NEED HELP: Filter Adjustment

[media2368]

I tried couple of hours to adjust filters but didn't have a good result of flattening the measurement. Can anyone here help me on this? It really making mad and tiring..

My measurement file can be downloaded from here.

[Wayne A. Pflughaupt]

It doesn’t look too bad, really. All those little ups and downs probably won’t be that noticeable, especially with movies.

The main problem I see is that your Target Curve was really low to start with (75 dB is typical), and you didn’t re-align the Curve before equalizing (I would have pushed it up to ~80 dB). What you’d really doing, with the Target set so low, is using the equalizer as a level adjuster – not the best use for one.

Regards,
Wayne

[brucek]

Redo your soundcard calibration. You must have some effect turned on. Note how it's all wavy...
And as Wayne points out - you need to measure and create filters with a 75dB target level..

brucek

[media2368]

Thanks Wayne and Brucek.

Actually, I measured at 80db but would like to target it at 70db...if there anything with my targeting???

[brucek]

Measure at 75dB and set the target level at 75dB.....................

[JohnM]

media2368 wrote:

Actually, I measured at 80db but would like to target it at 70db...if there anything with my targeting???

That's not the purpose of the target level - you would end up trying to use EQ filters as a subwoofer volume control.

The aim of the EQ filters is to counter the contributions of room resonances, to find correct settings for the filters we need to know (or estimate) what the level of the measurement would have been without the resonances, so that we can identify the contribution of the resonances and work out the filter settings to counter them. If the target level is too low the filters would over-correct, if too high they would under-correct. It is important to get the target level right for the best results.

[hasse_swe]

hi could Bruce? wayne? john? or any one else of the pro's answer this...
like this "edia2368's graph in this post for instance.....

whould it be better lo let REW auto adjust the peaks/filter(s) the first 1-2 runs?
like this for example :

first run:
1.adjust sub to 75db (or 85 or 105) ref level
2.use the "auto" set target level in REW
3.Find filter peaks
4.set filters
5.Optimize+adj. gains
6.send filter(s) to bfd

second run:
1.adjust sub back to 75db (or 85 or 105) ref level (since its lower after first run)
2.use the "auto" set target level
3.Find filter peaks again (this time REW probably finds more peaks)
5.set filters
6.Optimize+adj. gains
7.Put in the filter(s) in eq setting from previous run to this run manually
8.send all filter(s) to bfd
9.adjust sub back to ref level again

then make minor adjustements if needed....?
whould this be wrong way to go?
thanks/Hasse

[brucek]

I think the best way is to take a measure at 75dB and let REW recommend its filters with an appropriate target set. Then you can do a little manual tweaking on the screen if you like until you end up with the corrected response you want on the REW screen. The actual end result after entering those filters into the BFD will be very close to the corrected graph on the REW screen.

brucek

[Wayne A. Pflughaupt]

hasse_swe wrote:

hi could Bruce? wayne? john? or any one else of the pro's answer this...
like this "edia2368's graph in this post for instance.....

whould it be better lo let REW auto adjust the peaks/filter(s) the first 1-2 runs?
like this for example :

first run:
1.adjust sub to 75db (or 85 or 105) ref level
2.use the "auto" set target level in REW
3.Find filter peaks
4.set filters
5.Optimize+adj. gains
6.send filter(s) to bfd

second run:
1.adjust sub back to 75db (or 85 or 105) ref level (since its lower after first run)
2.use the "auto" set target level
3.Find filter peaks again (this time REW probably finds more peaks)
5.set filters
6.Optimize+adj. gains
7.Put in the filter(s) in eq setting from previous run to this run manually
8.send all filter(s) to bfd
9.adjust sub back to ref level again

then make minor adjustements if needed....?
whould this be wrong way to go?
thanks/Hasse

To expand a bit on brucek's advice, the purpose of a second run would be to see how well the EQ filters worked out in “real life” as compared to how REW predicted, not to see if or where “new” filters should be applied. If things didn’t quite turn out right with the first set of filters, you don’t want to “pile” new filters on top of the old ones. It’s better to go back and re-adjust the first filters to get the desired results. Make sense?

Regards,
Wayne

[hasse_swe]

Wayne A. Pflughaupt wrote:

To expand a bit on brucek's advice, the purpose of a second run would be to see how well the EQ filters worked out in “real life” as compared to how REW predicted, not to see if or where “new” filters should be applied. If things didn’t quite turn out right with the first set of filters, you don’t want to “pile” new filters on top of the old ones. It’s better to go back and re-adjust the first filters to get the desired results. Make sense?

Regards,
Wayne

yes makes sense...
but for instance me who have an huuuge peak at 50hz..if i just make an auto-set-filter run (with the auto target level rew sets)..REW will only see that peak and don't see rest peaks/dips i have that is atleast +-10db (which are located at much more lower level) ....those will rew take care of if make a second run 2....
but if manually lower the target level after first run and then find filters it will be ok of course...
/H

[brucek]

REW will only see that peak and don't see rest peaks/dips i have that is atleast +-10db (which are located at much more lower level) ....

That's fine. Simply tweak REW after the auto find of filters is complete. You can manually adjust the filters to your hearts content until the corrected waveform looks correct. Enter the filters and your done...........

brucek