This should be interesting [DEQX by two]

[Mr. Widget]

I do have a question, though. The Deqx site mentions speaker correction, room correction, and media correction. For room correction, it mentions up to 10 bands of graphic or parametric EQ, and apparently less for media correction, but I couldn't see where speaker correction might also have a specific filter selection.

There are only ten bands available for room and CD/media "correction". These ten bands can be used as a simple graphic equalizer (I can't imagine why anyone would ever do that... unless they simply can't get their head around the use of a parametric.) As an example of how powerful the DEQX parametric filters are, only two of the parametric bands are all that are required to duplicate the RIAA phono curve with an accuracy that is higher than most phono stages.

Anyway, you can access three of these bands and modify and store them for tweaking the sound of up to 99 different "room curves" or "CD curves"... these are all accessible via the remote control... or you can access all ten bands and use them to "correct" your room.

In addition to these parametric filters there are additional FIR filters that are used to "correct" your speakers. These filters are separate and are automatically generated based on your speaker measurements. Here are the correction filters that DEQX created for Project May. It is easy to see the mid band correction is roughly the mirror image of the curve supplied us from JBL (the band between the 800Hz and 10KHz crossover points). You can also see the DEQX "correction" filtering balancing out the naturally rising curve of the 1500ALs. (Below 800Hz)

For those who don't understand what this graph shows... it is not the response of the speakers, it is the electrical boost and cut created by the DEQX unit to balance the response of the speakers. Therefore when this curve goes up, that is where the speakers originally had too little output and where the curve goes down, the speakers originally had too much output.


Widget

[johnaec]

In addition to these parametric filters there are additional FIR filters that are used to "correct" your speakers. These filters are separate and are automatically generated based on your speaker measurements.

That was kinda' what my question was about; you've clarified that the speaker correction filters are a different set. I assume they're parametric, (as opposed to numerous fixed "graphic" filters that interact), but do you know how many bands of "automatic" parametric are available for this mode? I assume there must be a maximum number, (which is undoubtedly more than would likely ever be needed). Not that it would make any kind of difference in the real world - I'm just curious...

Or are these FIR, (??), filters a completely different animal altogether, (not band or frequency based, but rather curves somehow extracted mathematically from the speaker response)? I'll also try reading up some more...

John

[Mr. Widget]

I have no idea how the filters work. As you can see from those curves, you can not create such specific "corrections" with a handful or parametric or even dozens of graphic filters.


Widget

[johnaec]

I mqay have missed them, but have you posted CLIO curves of your system with the Deqx inline? (I'll try a search later.)

John

[Ken Pachkowsky]

If I am reading your question correctly it would be a maximum of 10 both graphic and parametric.

This is the first day I have been able to spend really using this and it is truly amazing. The fine tuning flexibility is incredible. That hollow sound I spoke of earlier is history. It just required a fine tuning of the filters.

Widget could be right. He told me if I were patient I would probably really like this unit..... .

Ken

[Ken Pachkowsky]

These are the current filters.

[Mr. Widget]

If I am reading your question correctly it would be a maximum of 10 both graphic and parametric.

Nope. There are ten filters, period. You essentially get to choose how you use them... if you choose graphic, the Qs are set to a fixed value with an even distribution of frequency centers... and then you have a graphic.

Widget could be right. He told me if I were patient I would probably really like this unit..... .

I think I said that if you could get it set up properly, you'd probably like it. Patience is certainly required, but so is a careful methodological implementation. Good to hear you are starting to get it to work for you.


Widget

[Ken Pachkowsky]

Nope. There are ten filters, period.

Widget

Got to be carefull how you word things.....yes 10 only, graphic or parametric is what I was saying.

Ken

[johnaec]

Got to be carefull how you word things.....yes 10 only, graphic or parametric is what I was saying.

I take it that's in addition to whatever the Deqx uses for its custom speaker correction filters?

John

[Ken Pachkowsky]

Yes, that is correct.

Ken

[Steve Schell]

Hi guys,

Mr. Widget, I sure wish I could bring you to my place for a day to show me the fine points of DEQX setup. My partner and I have been able to grunt through the filter correction process, but it has been like pulling teeth for us.

Our general approach has been to make and apply the correction filters for mono sub and left and right bass and midrange/high frequency horns and also add delay to the mains to let the long path sub catch up. Then we feed pink noise to the system and juggle individual driver levels and perhaps a bit of correction with the parametric to get the most even and pleasing response in the room, monitoring with the Phonic RTA.

I think we are getting close with it. I have recently made some recordings with a coincident pair of ribbon mics and an Alesis Masterlink running 24 bit/96kHz. A recording made at the Fullerton train station worked out well, and the freight trains are now rolling through the living room in a believable fashion.

[boputnam]

...and also add delay to the mains to let the long path sub catch up.

It'd be interesting to see the impulse measurement of each the mains and the subs, if you get them...

[boputnam]

For those who don't understand what this graph shows... it is not the response of the speakers, it is the electrical boost and cut created by the DEQX unit to balance the response of the speakers.

How is this measured, Widget? How do DEQX know?

Is there a mic used to "feedback" into the DEQX the speaker output, so that "it" knows what to do?

[Steve Schell]

Hi Bo,

I suppose that the impulse plots are in there, but I am not sure how to access them. I am still quite a novice with this thing, even after devoting considerable time to learning it. It's a beast, but then I have never been very good with computers.

I can take a stab at your question about the DEQX measurement process. When you make a measurement, the program feeds several bursts of signal, pink noise I think, to the particular driver being measured. The microphone supplied with the unit is placed in front and its signal is fed back into the DEQX. Once all the measurement data is recorded the attached computer running the program does some serious number crunching and after a couple of minutes produces the impulse response plots for each channel, with the individual drivers overlayed. It asks you to window the graph, then after more crunching produces the "correction filters" or inverse of the speakers' amplitude and phase errors, plus correction for their errors in arrival time. Once these correction filters are applied the speaker produces a virtually perfect amplitude and phase response, at least at the mic position. I have heard the unit work miracles, even with my rudimentary use of it.

[Mr. Widget]

How is this measured, Widget? How do DEQX know?

Is there a mic used to "feedback" into the DEQX the speaker output, so that "it" knows what to do?

Yes, when you buy the DEQX with the calibration software (an option) you get a calibrated mic. The curve is essentially the inverse of what was measured from the speaker. In this case it was two measurements, one for the left and one for the right speaker. In both cases I had the mic on a stand at mid horn height on axis and 1m away from the speaker.


Widget