Hi, cowanrg,
Sounds like a fantastic card, you won't be disappointed with PC XO when you've got it working.
I'm just watching Star Wars Episode I through DRC and the 'new' Bass Management plugin, and it sounds awesome. It's so hard to explain what the differences are, but I'm hearing things like they sounded in the cinema when I first watched it.
For instance, the light sabers made a seriously dynamic 'snap' sound in the cinema when they hit the robots etc. Hearing the same movie at home made me start thinking they re-mix the soundtrack completely for DVD, but now I've got DRC working, all the best bits are back!
OzOnE.
btw, thanks to Angelo Farina for the Bass Management code and everyone else who have helped me so far. My humble home cinema setup probably only cost me around £1600 if I include the HTPC, and it sounds more like an £8000 system now.
Sounds like a fantastic card, you won't be disappointed with PC XO when you've got it working.
I'm just watching Star Wars Episode I through DRC and the 'new' Bass Management plugin, and it sounds awesome. It's so hard to explain what the differences are, but I'm hearing things like they sounded in the cinema when I first watched it.
For instance, the light sabers made a seriously dynamic 'snap' sound in the cinema when they hit the robots etc. Hearing the same movie at home made me start thinking they re-mix the soundtrack completely for DVD, but now I've got DRC working, all the best bits are back!
OzOnE.
btw, thanks to Angelo Farina for the Bass Management code and everyone else who have helped me so far. My humble home cinema setup probably only cost me around £1600 if I include the HTPC, and it sounds more like an £8000 system now.
OzOnE_2k3 said:Hi, cowanrg,
Sounds like a fantastic card, you won't be disappointed with PC XO when you've got it working.
I'm just watching Star Wars Episode I through DRC and the 'new' Bass Management plugin, and it sounds awesome. It's so hard to explain what the differences are, but I'm hearing things like they sounded in the cinema when I first watched it.
For instance, the light sabers made a seriously dynamic 'snap' sound in the cinema when they hit the robots etc. Hearing the same movie at home made me start thinking they re-mix the soundtrack completely for DVD, but now I've got DRC working, all the best bits are back!
OzOnE.
btw, thanks to Angelo Farina for the Bass Management code and everyone else who have helped me so far. My humble home cinema setup probably only cost me around �1600 if I include the HTPC, and it sounds more like an �8000 system now.
i guess i have some good things to look forward to then. my system already sounds many times better than theaters, even the really nice ones. im lucky though, as my day job, im a dealer for many high end lines, so i get special deals. ive gotten some nice stuff along the way.
i got my lynx the other day and got it all hooked up. wow, what a NICE card. its still breaking in, but it sounds great out of the box.
anyways, im having one helluva time getting it setup in console. i CANNOT get anything to happen. using the ASIO ins and outs, i cant 'grab' the output from either foobar or theatertek. i just tried something simple, taking 1 and 2 inputs (left and right main outs) and patching them into 5 and 6 outputs. nothing. i tried all the ins, and nothing has sound on them.
in the ASIO setup for the lynx (in console), for the IN section, it only lets me choose the "record in", which dont seem to have anything going into them. the outputs, or the "play" has sound in the mixer, but the records are only active when the analog in jacks on the card. is there a way i can patch the output from foobar or whatever into the record in?
anyways, im having one helluva time getting it setup in console. i CANNOT get anything to happen. using the ASIO ins and outs, i cant 'grab' the output from either foobar or theatertek. i just tried something simple, taking 1 and 2 inputs (left and right main outs) and patching them into 5 and 6 outputs. nothing. i tried all the ins, and nothing has sound on them.
in the ASIO setup for the lynx (in console), for the IN section, it only lets me choose the "record in", which dont seem to have anything going into them. the outputs, or the "play" has sound in the mixer, but the records are only active when the analog in jacks on the card. is there a way i can patch the output from foobar or whatever into the record in?
Hi, cowanrg,
I think you've stumbled onto the common problem with most cards - The thing is, you can generally only use the ASIO outputs for one program at a time, so either Console itself is using the ASIO outputs, or foobar or whatever else is using the ASIO outputs.....
The only way around it for most people (including me) is the fact that my drivers allow me to loop the WDM / DirectSound 'outputs' into the ASIO 'inputs' so that Console can pick up the audio on the ASIO inputs, then of course apply the VST effects and send it to the ASIO outputs / speakers. (This is the so-called 'DirectWire' feature on the Audiotrak card.)
You would think that the more professional cards would have these type of loop-back features in the drivers, but they often don't afaik? This is one of the main reasons I was drawn to getting an Audiotrak (as well as the low price!)
So it goes something like: foobar > WDM > ASIO inputs > Console > ASIO outputs.
That way, only Console is using ASIO - any audio I play that I want to go through Console actually plays through the WDM / DS 'outputs' on my sound card drivers and does NOT play directly through ASIO.
This same 'clash' of program would happen if you tried to use Console to output to WDM / DS and then use foobar etc. to also use WDM / DS. The only way for me atm is to loop from WDM to ASIO using my drivers? I'm sure there must be other software out there which would fix the problem for other cards, but I've yet to find anything usable.
I know there are things like Virtual Audio Cable and ASIO4ALL, but I don't think they do want you'll want. I'll have a search around though, as I'd be interested to see which solutions are available.
Also, you'll often find that you can only get 2-channel inputs and outputs when using DS mode in Console. I can only get 8 channels in Console on my Audiotrak card when using ASIO.
When I play DVD's through Console, it obvisouly uses 8-channel DS outputs, so this is also looped from DS to ASIO to get Console to work on all 8 channels. I've attached a screenshot showing the basic flow of audio when I play DVD's so you can see how it works. Please don't laugh at my 'drawings'!
Sorry I can't be of more help atm.
OzOnE.
I think you've stumbled onto the common problem with most cards - The thing is, you can generally only use the ASIO outputs for one program at a time, so either Console itself is using the ASIO outputs, or foobar or whatever else is using the ASIO outputs.....
The only way around it for most people (including me) is the fact that my drivers allow me to loop the WDM / DirectSound 'outputs' into the ASIO 'inputs' so that Console can pick up the audio on the ASIO inputs, then of course apply the VST effects and send it to the ASIO outputs / speakers. (This is the so-called 'DirectWire' feature on the Audiotrak card.)
You would think that the more professional cards would have these type of loop-back features in the drivers, but they often don't afaik? This is one of the main reasons I was drawn to getting an Audiotrak (as well as the low price!)
So it goes something like: foobar > WDM > ASIO inputs > Console > ASIO outputs.
That way, only Console is using ASIO - any audio I play that I want to go through Console actually plays through the WDM / DS 'outputs' on my sound card drivers and does NOT play directly through ASIO.
This same 'clash' of program would happen if you tried to use Console to output to WDM / DS and then use foobar etc. to also use WDM / DS. The only way for me atm is to loop from WDM to ASIO using my drivers? I'm sure there must be other software out there which would fix the problem for other cards, but I've yet to find anything usable.
I know there are things like Virtual Audio Cable and ASIO4ALL, but I don't think they do want you'll want. I'll have a search around though, as I'd be interested to see which solutions are available.
Also, you'll often find that you can only get 2-channel inputs and outputs when using DS mode in Console. I can only get 8 channels in Console on my Audiotrak card when using ASIO.
When I play DVD's through Console, it obvisouly uses 8-channel DS outputs, so this is also looped from DS to ASIO to get Console to work on all 8 channels. I've attached a screenshot showing the basic flow of audio when I play DVD's so you can see how it works. Please don't laugh at my 'drawings'!
Sorry I can't be of more help atm.
OzOnE.
Attachments
cowanrg said:
Internal loopback is done using the jumpers on the card. Its been awhile now but if I remember you lose an 8 channel ADAT to allow it serve as the loopback. Its in the manual so take a look at that first.
Route the playback into the ADAT loopback, pickup the ADAT in console, process then output and route to physical outs. I'm pretty sure that's what how I worked it with mine.
Fr. Allocator, Acorate or DRC with selfmade XO with Console
Have anyone compared the solution offered by Thuneau (Frequency Allocator) with FIR filters made with Acourate or self made XO made in Octave / Matlab and room correction with DRC?
Both filters made with Acourate and XO made with Octave/ Matlab can be run in Console using the ConvolverVST or Voxengo Prestine Space.
What I want to do:
Use a software cross over between my stereo bass speaker elements and the top speakers. (I use a self-made passive filter between the descant element and the midrange.)
Use a software room correction system.
I own a Lynx Two B soundcard which is connected directly to four B&O Icepower 600W amplifiers. I use the Console program to connect the VST plugins.
I have already tried the following solutions:
Foobar player with different XO-plugins and convolver plugin with filters made with DRC.
Waves LinEq as cross over with Voxengo Prestine Space with filters made with DRC.
I have started experimenting with the Frequency Allocator and I think this solution sounds better then Waves LinEq as XO.
I would really much have liked to compare this solution with the Acourate solution but I think this program is way too expensive to buy just for a comparison test.
The last solution is to try to build my own filters using Octave or Matlab and create a correction filter using DRC. Does anyone know how to merge (convolve) an XO fir filter with the DRC FIR filter?
Have anyone compared the solution offered by Thuneau (Frequency Allocator) with FIR filters made with Acourate or self made XO made in Octave / Matlab and room correction with DRC?
Both filters made with Acourate and XO made with Octave/ Matlab can be run in Console using the ConvolverVST or Voxengo Prestine Space.
What I want to do:
Use a software cross over between my stereo bass speaker elements and the top speakers. (I use a self-made passive filter between the descant element and the midrange.)
Use a software room correction system.
I own a Lynx Two B soundcard which is connected directly to four B&O Icepower 600W amplifiers. I use the Console program to connect the VST plugins.
I have already tried the following solutions:
Foobar player with different XO-plugins and convolver plugin with filters made with DRC.
Waves LinEq as cross over with Voxengo Prestine Space with filters made with DRC.
I have started experimenting with the Frequency Allocator and I think this solution sounds better then Waves LinEq as XO.
I would really much have liked to compare this solution with the Acourate solution but I think this program is way too expensive to buy just for a comparison test.
The last solution is to try to build my own filters using Octave or Matlab and create a correction filter using DRC. Does anyone know how to merge (convolve) an XO fir filter with the DRC FIR filter?
harruharru,
The trial version of Acourate will do Butterworth and Linkwitz-Riley linear filters.
You can also create Transient Perfect crossovers using the Excel Spreadsheets from John Kreskovsky and then convert these into filters for a convolver using oehlrich’s program frdconv. Have a look at page 39 of this thread.
The trial version of Acourate will do Butterworth and Linkwitz-Riley linear filters.
You can also create Transient Perfect crossovers using the Excel Spreadsheets from John Kreskovsky and then convert these into filters for a convolver using oehlrich’s program frdconv. Have a look at page 39 of this thread.
I tried to create Transient Perfect crossovers using the Excel Spreadsheets from John Kreskovsky. I made a simple 2 way cross over at 80 Hz with 10th order Bessel caracteristics . The filters worked but did not sound as good as the simple filters made with this script using Octave:
n=10; # exponent for filter size
fxo1=80; # sub crossover frequency in Hz
fs=44100; # sample rate in Hz
# End of user parameters
k=2^n; # order of filter
fn1=2*fxo1/fs; # normalized subsonic cutoff frequency
i=linspace(1,k,k); # k-tap filter array
f_lo=linspace(0,200,512); # for plot of low end of freq response
hir=fir1(k,fn1,'high','scale'); # high-pass impulse response
sub=fir1(k,fn1,'low','scale'); # low-pass ir
hirtxt=hir(i); # my klugey way of taking k elements
subtxt=sub(i);
save -ascii mid_high.txt hirtxt
save -ascii sub_low.txt subtxt
(For those of you who want to try, download Octave (free under GNU) , use notepad to save script as test.m (filename with .m extension), type test (filename) in Octave)
I used convolverVST to run the filters. (To use the filters in convolverVST you must use the makeir program. Documentation can be found at the convolver download site.)
I might have done something wrong when I produced the filters using the Excel Spreadsheets from John Kreskovsky. I do not know. But the filters made with octave sounded best.
I have then done a comparison between the Octave filters and the Fr. Allocator. They both sound good, and quite similar but I believe the Octave filter sounded a bit more detailed.
Do anyone knowif it is possible to merge DRC filters with the XO filters to create one set of filters? It is only possible to use one instance of convolverVST when using console.
I have also tried to find some comparison tests between DRC and Acourat but I have so far found none.
n=10; # exponent for filter size
fxo1=80; # sub crossover frequency in Hz
fs=44100; # sample rate in Hz
# End of user parameters
k=2^n; # order of filter
fn1=2*fxo1/fs; # normalized subsonic cutoff frequency
i=linspace(1,k,k); # k-tap filter array
f_lo=linspace(0,200,512); # for plot of low end of freq response
hir=fir1(k,fn1,'high','scale'); # high-pass impulse response
sub=fir1(k,fn1,'low','scale'); # low-pass ir
hirtxt=hir(i); # my klugey way of taking k elements
subtxt=sub(i);
save -ascii mid_high.txt hirtxt
save -ascii sub_low.txt subtxt
(For those of you who want to try, download Octave (free under GNU) , use notepad to save script as test.m (filename with .m extension), type test (filename) in Octave)
I used convolverVST to run the filters. (To use the filters in convolverVST you must use the makeir program. Documentation can be found at the convolver download site.)
I might have done something wrong when I produced the filters using the Excel Spreadsheets from John Kreskovsky. I do not know. But the filters made with octave sounded best.
I have then done a comparison between the Octave filters and the Fr. Allocator. They both sound good, and quite similar but I believe the Octave filter sounded a bit more detailed.
Do anyone knowif it is possible to merge DRC filters with the XO filters to create one set of filters? It is only possible to use one instance of convolverVST when using console.
I have also tried to find some comparison tests between DRC and Acourat but I have so far found none.
harruharru said:
You load the convolver with the drc filter and play the xo filter through the convolver - the output is the convolution between the the drc and the xo filter - this must be saved and then used as the new filter set
You might want to look into the aurora plug-ins;
http://pcfarina.eng.unipr.it/Aurora_XP/index.htm
Please note that the length of this convolution result is n+m-1 with n=length of XO filter (number of taps) and m=length of DRC filter.
This will cause more load to ConvolverVST unless you cut the result to a proper length accompanied by a proper windowing.
Uli
www.acourate.com
Hi all,
I've been searching and searching for months now for a way of doing DRC without using a PC. In other words, I've been looking for a simple (and cheap) DSP board which could do the convolution with chosen DRC filter impulses and process the many channels needed for surround tracks.
From what I've heard, even with a DSP, convolution is quite intensive, and it may require more than one chip to do the work. I don't have any experience with DSP programming, but I program FPGA chips and PIC chips quite a lot. Maybe an FPGA could be used for this?
The idea would be to build or find a module which could be used as a standalone DRC processor. This would need to be cheaper than a PC, or it would defeat the object. Is it even possible to do multi-channel convolution relatively cheaply?
If this could be achieved, the processor could be used directly on digital inputs and outputs. The idea of this being to modify existing AV receivers so the DRC processing is kept in the digital domain right up to the receiver's own DACs.
I have a few old ECHO Gina / Layla PCI boards here with Motorola 56301 DSP chips on, which I could use for experimenting with, but I can't find any software simple enough to program the chips with. I can build simple parallel port JTAG interfaces, or program EPROM chips directly, but I don't know how to go about communicating with the Motorola chip? Does anyone have any experience in this area?
I would consider buying some other DSP chips, as I know there are much faster ones available now. I know there are units like the Sony DRE-S777, and Yamaha SREV1, but these units are hideously expensive. Surely it can be done for a few hundred dollars / Euros / Pounds?
http://vintageking.com/New-Brands/Sony/Sony-DRE-S777-2
http://aes.harmony-central.com/109AES/Content/Yamaha/PR/SREV1.html
On a side topic, I've just bought a small MU502 mixer (don't laugh!).....
http://www.activemusician.com/item--EM.PHN-MU502
And I've bodged together a quick 'n' nasty preamp board using the following transistor circuit (figure 6)....
http://sound.westhost.com/project93.htm
The preamp seems to work extremely well with the mixer and a WM-61A capsule (especially considering what my prototype looks like!). I'm now just waiting for an XLR cable and socket so I can do some sweep recordings.
If the home-built WM-61A stuff doesn't work well soon, I'll have to buy an ECM8000. Having said that, I'm listening to music through a DRC filter made with the WM-61A, and it sounds amazing so far (this was before I got the new mixer, but the DRC results were always inconsistant.)
So, does anyone have any suggestions for the DSP stuff? I'm itching to get something working, and I can dedicate a fair amount of time to the project once I know where to start.
OzOnE.
I've been searching and searching for months now for a way of doing DRC without using a PC. In other words, I've been looking for a simple (and cheap) DSP board which could do the convolution with chosen DRC filter impulses and process the many channels needed for surround tracks.
From what I've heard, even with a DSP, convolution is quite intensive, and it may require more than one chip to do the work. I don't have any experience with DSP programming, but I program FPGA chips and PIC chips quite a lot. Maybe an FPGA could be used for this?
The idea would be to build or find a module which could be used as a standalone DRC processor. This would need to be cheaper than a PC, or it would defeat the object. Is it even possible to do multi-channel convolution relatively cheaply?
If this could be achieved, the processor could be used directly on digital inputs and outputs. The idea of this being to modify existing AV receivers so the DRC processing is kept in the digital domain right up to the receiver's own DACs.
I have a few old ECHO Gina / Layla PCI boards here with Motorola 56301 DSP chips on, which I could use for experimenting with, but I can't find any software simple enough to program the chips with. I can build simple parallel port JTAG interfaces, or program EPROM chips directly, but I don't know how to go about communicating with the Motorola chip? Does anyone have any experience in this area?
I would consider buying some other DSP chips, as I know there are much faster ones available now. I know there are units like the Sony DRE-S777, and Yamaha SREV1, but these units are hideously expensive. Surely it can be done for a few hundred dollars / Euros / Pounds?
http://vintageking.com/New-Brands/Sony/Sony-DRE-S777-2
http://aes.harmony-central.com/109AES/Content/Yamaha/PR/SREV1.html
On a side topic, I've just bought a small MU502 mixer (don't laugh!).....
http://www.activemusician.com/item--EM.PHN-MU502
And I've bodged together a quick 'n' nasty preamp board using the following transistor circuit (figure 6)....
http://sound.westhost.com/project93.htm
The preamp seems to work extremely well with the mixer and a WM-61A capsule (especially considering what my prototype looks like!). I'm now just waiting for an XLR cable and socket so I can do some sweep recordings.
If the home-built WM-61A stuff doesn't work well soon, I'll have to buy an ECM8000. Having said that, I'm listening to music through a DRC filter made with the WM-61A, and it sounds amazing so far (this was before I got the new mixer, but the DRC results were always inconsistant.)
So, does anyone have any suggestions for the DSP stuff? I'm itching to get something working, and I can dedicate a fair amount of time to the project once I know where to start.
OzOnE.
Hi, peufeu,
Does that mean it would be very expensive to do multi-channel convolution using DSP chips?
It seems a shame that a PC needs to be used to make DRC cost effective. The main problems with a PC that I see are power consumption, the time it would take to boot up, and over-complexity. How long does it take for a 1GHz Via EPIA board to boot BruteFIR from USB stick (or any other PC)?
Does anyone know how to roughly calculate the DSP MIPS required for a convolution on a single stream of say 48KHz audio on your average £35 / $70 DSP chip?
You'd need to guesstimate the length of the filter and processing precision etc. I personally don't know how to go about calculating this, but it would give us a much better idea of where we are if anyone knows how to to this.
Thanks,
OzOnE.
Does that mean it would be very expensive to do multi-channel convolution using DSP chips?
It seems a shame that a PC needs to be used to make DRC cost effective. The main problems with a PC that I see are power consumption, the time it would take to boot up, and over-complexity. How long does it take for a 1GHz Via EPIA board to boot BruteFIR from USB stick (or any other PC)?
Does anyone know how to roughly calculate the DSP MIPS required for a convolution on a single stream of say 48KHz audio on your average £35 / $70 DSP chip?
You'd need to guesstimate the length of the filter and processing precision etc. I personally don't know how to go about calculating this, but it would give us a much better idea of where we are if anyone knows how to to this.
Thanks,
OzOnE.
About MIPs : well if you use FFT it's not that simple, but if you don't, it's (impulse response length in samples) x (sample rate), nothing more, and the result is in MACs per second (multiply/accumulate).
As for the length of the impulse response, I never tried DRC so I have no idea. Google ?
Note that VIA cpus suck at FP calculations... or you could stick a $100 GeForce and patch brutefir to use GPUFFTW and kick some cray azz... basically 1 GeForce 7900 = 8 dual core xeons)
(http://gamma.cs.unc.edu/GPUFFTW/results.html)
Don't power it off, put it in suspend-to-RAM mode, it will wake up a lot faster.
As for the length of the impulse response, I never tried DRC so I have no idea. Google ?
Note that VIA cpus suck at FP calculations... or you could stick a $100 GeForce and patch brutefir to use GPUFFTW and kick some cray azz... basically 1 GeForce 7900 = 8 dual core xeons)
(http://gamma.cs.unc.edu/GPUFFTW/results.html)
Don't power it off, put it in suspend-to-RAM mode, it will wake up a lot faster.
ShinOBIWAN said:
ShinOBIWAN, I realize this is asking too much to remember how you configured your 1820m but would you please explain the routing procedure for it, it seems confusing, but I'll never give up.
Thanks a bunch,
otmopo3ok
PS: or anyone else using this card for that matter may chime in. Thx.
EMU 1820 m Config
Open up EMU mixer application
First you need an output strip that is capabel of receiving the output of your media player - i. e Waveout type.
Make sure that the output strip received the output from your media player. Now in this Output strip make an insert with asio1/2 output. This will be your input in console.
Mute the Output at the bottom of the strip( main out)
Now add asio output strips to your likeing and channel count.
These Asio output will whow up in console
In each of these make an insert with the desired physical output i.e breakout box 1/2 etc. Mute the output strip in the bottom.
That should do it save the mixer configuration with the desired sample rate you want to run at.
Fire up console and the Asio input/outputs should show up
Open up EMU mixer application
First you need an output strip that is capabel of receiving the output of your media player - i. e Waveout type.
Make sure that the output strip received the output from your media player. Now in this Output strip make an insert with asio1/2 output. This will be your input in console.
Mute the Output at the bottom of the strip( main out)
Now add asio output strips to your likeing and channel count.
These Asio output will whow up in console
In each of these make an insert with the desired physical output i.e breakout box 1/2 etc. Mute the output strip in the bottom.
That should do it save the mixer configuration with the desired sample rate you want to run at.
Fire up console and the Asio input/outputs should show up
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.