer



  • 5,965 posts
  • Join Date: Aug 05 1999
  • Real Name:Wayne
  • LocationKaty, TX

Posted April 17 2002 - 05:30 PM

Manuel, [quote]

The question then would be - how do you include the mains frequency response when EQing so that you get it flat or have a house curve for the whole system (L+R+sub)?

[quote]The mains generally need limited house curve, too. In my system 50% of the total curve takes place above 100Hz. The main’s curve can usually be achieved with the receiver’s tone controls and/or speaker placement. A lot of audiophiles like to pull their speakers out from the back wall; there are benefits in this practice, but it tends to reduce the speaker’s bass response. As I’ve noted in other posts, if the mains already sound warm with adequate bass response (taking into account the effects of the crossover) then you already have the house curve in place.



To answer your question, Ranga, set up your house curves in the mains and sub separately. First get the mains to where they sound good full-range without the sub. Then EQ the sub separately, with crossover engaged. Final house curve tweaking should be with the subs and mains running together with any and all crossovers engaged.



Regards,

Wayne A. Pflughaupt

My Equipment List
“A nice mid-fi system,” according to an audiophile acquaintance.

My Tech / DIY Articles and Reviews

#17 of 34 OFFLINE   Ranga

Ranga

    Stunt Coordinator



  • 109 posts
  • Join Date: Nov 30 2000

Posted April 17 2002 - 09:31 PM

[quote]

Final house curve tweaking should be with the subs and mains running together with any and all crossovers engaged.

[quote]



Does the EQ come into the picture here? At this stage, I probably would need to use the Test CD again and manually tweak the response and re-adjust the EQ. Correct?



While the ETF software makes it a breeze to monitor and EQ the frequency response, it seems that I may not be able to utilize it fully for the kind of measurements and correction that I need to make. BruceD/Brucek??

#18 of 34 OFFLINE   Ranga

Ranga

    Stunt Coordinator



  • 109 posts
  • Join Date: Nov 30 2000

Posted April 18 2002 - 10:54 AM

Wayne/BruceD ??Posted Image

#19 of 34 OFFLINE   Wayne A. Pflughaupt

Wayne A. Pflughaupt

    Producer



  • 5,965 posts
  • Join Date: Aug 05 1999
  • Real Name:Wayne
  • LocationKaty, TX

Posted April 18 2002 - 10:56 AM

Ranga,[quote]

Does the EQ come into the picture here? At this stage, I probably would need to use the Test CD again and manually tweak the response and re-adjust the EQ. Correct?

[quote] Of course, part of the house curve equation is simply adjusting sub/main levels. Beyond that, this is difficult to answer because everyone’s situation is different. I’ve seen people post pre-EQ response curves where they would pretty much have their house curve in place just by smoothing response. Other’s unequalized response was such that they needed to first flatten response, then apply a separate, specially set-up filter to induce the house curve (this is what brucek did). Still others are between these two extremes.



Judging from your graphs, Ranga, it looks like you would fall in the “first get flat response” category. Brucek has some helpful information on setting up a filter to induce a house curve to flat response – I’m sure he’ll be happy to pass it on to you when the time comes.



I’ll defer the ETF questions to the capable “Bruces.”



Regards,

Wayne A. Pflughaupt

My Equipment List
“A nice mid-fi system,” according to an audiophile acquaintance.

My Tech / DIY Articles and Reviews

#20 of 34 OFFLINE   Ranga

Ranga

    Stunt Coordinator



  • 109 posts
  • Join Date: Nov 30 2000

Posted April 18 2002 - 11:02 AM

Thanks Wayne.




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users