For personal use reaper is also well priced. I have only tried using DRC with DRC Designer, Use a flat target, using the normal correction but with a bit more correction in the bass and I up the maximum correction to the full 12db as I have no headroom issues.
For the room curve I use one of the EQs built into reaper, reafir.
Listening position response before DRC:
After DRC:
For the room curve I use one of the EQs built into reaper, reafir.
Listening position response before DRC:
After DRC:
I've always used REW measurements within DRC Designer. That way you get to look at the basics results you are using your correction on.
Don't expect FIR to solve everything. Solve what you can with other means like damping panels etc. Get the best result you can and then take it a step further with FIR. At least, that's what I would recommend.
Look at the thread:
http://www.diyaudio.com/forums/full-range/275730-convolution-based-alternative-electrical-loudspeaker-correction-networks.html for more info on how to use REW with DRC-FIR or DRCDesigner.
Don't expect FIR to solve everything. Solve what you can with other means like damping panels etc. Get the best result you can and then take it a step further with FIR. At least, that's what I would recommend.
Look at the thread:
http://www.diyaudio.com/forums/full-range/275730-convolution-based-alternative-electrical-loudspeaker-correction-networks.html for more info on how to use REW with DRC-FIR or DRCDesigner.
Hi Wesayso, inspired partly by your Two Towers thread I built a pair of 24 driver line arrays using Faital Pro 3FE22 (crossed to 15" subs)
I have iir pre EQ (measured outdoors) which also helped phase response but am now trying to use fir for room correction. I have damping panels at first reflection points and behind the listening position (only 3 feet from back wall)
Audio is my thing but computers and software not so much! I tried installing Sox to convert REW WAV files to PCM for DRC Designer but for some reason it won't install on my machine..
I also have Audacity but can't see a way of converting to .PCM?
When you say you get to look at the results, does DRC designer show an imported curve then?
I have iir pre EQ (measured outdoors) which also helped phase response but am now trying to use fir for room correction. I have damping panels at first reflection points and behind the listening position (only 3 feet from back wall)
Audio is my thing but computers and software not so much! I tried installing Sox to convert REW WAV files to PCM for DRC Designer but for some reason it won't install on my machine..
I also have Audacity but can't see a way of converting to .PCM?
When you say you get to look at the results, does DRC designer show an imported curve then?
Glad to read my thread was inspiring, part of my goal was to give back for what I've learned here.
No, that's the point. DRCDesigner does not show anything. But the REW measurement does. You can look at the RAW sweep in REW and determine if it's doing well enough for further treatment with FIR filters. DRCDesigner only runs the sweep and will base the correction on that sweep. But you won't know if the starting base point is any good. Lots of things can go wrong. I like to see every step along the way.
As far as I know Sox is included in DRCDesiger and simply is a command line tool. Personally I use an old Cool edit version for that step (again to have a visual confirmation) but I can imagine this stuff getting a bit confusing without the proper guidance. I believe either Greg (gmad) or Perceval has written a step by step guide which includes converting with Audacity. I believe it's as simple as importing RAW data (the REW exported wav file) and exporting it to the proper PCM file type.
No, that's the point. DRCDesigner does not show anything. But the REW measurement does. You can look at the RAW sweep in REW and determine if it's doing well enough for further treatment with FIR filters. DRCDesigner only runs the sweep and will base the correction on that sweep. But you won't know if the starting base point is any good. Lots of things can go wrong. I like to see every step along the way.
As far as I know Sox is included in DRCDesiger and simply is a command line tool. Personally I use an old Cool edit version for that step (again to have a visual confirmation) but I can imagine this stuff getting a bit confusing without the proper guidance. I believe either Greg (gmad) or Perceval has written a step by step guide which includes converting with Audacity. I believe it's as simple as importing RAW data (the REW exported wav file) and exporting it to the proper PCM file type.
Yep, it is on the other convolution thread:
http://www.diyaudio.com/forums/full...dspeaker-correction-networks.html#post4358541
http://www.diyaudio.com/forums/full...dspeaker-correction-networks.html#post4358541
Just read all of the Two Towers, DRC-FIR and Rephase threads and since I've made about 300 sweeps in REW this month, I thought I would share and join the fun!
BTW, great job Wesayso on the Towers and the journey afterwards...
My system consists of an HTPC (i7700K) Running into an older AMD (ATI) video card and then HDMI out to a Transformative Engineering HDS-42AVR 4K HDMI audio/video splitter.
HDMI video goes to the TV and HDMI audio goes to a MiniDSP nanoAVR HD for crossover and IIR PEQ duties.
Then onto an Outlaw Audio Model 975 Pre/pro for preamp and DAC. Only one D/A stage in the whole chain...
Low end consists of a Crown XLS-1502 driving a pair of dual NHT 1259 woofers in cabinets I built (Q of around .707)
Currently crossed over ( LR48 150Hz.) to 60" Carver planar magnetic ribbon dipoles in slabs that I also built about 20 years ago. I like these so much that I have felt no need for anything else...
Surrounds are Optimus LX-5's modded with Seas mid-woofers that are surprisingly flat in response powered by a homemade SS amp.
Here is my system as of today:
Managed to get EQ/APO running in the Windows chain and went to YouTube for some quick music to check it, found some David Gilmour Live at Pompeii 2016 that I found so enthralling that I just listened for the rest of that night...
On to some REW caps...
MiniDSP only. SPL/Phase
DRC-FIR ERB
Manually done in RePhase
And since I accidentally found out that you can listen to both convos in series ( or is it parallel?) in EQ-APO here is a shot of that...
Now the same set with a 4 cycle FDW
Mini DSP alone
DRC-FIR-ERB
Rephase
Both at once....
And now for something completely different, some measured impulse responses...
Just MiniDSP. I get way better looking ones if I vector average a bunch of different mic positions
DC-FIR-ERB ( FWIW, I actually like the soft filter better now, but a month ago, on first impression, I liked this better and put it on APO..
I'm on windows beta preview on this machine that proceeded to break APO and i just got it working again with the latest update..
Living dangerously...
And the interesting one... Rephase, with a big pre- echo, but a nice central spike... I think i over corrected here...
Anyway I thought that I'd throw that out there and put a dipole into the mix on here. I do get some inherent side cancellations due to the figure 8 response of a dipole and it really does help.
Wesayso doesn't get that, but he also deals with the backwave nicely in his towers...
Reply's are appreciated!
BTW, great job Wesayso on the Towers and the journey afterwards...
My system consists of an HTPC (i7700K) Running into an older AMD (ATI) video card and then HDMI out to a Transformative Engineering HDS-42AVR 4K HDMI audio/video splitter.
HDMI video goes to the TV and HDMI audio goes to a MiniDSP nanoAVR HD for crossover and IIR PEQ duties.
Then onto an Outlaw Audio Model 975 Pre/pro for preamp and DAC. Only one D/A stage in the whole chain...
Low end consists of a Crown XLS-1502 driving a pair of dual NHT 1259 woofers in cabinets I built (Q of around .707)
Currently crossed over ( LR48 150Hz.) to 60" Carver planar magnetic ribbon dipoles in slabs that I also built about 20 years ago. I like these so much that I have felt no need for anything else...
Surrounds are Optimus LX-5's modded with Seas mid-woofers that are surprisingly flat in response powered by a homemade SS amp.
Here is my system as of today:
Managed to get EQ/APO running in the Windows chain and went to YouTube for some quick music to check it, found some David Gilmour Live at Pompeii 2016 that I found so enthralling that I just listened for the rest of that night...
On to some REW caps...
MiniDSP only. SPL/Phase
DRC-FIR ERB
Manually done in RePhase
And since I accidentally found out that you can listen to both convos in series ( or is it parallel?) in EQ-APO here is a shot of that...
Now the same set with a 4 cycle FDW
Mini DSP alone
DRC-FIR-ERB
Rephase
Both at once....
And now for something completely different, some measured impulse responses...
Just MiniDSP. I get way better looking ones if I vector average a bunch of different mic positions
DC-FIR-ERB ( FWIW, I actually like the soft filter better now, but a month ago, on first impression, I liked this better and put it on APO..
I'm on windows beta preview on this machine that proceeded to break APO and i just got it working again with the latest update..
Living dangerously...
And the interesting one... Rephase, with a big pre- echo, but a nice central spike... I think i over corrected here...
Anyway I thought that I'd throw that out there and put a dipole into the mix on here. I do get some inherent side cancellations due to the figure 8 response of a dipole and it really does help.
Wesayso doesn't get that, but he also deals with the backwave nicely in his towers...
Reply's are appreciated!
Hi Gary,
Nice speakers you have there! Must have been quite the task to wade trough the Two Towers thread, sorry for that.
If measuring at different spots and averaging that gives you a much cleaner impulse, it means you have some reflections/diffraction when measuring at a single spot.
Best way to go forward from that is to use the average of multiple measurements, close to the listening sweet spot and correct that with DRC.
DRC has a lot of settings, looks like your ERB template isn't correcting much of the phase turn.
You might want to try Greg's recipe for correction, I'd still use an averaged set of measurements for that to clean up placement dependant anomalities.
Greg's thread can be found here: A convolution based alternative to electrical loudspeaker correction networks
ERB uses quite long windows in the midrange, which can sound funny. For my arrays I use my own template with a longer correction window up top, but about 4 cycles trough the midrange. My phase correction window is custom too. As short as I could get away with. Seeing your couch in the picture, I can't help but think, that too will probably show up in your measurements as reflections.
I spend quite some time firing IR sweeps to track down my reflections and soften them up with damping panels. That's the reason my IR's look pretty clean wherever I measure. I still get benefits from using averaging of multiple measurements. I used 3 spots for each channel, about 8.5 cm apart. More is even better when there are lots of tiny differences. That way you average out the placement specific reflections and stand less of a chance to overcorrect anything.
Be aware not to use long windows in RePhase to correct the phase. You could use a 4 or 6 cycle window to correct the phase part.
To get a 6 cycle window in REW for output to RePhase you can use a FDW of 6 cycles on your measurement, next go to the All SPL tab and under Trace Arithmetic choose that filtered measurement in the "A" and "B" windows and use A+B to sum the (A and B, both pointing to the same trace with that 6 cycle FDW applied) measurement, next you offset that newly created A + B trace by -6 dB (on the left side of the controls window) and if you export that trace it will represent the 6 cycle version.
I use that trick because applying a frequency dependent window and exporting that, still gives you an unfiltered export.
Nice speakers you have there! Must have been quite the task to wade trough the Two Towers thread, sorry for that
.If measuring at different spots and averaging that gives you a much cleaner impulse, it means you have some reflections/diffraction when measuring at a single spot.
Best way to go forward from that is to use the average of multiple measurements, close to the listening sweet spot and correct that with DRC.
DRC has a lot of settings, looks like your ERB template isn't correcting much of the phase turn.
You might want to try Greg's recipe for correction, I'd still use an averaged set of measurements for that to clean up placement dependant anomalities.
Greg's thread can be found here: A convolution based alternative to electrical loudspeaker correction networks
ERB uses quite long windows in the midrange, which can sound funny. For my arrays I use my own template with a longer correction window up top, but about 4 cycles trough the midrange. My phase correction window is custom too. As short as I could get away with. Seeing your couch in the picture, I can't help but think, that too will probably show up in your measurements as reflections.
I spend quite some time firing IR sweeps to track down my reflections and soften them up with damping panels. That's the reason my IR's look pretty clean wherever I measure. I still get benefits from using averaging of multiple measurements. I used 3 spots for each channel, about 8.5 cm apart. More is even better when there are lots of tiny differences. That way you average out the placement specific reflections and stand less of a chance to overcorrect anything.
Be aware not to use long windows in RePhase to correct the phase. You could use a 4 or 6 cycle window to correct the phase part.
To get a 6 cycle window in REW for output to RePhase you can use a FDW of 6 cycles on your measurement, next go to the All SPL tab and under Trace Arithmetic choose that filtered measurement in the "A" and "B" windows and use A+B to sum the (A and B, both pointing to the same trace with that 6 cycle FDW applied) measurement, next you offset that newly created A + B trace by -6 dB (on the left side of the controls window) and if you export that trace it will represent the 6 cycle version.
I use that trick because applying a frequency dependent window and exporting that, still gives you an unfiltered export.
Last edited:
Gary, great job following the footsteps of giants like Wesayso and gmad.
So, you have ribbons that can reach 150Hz, eh? Not many of those these days.
I'll follow you on your journey, as I also have an OB setup, and of course, a lot more reflections showing up than the Two Towers setup.
Keep it up!
So, you have ribbons that can reach 150Hz, eh? Not many of those these days.
I'll follow you on your journey, as I also have an OB setup, and of course, a lot more reflections showing up than the Two Towers setup.
Keep it up!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.