A how to for a PC XO.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
diyAudio Member
Joined 2004
Audiophilenoob said:
you need to take measurements with a RTA if I'm not mistaken

Its entirely possible to do that and enter the values by hand. But its tedious and inaccurate.

m0tion said:
Ok, i've read through the CurveEQ docs. Are you capturing your room response in CurveEQ or importing it from another program? I'm ready to give this aspect a shot, but I'm pretty confused as to how to get my room's response into CurveEQ. I know once I get it in there I can just hit the 'I' to invert it and then it should give me flat response.

Set CurveEQ to capture and route the input the mic is plugged into to CurveEQ. Play a sine sweep from 20hz-20khz for a 1 minute duration. At this point you'll see Curve EQ draw your room response, then hit invert, save and do the same for the other speakers.

Reconnect CurveEQ back to the correct inputs and then play the sine sweep again but in TrueRTA(1/24th Octave). You'll now have a flat response from that speaker.

Easy :)
 
I went ahead and bit the bullet and shelled out the cashola for the software...

got Waves Platinum
Console 1.4
and curveEQ from voxengo

haven't messed with it yet... this is awaiting some reading and a new sound card

something I'm considering doing here

I may just keep my DBX for xover duty... as likely this sound between this xover and the comp one will not be audibly superior either way

I can then go... from the 1st output from the sound card to the xover... then back in 6channels of input to the card then out again after DRC and phase adjustment???

anything wrong with that???

or will the sound card not be able to have one output seperated from the others?

this will keep my CPU usage low... since I plan to use this for comp duty during listening and gaming... of course 60% CPU usage is unacceptable here

even though my comp is a beast

(2.5ghz Oc'd A64 3500+, 1gb Oc'd ram... 250gb HD, etc etc)

I don't think it'll perform where it needs to be to have a full scale FIR filtering on the comp
 
diyAudio Member
Joined 2004
kan3 said:
Probably not a concern since the topic is audio and not video, but a 6200 will not do HiDef video. It'll work just fine on standard def but once you switch out of overlay into vmr9 and do ffdshow processing it's toast.

I think your referring to the 6200 in the PC XO?

Its not a problem at all, since the only thing that computer does is filter audio. I could have used a 13 year old Trident PCI graphics card and it would have still been fine :)

The main PC does the video and thats a 7800 GTX so Hidef isn't a problem.
 
Ok, I feel pretty retarded, but I'm going to need some more help before I can correctly use CurveEQ (yes, I read the docs, I think they suck). Attached is a picture of the setup I'm currectly using, just a simple test setup with one instance of CurveEQ per channel. I have my line-input routed to input 7 and 8 and my microphone is connected to the left side of the line-input (coresponds to input 7 in console). I made a sweep wave file and I'm playing it through Winamp.

CurveEQ wouldn't register any input unless it was connected to an output (is this right?) so I connected it to an output on the card I don't use. My plan was to capture the input from the line-in and then once the spectrum was captured, save it, and use it in an instance of CurveEQ that was then actually connected to my real input. I used the little 'C' button and then played the WAV file in an instance of CurveEQ and it seems to work fine, from the monitoring lines I could see it slowly work it's way through the entire spectrum. One thing I found strange is that the spectrum was sloping to the upper-right as it went along and I know that my system can't have a constantly rising response, can it?

Anyway, once I captured the spectrum it told me that I had a spectrum available, but nothing appeared in the CurveEQ graph window, so I assume I use the 'M' button to match the specturm (how many points should I use? I figured about 30). Once I told it to match the spectrum it seemed to start another capture process and it said "matching spectrum". At this point I could again seem the input from my microphone being captured. I wasn't sure what to do so I figured I'd play the sweep again. Once it was through I hit the 'S' to stop the matching. Then what appeared was 30 points outlining what I can only assume is the difference between what was captured the first time and what was captured the second time. This really doesn't help me =).

I swear I've read the docs, but they seem really really brief on the SpectraMatch section. Shin, HELP! =)
 

Attachments

  • console2.jpg
    console2.jpg
    35 KB · Views: 1,443
Compatible soundcards

Hello.

I was looking for alternatives to the RME solution, and I came across this convenient list of 2004's cards: http://www.digit-life.com/articles2/digests/sound2k4.html

Of note:

E-MU 1820m --- Looks like a good alternative if it'll work
E-MU 1212m

ESI juli@ ( http://www.digit-life.com/articles2/esi-julia/index.html )

- The E-MU boards support ASIO2, but I'm not sure on the other requirements.
- The E-MU 1212m doesn't seem to have enough analog outputs, but the 1820m is looking good
- ESI juli@ seems to be from the same people as the audiotrak prodigy 7.1; it looks like it could be a higher quality solution though. I'm not sure if it'll handle a 3-way system though. It does handle the ASIO and MME, not sure about "DX."

Does anyone feel like looking into these options? The E-MU boards are supposed to test a little better than the Lynx2 for some odd reason. The ESI Juli@ seems to test fairly well too.

Anyone have any information on this?

Also, the Prodigy 7.1LT vs. the 7.1; I'm trying to figure out the differences here. It seems the LT version can still handle 8 outputs via a G9 connector ?

All this is making me :bawling:

Thanks for any help. :)
 
Sure, I've looked at all of these cards actually:

ESI Juli@: Good card, uses Directwire and has ASIO2 support. Essentially the same from a software prespective as the Audiotrak Prodigy 7.1LT which works great. Only problem is that this card is 2-analog-in, 2-analog-out. So all you could really do is DRC a stereo setup, no crossovers.


E-MU Series: Looks like great hardware, it supports ASIO which is the important part. It is not known whether or not you can route inputs through software ("internal routing") so you'd have to buy it and see or call Creative and ask.


Audiotrak Prodigy: Audiotrak Prodigy 7.1 and 7.1LT are essentially the same card that has been laid out differently. The 7.1LT is the successor to the 7.1 and has a smaller form-factor (and is cheaper). It is also considered to have superior sound quality because of the new layout. The 7.1 is no longer available. The 7.1LT is available at newegg and other places. If you see references to the 7.1 at online stores, check their inventory, they're out of stock and haven't update their websites yet.
 
OK, more to post:

Shin: Have you been adjusting the response of the Microphone?

- Measurement Microphones should be re-zero'ed (similar to curveEQ inversion, but for a microphone) by way of a stored file of the mic's frequency response.
- I'm sure you knew about that, but has that been worked into the solution?


- I've partially verified that the E-MU 1212m/1820m cards support all the plugin types we need: Dx, MME, ASIO2, VST...
- There's a good chance it'll do the internal digital routing just because it has these plugin features and is marketed toward the same end
- It has 8 balanced analog outputs, a CLEAN design, and an attractive pricepoint
- Not sure if you can increase the number of outputs (2-cards?)

Any input guys?
 
Thanks for the info m0tion

- I've e-mailed E-MU about this, hopefully I'll get a response
- I'm not willing to purchase it blindly...heh, I'd rather just get a prodigy in that case

I've read several articles about the E-DSP chip and the 1820m in general that seem to indicate that it'd work...but I can't be certain.
 
diyAudio Member
Joined 2004
Archmage said:
OK, more to post:

Shin: Have you been adjusting the response of the Microphone?

- Measurement Microphones should be re-zero'ed (similar to curveEQ inversion, but for a microphone) by way of a stored file of the mic's frequency response.
- I'm sure you knew about that, but has that been worked into the solution?


Yes I'm aware of this, many RTA's use calibration files for mics. However the mic I used really didn't need one.

I used a an Earthworks M50 Mic to do my original measurements - these are good mics BTW. Unfortunately it was hired as the cost was way out of my 'acceptable money for a mic' range.

I've since started rebuilding the speakers so I'll need to hire out one again for the initial setup.

I also have a Behringer ECM8000 mic and that does need some equalisation on the top end of its measurement range. I'm quite sure the differences with and without the calibration file are very minimal though.

Sorry I've not updated anything on this thread but I've been rather ill the last week and its only the past couple of days I've felt like doing anything. I've also got speakers that I'd like to finish as well as other daft things like making up for lost time at work.
 
Hmm

So would it be easy to use a mic-calibration file with the proposed software in this thread (namely CurveEQ); I'm assuming it's something so basic that it would have been an easy feature to implement. :confused:

Also:

- E-MU confirmed that the 1820m would work for this.
- The card is limited to 8 analog outputs, but perhaps there's a way to get more (2 cards, would that work?)


It looks like I may be trying this whole thing out as it's quite affordable considering that I have a spare PC just for this...the Sound card replaces the cost of the planned DCX 2496's, and well I guess there are other expenses, but I would have spent that money anyway :)

- Anyone think an Athlon XP Thoroughbred at about 2.15 Ghz and 512 or 768mb of RAM might handle it? Dedicated Audio processing; I think it could do it...
 
Waves EQ alternatives & DSP/FPGA boards

Hi,

A couple of thoughts :

- Is there a way of utilising the relatively cheap FPGA or DSP PCI boards to do the FIR filter processing for the XO and greatly reduce the CPU requirements as well ?
I'm no expert in this area but I got a lot of hits when I Google'd "FIR and FPGA/DSP"

- If not, is there a cheaper alternative to Waves EQ ?
 
diyAudio Member
Joined 2004
Re: Waves EQ alternatives & DSP/FPGA boards

jacklelad said:
Hi,

A couple of thoughts :

- Is there a way of utilising the relatively cheap FPGA or DSP PCI boards to do the FIR filter processing for the XO and greatly reduce the CPU requirements as well ?
I'm no expert in this area but I got a lot of hits when I Google'd "FIR and FPGA/DSP"

- If not, is there a cheaper alternative to Waves EQ ?

The RME and most most other cards nowadays use DSP's. The problem is that whilst they will provide virtually zero CPU load on supported instructions and when it comes to FIR very few(I don't know of any) support this in some form of native instruction set. Aside from this you'd need software specifically written to take advantage of such a set of instructions.
Its not all bad because the Waves and Voxengo stuff is SSE optimised.

What happens here is the PC CPU does the bulk of the processing which takes time because its non-specialised for the task. The stuff like conversion, sampling and routing is done by the card and doesn't take up any CPU overhead.
 
jacklelad said:
Waves is very expensive !

Would this :-

http://www.elementalaudio.com/products/eqium/index.html

at 129 dollars do the same thing ? Has anyone taken a look ?
They also have a linear phase EQ at the same price.
They do not mention FIR explicitly, so I can email them to ask if it looks promising.

Jack


shoot me an email about waves
 
I think I'm going to settle for an E-MU 1820m

Anyone know a good source for them (price)? I'm looking at YOU AudiophileNoob! ...

I need to order by the end of the week, so I better find a good place to order from pronto :smash:

- Heh, I found a source for just ONE of those PHL 1120's, so I'm sticking with my Excels for Revision1 - maybe I'll like them :cannotbe:

Edit: Actually the guy with (1) 1120 driver is going to hold onto it, but at the same time I found another source with possibly many. I'll know soon enough...
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.