mbon said:
Dennis S Bragion's DRC documentation details the same spectral time frequence plots.
Other programs that I've used to sucessfully do this are ETF and SIA Smaart Acoustic Analysis.
Besides crossover tuning, its especially useful for fine tuning the DRC parameters while measuring in-room.
mbon said:
IMHO spectral view is not enough precise to align the HP, that can be used to foresee the large defects but it is ineffective since the shift is about some samples.
An example, here I have create a misalignment of 20 samples (is about 16 cm) at 700Hz (and it is worse when one goes down in frequency): it's completely invisible in spectral view, even by triturating the adjustments, and quite visible if the high and bass speaker are are measured separately (preferably with the convolution over the dirac method), or with a High and low pass filter in the global impulse in sample view mode.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Hello, I use high resolution 192Khz samples, it is easy to align two drivers by tuning the display parameters, for fine tuning it is strongly recommanded to apply FFT filters as you mention but the spectral view should always give the same information with a good measurement and good display tuning IMHO.heresie said:
mbon said:ShinOBIWA ???? and the DEQX review ??? what are you doing ? we are all waiting for
Hi but...
I haven't even touched it since I posted that it solved the distortion problem I was having - sorry.
The speakers are in bits (again!) and aren't going back together till I've sprayed them up and can finally consider them finished. I have the odd day of decent weather and the patio heater helps too but spray on a bad day however and its all a waste of time.
I'll do a summary of what I think when I've got some opinions to post. My guess is a couple more months at this rate.
yup. looks good. Only stereo from what I can tell, though. Can anyone comment on what type cpu resources it needs? I'm running on a 1GHz PIII, which is obviously underpowered by todays standards.
Link:
http://www.knufinke.de/sir/index_en.html
Link:
http://www.knufinke.de/sir/index_en.html
Within Console I dragged out three SIR boxes.
One for the front L/R channels
Another for rear L/R
And a third for Center and Sub
I did this just to see if it was feasible. I have only actually produced an Impulse response file for the front two channels.
The CPU was struggling with this set up and it's a P4 at 3ghz.
You've raised a basic question for me that I realized I hadn't understood.
In a multi-channel set up should you use separate Impulse response files to correct each channel individually or by stereo pairs as above, or should you produce one file to correct all 6 channels. If so which software makes this possible?
Thanks
Geoff
One for the front L/R channels
Another for rear L/R
And a third for Center and Sub
I did this just to see if it was feasible. I have only actually produced an Impulse response file for the front two channels.
The CPU was struggling with this set up and it's a P4 at 3ghz.
You've raised a basic question for me that I realized I hadn't understood.
In a multi-channel set up should you use separate Impulse response files to correct each channel individually or by stereo pairs as above, or should you produce one file to correct all 6 channels. If so which software makes this possible?
Thanks
Geoff
geofstro said:
Hi again Geoff
Best results are gained from individual impulse responses for each channel. Do not use the main's or any other measurement that wasn't taken for that particular speaker - it will sound phasey and unnatural. If processing power is limited them you can just do your fronts and this will provide the biggest gains but I noticed an increase in the spaciousness and size of the rear soundfield when I applied DRC to them, its not as dramatic as doing the front with 2 channel music but it just moves the surround experience on another notch.
You can also shave off a few cycles from the audio processing by using simple parametric EQ on the sub to flat the response in conjunction with an RTA like TrueRTA.
You do need a fair amount of CPU power to run a 5.1 DRC setup with XO's. I use an Athlon X2 dual core CPU which has been overclocked to 2.8Ghz. This would make for a P4 rating of over 5Ghz if AMD speed rating is to be trusted (which it isn't). So I can run all the options and the system doesn't glitch. You can never have too much CPU power.
Hi ShinOBIWAN,
Thanks for the clear explanation. When you use Denis's DRC program you measure each channel separately. I saved the separate IR file for each of the L and R front channels. I also combined these into a stereo IR file, which you have to do if you intend to use a convolver which only supports stereo IR files, as is the case with SIR.
Providing each channel was measured separately, I believe the phase related problems you were referring to would not occur just because the IR for each channel is combined into a stereo IR for playback.
Am I right?
If so a stereo convolver such as SIR could be used in a 5.1 set-up when using console to route the channels via separate SIR boxes for each pair of channels as I described.
I can see how this might work well for both the front L + R channels and the rear L + R channels. What happens though when the Centre and Sub channels are treated as a stereo pair?
I'm starting to think there may be a problem there since the volume for the center and sub are adjusted separately. This is unlike the other channels where the volume is adjusted for both channels of the stereo pair together.
How does volume relate to DRC in such a situation?
Thanks again
Geoff
Thanks for the clear explanation. When you use Denis's DRC program you measure each channel separately. I saved the separate IR file for each of the L and R front channels. I also combined these into a stereo IR file, which you have to do if you intend to use a convolver which only supports stereo IR files, as is the case with SIR.
Providing each channel was measured separately, I believe the phase related problems you were referring to would not occur just because the IR for each channel is combined into a stereo IR for playback.
Am I right?
If so a stereo convolver such as SIR could be used in a 5.1 set-up when using console to route the channels via separate SIR boxes for each pair of channels as I described.
I can see how this might work well for both the front L + R channels and the rear L + R channels. What happens though when the Centre and Sub channels are treated as a stereo pair?
I'm starting to think there may be a problem there since the volume for the center and sub are adjusted separately. This is unlike the other channels where the volume is adjusted for both channels of the stereo pair together.
How does volume relate to DRC in such a situation?
Thanks again
Geoff
I thought I'd chime in since this thread was so popular. I have a rather simple, IIR-based 4-way crossover software that doesn't strain a 1GHz processor AND produces linear phase, transient-perfect crossovers when dialed in correctly. You can have pretty steep slopes and no ringing associated with FIR filters. The program is a VST plugin that can be used inside the Console, or with its own VST host - called VSTShell.
You can read about it at www.thuneau.com. I'm still in Beta stages, but the current version works pretty good.
You can sign up to be a Beta tester at my forum. All I ask, is that you have an ASIO sound card and take some time to use the software and dilligently report any bugs you find.
You can read about it at www.thuneau.com. I'm still in Beta stages, but the current version works pretty good.
You can sign up to be a Beta tester at my forum. All I ask, is that you have an ASIO sound card and take some time to use the software and dilligently report any bugs you find.
I am unsure whether my question is on topic, however, here it is anyway.
It appears that to output the filtered signals for, e.g., three separate frequency ranges, each for a left and right channels, out of the PC, a rather expensive cards are required. Would it be possible to run three separate cheaper cards, e.g., Chaintech 710 and send each sfiltered signal to one of them?
Thank you,
M
It appears that to output the filtered signals for, e.g., three separate frequency ranges, each for a left and right channels, out of the PC, a rather expensive cards are required. Would it be possible to run three separate cheaper cards, e.g., Chaintech 710 and send each sfiltered signal to one of them?
Thank you,
M
>>>The program is a VST plugin that can be used inside the Console, or with its own VST host - called VSTShell.
You can read about it at www.thuneau.com.
Thats interesting ! I tried almost every Linear phase EQ on the market , but none of those were optimized for xover duties .
You can read about it at www.thuneau.com.
Thats interesting ! I tried almost every Linear phase EQ on the market , but none of those were optimized for xover duties .
mefistofelez said:
No, one card per instance. It's due to sync issues. The crossover legs have to mach to the same clock, and that is not easilly possible with three cards. You can get get a very nice multichannel card like the Audiotrak ProDigy 7.1 for under $100.00. Not much more than three $20.00 cards.
Vil said:
It's not really a linear phase EQ. It's reverse phase filtering. The signal leaves the processor with its phase mangled, but in the opposite direction than the phase roll of the loudspeaker crossover filters. They both offset eachother.
Thunau said:
Hi,
But in many case 8 channels are not enough
With stream boy and at lest one card with directwire/patchmix you can use 1, 2, 3...different card.. but indeed you need to correct the delay precisely (I suggest to use one card for a couple of speaker if active filter)
An externally hosted image should be here but it was not working when we last tested it.
Regards,
Avi
Multiple cards just don't work unless you can sync all their clocks. Consider that if the clock is off even a fraction of a tick longer on one card (eg one sample per second), then after only 44000 secs (ie 12 hours) the sync is 1 *second* out. Even less than that is a going to be a gross sync error of course!
So pro cards can take an spdif (or similar) between two cards and use that to keep the clocks roughly the same.
I am using a slight trick where I have extra digital channels out on one of my cards and feed those into the spdif in on another audio card and cross that back to the analogue out on that card. ie basically use the spare soundcard as an spdif fed DAC.
There are a tiny number of software programs which try and resample the audio to keep a couple of soundcards in sync, but I honestly don't think you will get good results in general using this method (perhaps useful for front and surround speakers...)
Good luck
So pro cards can take an spdif (or similar) between two cards and use that to keep the clocks roughly the same.
I am using a slight trick where I have extra digital channels out on one of my cards and feed those into the spdif in on another audio card and cross that back to the analogue out on that card. ie basically use the spare soundcard as an spdif fed DAC.
There are a tiny number of software programs which try and resample the audio to keep a couple of soundcards in sync, but I honestly don't think you will get good results in general using this method (perhaps useful for front and surround speakers...)
Good luck
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.