I looked into DRC a few years ago but never quite got it to work the way I wanted. Is anyone here using DRC and has there been any advancement in DRC in the past few years?
For those of you who have not heard of DRC (Digital Room Correction) here are some links.
DRC: Digital Room Correction
Digital Room Correction | Download Digital Room Correction software for free at SourceForge.net
Digital room correction - Wikipedia, the free encyclopedia
DRC
For those of you who have not heard of DRC (Digital Room Correction) here are some links.
DRC: Digital Room Correction
Digital Room Correction | Download Digital Room Correction software for free at SourceForge.net
Digital room correction - Wikipedia, the free encyclopedia
DRC
I believe many users of brutefir / brutefirdrc use Denis Sbragion's DRC to generate the room correction filters required. There are alternatives to DRC (Inguz, Room EQ Wizard, Acourate, etc, some of which are non-free).
The first link you provided leads you to documentation, which has an extensive changelog. As it is, there seem to be so many options in setting up filters that non DSP experts like me find it challenging to comprehend and optimize. I'm not aware of any major advance in the science of DSP / room correction that calls out for enhancements in DRC (the software by Denis).
For me (brutefirdrc + squeezebox Touch), drc works very well.
The first link you provided leads you to documentation, which has an extensive changelog. As it is, there seem to be so many options in setting up filters that non DSP experts like me find it challenging to comprehend and optimize. I'm not aware of any major advance in the science of DSP / room correction that calls out for enhancements in DRC (the software by Denis).
For me (brutefirdrc + squeezebox Touch), drc works very well.
I did tried and used DRC based convolution for a while last fall. I took me a while to figure things out.
I've been running brutefir on a Squeezboxserver.
I used a calibrated Behringer ECM8000 and an EMU0404 for the recording.
aljordan from audio asylum has prepared a nice toolbox to prepare the pulse responses for your fir-filters based on DRC on a Windows system.
The result was quite "impressive" - at the first glance. Pretty clean, sterile and impulsive presentation.
However:
1.
I put quite some effort into it. Though I skipped buying a better mic for 500$+ and even more expensive AD converter to get a rather flawless room response recorded.
Join the Acourate usr group at Yahoo. There you'll find quite some expert talk about all that.
I consider a real good room-response recording the first big hurdle. And the recording is the base for everything which
comes later on.
Not to forget: You'd theoretically have to redo it everytime you change anything in your room.
2.
Consider that you need different filters for different sample rates.
Without resampling you'll have a hard time to get multiple samplerates supported. Your convolution application/plugin has to be able to switch automatically samplerates
or you need to resample everything which will cause losses on the resampling side. Manual switching I consider not acceptable.
3.
The room response is recorded at a single point in the room. This single-point-shot will impact the result if you don't sit at that exact recording position later on.
However if your base system or room is that bad sounding, those in comparison rather little flaws introduced by that measurement position might be a better compromise.
4.
Not one filter is lossless. You can't avoid introducing losses to the signal by filtering it.
5.
You have 2 milllion options to tweak another million parameters. Or you just take rather standard configs.
I for myself skipped all that equalization after testing it. Perhaps I gave up to early on it.
I think it is much better strategy to get the flaws out of your system and room first and leave the base data untouched.
I actually found the better your system the less room interaction you'll see. I always felt that micro-details, timbre of instruments and reverberation went
down the drain when applying this equalization. I never had the feeling being able to get all this under control by putting even more effort (time&money)
into it.
Just my 2 cents.
I've been running brutefir on a Squeezboxserver.
I used a calibrated Behringer ECM8000 and an EMU0404 for the recording.
aljordan from audio asylum has prepared a nice toolbox to prepare the pulse responses for your fir-filters based on DRC on a Windows system.
The result was quite "impressive" - at the first glance. Pretty clean, sterile and impulsive presentation.
However:
1.
I put quite some effort into it. Though I skipped buying a better mic for 500$+ and even more expensive AD converter to get a rather flawless room response recorded.
Join the Acourate usr group at Yahoo. There you'll find quite some expert talk about all that.
I consider a real good room-response recording the first big hurdle. And the recording is the base for everything which
comes later on.
Not to forget: You'd theoretically have to redo it everytime you change anything in your room.
2.
Consider that you need different filters for different sample rates.
Without resampling you'll have a hard time to get multiple samplerates supported. Your convolution application/plugin has to be able to switch automatically samplerates
or you need to resample everything which will cause losses on the resampling side. Manual switching I consider not acceptable.
3.
The room response is recorded at a single point in the room. This single-point-shot will impact the result if you don't sit at that exact recording position later on.
However if your base system or room is that bad sounding, those in comparison rather little flaws introduced by that measurement position might be a better compromise.
4.
Not one filter is lossless. You can't avoid introducing losses to the signal by filtering it.
5.
You have 2 milllion options to tweak another million parameters. Or you just take rather standard configs.
I for myself skipped all that equalization after testing it. Perhaps I gave up to early on it.
I think it is much better strategy to get the flaws out of your system and room first and leave the base data untouched.
I actually found the better your system the less room interaction you'll see. I always felt that micro-details, timbre of instruments and reverberation went
down the drain when applying this equalization. I never had the feeling being able to get all this under control by putting even more effort (time&money)
into it.
Just my 2 cents.
Last edited:
All very valid points, soundcheck.
I've further tweaked the wrapper so that I can transmit flac as flac or pcm. I use pcm (for 44.1kHz->96kHz) over wired ethernet, 1gB/s to router.
Here in the tropics full carpeting is best avoided (very high humidity leads to mould), so concrete or marble, sometime hardwood timbre floors are common. Some rugs are used, but the acoustics of our living areas are generally much more live and reverberant than homes I have visited in US and Europe. Sometimes I think manufacturers should specify what sort of home acoustic their speakers are designed for..
For brutefirdrc, which I use in conjunction with your toolbox (thanks!), the switching for different sample rates is automatic. No resampling done.
I've further tweaked the wrapper so that I can transmit flac as flac or pcm. I use pcm (for 44.1kHz->96kHz) over wired ethernet, 1gB/s to router.
I've been working on the room and system for 23 years, and for the near future I have no option to fix the floors (terrazo on concrete) or ceiling (ceiling). The floor / ceiling gives a huge room mode at around 57Hz, 114Hz etc, measured and easily heard (around a concert B-flat). Walls and other fixed features add to it. Bottom line is that I've put in as much treatment and adjustments as the family will take, so drc has to come in.
Here in the tropics full carpeting is best avoided (very high humidity leads to mould), so concrete or marble, sometime hardwood timbre floors are common. Some rugs are used, but the acoustics of our living areas are generally much more live and reverberant than homes I have visited in US and Europe. Sometimes I think manufacturers should specify what sort of home acoustic their speakers are designed for..
Folks.
I just saw mb's response, which reminded me to add another issue which
you step into when doing drc:
To avoid clipping during the filter conversion you need to digital attenuate
the signal prior to processing. I've seen examples reaching from 8!! to 3 db.
Then you'll filter and add the filter losses to that already manipulated signal.
Which then might get manipulated by another digital volume control afterwards.
Let's hope that you didn't upsample or your DAC upsamples
that material to add some more losses.
brutefirdrc on Squeezeboxserver afaik converts everything to flac, which causes another issue ( see my SB Touch Toolbox).
You see. There are some challenges associated to room equalization.
Good luck.
I just saw mb's response, which reminded me to add another issue which
you step into when doing drc:
To avoid clipping during the filter conversion you need to digital attenuate
the signal prior to processing. I've seen examples reaching from 8!! to 3 db.
Then you'll filter and add the filter losses to that already manipulated signal.
Which then might get manipulated by another digital volume control afterwards.
Let's hope that you didn't upsample or your DAC upsamples
that material to add some more losses.
brutefirdrc on Squeezeboxserver afaik converts everything to flac, which causes another issue ( see my SB Touch Toolbox).
You see. There are some challenges associated to room equalization.
Good luck.
Thanks. Yes, there are lots of challenges to drc
, especially if you want to have the digital signal kept as faithfully as possible, which is something I'm aiming for. It's an oxymoron, since drc by definition will mangle the signal. I guess I'm aiming for 'minimalist' drc!- in brutefirdrc at present, AFAIK there is no resampling (except if the player does not support the rate. Eg. if it is 192kHz, then there's no choice by for sox to bring it down to 96kHz. This is something done by sbserver, not by bfdrc.
- the need to reduce the input signal, and to be careful to avoid clipping is definitely real. Bfdrc sends audio in 24-bits to the convolver, and gets it back in 24-bits, even if the audio is originally 16-bits, so that's not too bad. A lot of the correction is attenuation, and gain is limited to N dB (user defined), so overall the loss is around 2 bits. 22-bits of audio is ok for me, in comparison to a 12 dB room mode(!!).
- by default bfdrc sends everything to the player in flac after correction. However I like your idea of playing in pcm on the sbtouch, so I've adapted the wrapper to send pcm to the player (no resampling, just flac decode -> fir filtering -> pcm to sbtouch). The sbtouch sees pcm exactly as it would see pcm from a server configured to send flc->pcm.
custom-conf pipeline:
(flac source) sox -> (24-bit) -> modified brutefir wrapper (64-bit fp internal processing) -> 24-bit pcm -> (wire) to sbtouch with soundcheck mods
Prior to bfdrc I used a Behringer DEQ with a few notches on the parametric eq to manage the main mode + 1st harmonic. It sucked a lot of life out of the sound, but the payback was that bass boom was less intrusive.
With bfdrc and ~6 weeks of filter tuning, soundcheck mods, wrapper tweaking, etc I get:
- much greater neutrality compared to 'direct' playback (no drc, no Behringer deq). The room signature is reduced, but not eliminated, so it still sounds natural
- dramatically improved ability to resolve acoustics between recordings
- improved ability to hear sonic signature of components, eg. its fairly easy now to hear how my four dacs sound different from each other
Hi all,
It's the first time I reply to a thread so I hope you'll pardon me if I don't do things right.
I've played with DRC (the concept) for a few years now, never in depth as, as soundcheck quoted, there're millions of parameters to play with and I don't feel self-confident and knowledgeable enough to experiment that far.
I've played with ACXO, DRC (Denis Sbragion's), Inguz and currently I'm experimenting with Audiolense.
I started with ACXO because at the time it was the easiest way for me to get a hand on DRC. The GUI is simple, well done and makes things very easy for a beginner. And it worked, at least for me. The reason I looked elsewhere is that it is limited to already stored 16/44.1 wav files.
I then tried DRC (Denis Sbragion's) but, as it is proposed, it's too tough for me and because I have a Squeezebox3 I switched to Inguz. Unfortunately because my SB software is implemented within a NAS and not within a computer I was not able to make Inguz work directly within the SB3.
All was not lost as I currently use it as a plugin within foobar 2000.
Both with ACXO and Inguz I had results which I found very positive ie the so-called digital losses (digital artefacts + digital manipulations) were by far outweighted by the gain of
- a lot more balanced frequency response (less dips and bumps) which allowed me to re-discover some recordings (yes I know it's a cliche but I actually discovered that on some guitar lines played by Frank Zappa which I thought were played by another guitar were actually played by a delay+reverb pedal of the main guitar, or that this instrument which I thought was a clarinet within the Adelaide Symphony Orchestra was actually an oboe and the list goes on and on and ...)
- the Prat (Pace, Rhythm And Timing) has been put on turbo due to close to 100% correct phasing
- the soundstage got from a little bit better from 2D to close to 100% 3D especially in depth (with the ASO the percussions were actually 20m further back from the first violins and cellos instead of 3m)
I have to specify one important thing. I don't make any sound test or judgement based on commercial music. It's a principle of mine and the reason is that I was not there when the mastering engineer made the final master. Therefore who am I to make a judgement on how a track should sound if don't have any reference on how it should sound?
I'm lucky enough that I do on average 50 recordings a year, 80% of them being live acoustic concert recordings. All my hardware and software evaluations are done using my recordings or recordings I have been involved in. In any case I was there and I know how it should sound.
On top of that, being my recordings, I don't listen to them from ripped CDs but directly from the computer which actually made them. So to paraphrase soundcheck I suppress some 'digital manipulations': let's say that a CD you use as a test CD has been recorded in 24/96 (which is quite common these days), it has then been downsampled and dithered to 16/44.1 to conform to the Red Book specs. Therefore the CD you're ripping is already a 'digitally manipulated' file. My recordings are not.
soundcheck is right with the facts that
- recording usefull test signals, whichever versions of DRC you use, is a minutious, tedious and lengthy process (and sometimes quite frustrating)
- the best you can afford as far as recording equipment (measurement mic, mic preamp and ADC) the better BUT you can already have obvious audible improvement just with the Behringer ECM 8000 as a measurement mic. Also this recording gear you can borrow or hire as you will need it for a few hours anyway (until you change something in your playback system or in your listening room as soundcheck points it out)
On the fact that outside of the 'sweet spot' DRC is creating a mess it had never happened in my room. Outside of the 'sweet spot' the sound has always been good enough and certainly not worse than without correction.
To sum it up, my listening room being the family living room, although being equipped by some eightnerve acoustical treatment, Digital Room Correction is mandatory for me. The pluses far outweight the minuses.
I hope this can help you DJNUBZ
It's the first time I reply to a thread so I hope you'll pardon me if I don't do things right.
I've played with DRC (the concept) for a few years now, never in depth as, as soundcheck quoted, there're millions of parameters to play with and I don't feel self-confident and knowledgeable enough to experiment that far.
I've played with ACXO, DRC (Denis Sbragion's), Inguz and currently I'm experimenting with Audiolense.
I started with ACXO because at the time it was the easiest way for me to get a hand on DRC. The GUI is simple, well done and makes things very easy for a beginner. And it worked, at least for me. The reason I looked elsewhere is that it is limited to already stored 16/44.1 wav files.
I then tried DRC (Denis Sbragion's) but, as it is proposed, it's too tough for me and because I have a Squeezebox3 I switched to Inguz. Unfortunately because my SB software is implemented within a NAS and not within a computer I was not able to make Inguz work directly within the SB3.
All was not lost as I currently use it as a plugin within foobar 2000.
Both with ACXO and Inguz I had results which I found very positive ie the so-called digital losses (digital artefacts + digital manipulations) were by far outweighted by the gain of
- a lot more balanced frequency response (less dips and bumps) which allowed me to re-discover some recordings (yes I know it's a cliche but I actually discovered that on some guitar lines played by Frank Zappa which I thought were played by another guitar were actually played by a delay+reverb pedal of the main guitar, or that this instrument which I thought was a clarinet within the Adelaide Symphony Orchestra was actually an oboe and the list goes on and on and ...)
- the Prat (Pace, Rhythm And Timing) has been put on turbo due to close to 100% correct phasing
- the soundstage got from a little bit better from 2D to close to 100% 3D especially in depth (with the ASO the percussions were actually 20m further back from the first violins and cellos instead of 3m)
I have to specify one important thing. I don't make any sound test or judgement based on commercial music. It's a principle of mine and the reason is that I was not there when the mastering engineer made the final master. Therefore who am I to make a judgement on how a track should sound if don't have any reference on how it should sound?
I'm lucky enough that I do on average 50 recordings a year, 80% of them being live acoustic concert recordings. All my hardware and software evaluations are done using my recordings or recordings I have been involved in. In any case I was there and I know how it should sound.
On top of that, being my recordings, I don't listen to them from ripped CDs but directly from the computer which actually made them. So to paraphrase soundcheck I suppress some 'digital manipulations': let's say that a CD you use as a test CD has been recorded in 24/96 (which is quite common these days), it has then been downsampled and dithered to 16/44.1 to conform to the Red Book specs. Therefore the CD you're ripping is already a 'digitally manipulated' file. My recordings are not.
soundcheck is right with the facts that
- recording usefull test signals, whichever versions of DRC you use, is a minutious, tedious and lengthy process (and sometimes quite frustrating)
- the best you can afford as far as recording equipment (measurement mic, mic preamp and ADC) the better BUT you can already have obvious audible improvement just with the Behringer ECM 8000 as a measurement mic. Also this recording gear you can borrow or hire as you will need it for a few hours anyway (until you change something in your playback system or in your listening room as soundcheck points it out)
On the fact that outside of the 'sweet spot' DRC is creating a mess it had never happened in my room. Outside of the 'sweet spot' the sound has always been good enough and certainly not worse than without correction.
To sum it up, my listening room being the family living room, although being equipped by some eightnerve acoustical treatment, Digital Room Correction is mandatory for me. The pluses far outweight the minuses.
I hope this can help you DJNUBZ
I've just realized that I haven't answer properly to DJNUBZ.
Here is a link on DRC only few months ago: http://www.diyaudio.com/forums/pc-based/153446-digital-room-correction-project.html
It appears that diy digital room correction is interesting a lot less people than before while more and more people use computer-based system as their main music source. It's a bit of a contradiction to me.
Furthermore, if one puts room correction per se aside, DRC could be used to 'correct' loudspeakers which are by far the less accurate elements of a playback system. By measuring drivers one by one at 1m on axis it should be possible to DRC-correct it within its own range in frequency, time and phase then, if the physical specs of the speaker allows it, to do the same for the whole speaker at let's say 3m on axis in a pseudo anechoic set up ie as far as possible away from any boundary.
As many DRC softwares allow for digital Xover it could be a way of improving by a huge margin one who'd go active.
I don't have the academic knowledge to know if this is possible or useful so feel free to rectify my weird and wandering ideas.
Here is a link on DRC only few months ago: http://www.diyaudio.com/forums/pc-based/153446-digital-room-correction-project.html
It appears that diy digital room correction is interesting a lot less people than before while more and more people use computer-based system as their main music source. It's a bit of a contradiction to me.
Furthermore, if one puts room correction per se aside, DRC could be used to 'correct' loudspeakers which are by far the less accurate elements of a playback system. By measuring drivers one by one at 1m on axis it should be possible to DRC-correct it within its own range in frequency, time and phase then, if the physical specs of the speaker allows it, to do the same for the whole speaker at let's say 3m on axis in a pseudo anechoic set up ie as far as possible away from any boundary.
As many DRC softwares allow for digital Xover it could be a way of improving by a huge margin one who'd go active.
I don't have the academic knowledge to know if this is possible or useful so feel free to rectify my weird and wandering ideas.
@chaimovitch
I guess we pretty much got the same view about the subject.
If you talk about your studio as your reference. I'm wondering if you run DRC
also at your studio. How do you make sure that your mix is not "colored" by your studio-gear and studio itself?
Activation: Going active means quite an effort without being sure where all this ends up. If you consider all the discussions about "audiophile" stereo systems, dacs etc. -- going the multichannel route could end up being a major challenge.
Cheers
I guess we pretty much got the same view about the subject.
If you talk about your studio as your reference. I'm wondering if you run DRC
also at your studio. How do you make sure that your mix is not "colored" by your studio-gear and studio itself?
Activation: Going active means quite an effort without being sure where all this ends up. If you consider all the discussions about "audiophile" stereo systems, dacs etc. -- going the multichannel route could end up being a major challenge.
Cheers
Hello,
I am currently using a SB touch with a qnap ts-219 nas. I would like to add a drc plugin. chaimovitch have you find a way to implement drc working with your squeezebox server on your nas ?
I am currently using a SB touch with a qnap ts-219 nas. I would like to add a drc plugin. chaimovitch have you find a way to implement drc working with your squeezebox server on your nas ?
I have DRC working for 2 channel 44.1khz bit all the way to 8 channel 96khz, at 16 and 24 bits.
All without fiddling.
It is documented here:
http://www.diyaudio.com/forums/multi-way/63078-how-pc-xo-132.html#post2358684
All without fiddling.
It is documented here:
http://www.diyaudio.com/forums/multi-way/63078-how-pc-xo-132.html#post2358684
my studio
If you talk about your studio as your reference. I'm wondering if you run DRC also at your studio. How do you make sure that your mix is not "colored" by your studio-gear and studio itself?
[/QUOTE]
Because 'my studio' is actually the actual venue. I record on location that is 'my studio' is most often a church, a cathedral, a school auditorium etc.
Chaimovitch
If you talk about your studio as your reference. I'm wondering if you run DRC also at your studio. How do you make sure that your mix is not "colored" by your studio-gear and studio itself?
[/QUOTE]
Because 'my studio' is actually the actual venue. I record on location that is 'my studio' is most often a church, a cathedral, a school auditorium etc.
Chaimovitch
Hi, I'm having troubles installing drc. Even with the most basic CFLAG settings, I get:
(the message reads "install: skipping directory 'config/*KHz' make: ***[install] error 1")
Any hints what to try?
Rüdiger
Code:
jamballa source # make install
install drc lsconv glsweep /usr/local/bin
install -d /usr/local/share/drc/config/
install config/* /usr/local/share/drc/config/
install: Verzeichnis âconfig/44.1 KHzâ ausgelassen
install: Verzeichnis âconfig/48.0 KHzâ ausgelassen
install: Verzeichnis âconfig/88.2 KHzâ ausgelassen
install: Verzeichnis âconfig/96.0 KHzâ ausgelassen
make: *** [install] Fehler 1(the message reads "install: skipping directory 'config/*KHz' make: ***[install] error 1")
Any hints what to try?
Rüdiger
Linux...next you'll have to compile your damn soundcard driver..
The windows version is already compiled. It's a tricky enough process as it is.
I just can't see how linux is worth the pain. I've tried Ubuntu twice and gone back to Windows. Linux just has a way of making simple things impossible.
The windows version is already compiled. It's a tricky enough process as it is.
I just can't see how linux is worth the pain. I've tried Ubuntu twice and gone back to Windows. Linux just has a way of making simple things impossible.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.