A how to for a PC XO.

[Daveis]

Most of my music is 16-bit 44.1 khz CD's. I'd like to dispense with my 6-channel or 8-channel volume control and control my volume digitally. Now, my DAC's are 24/96 capable. So what I was thinking is that I could set my system volume to the max I ever intend to listen it at. And I figure out a way to shift the 16-bit music up into the 24-bits of resolution I have to work with. Then I can lower the volume digitally without losing any of the original 16-bits.

Is there a way to do that with either software (on PC) or a hardware device?

[DomKo]

Hi diyer,

I have a problem about console. There is a 'pick-pop' noise when I play with the music through the console.

If I play directly without console, there is no 'pick-pop' noise. The audiotrak sound card generate the good music.

When I play through the console with asio, there is 'pick-pop' noise RANDOMLY happening on each channel.

I think the problem is coming from console setup or configuration. Can you help me? OR, I need to reinstall sound card driver and console again.

My System
2.4G P4
512M RAM
Audiotrak 7.1 HiFi Sound Card

Console 1.6.1
In Console, Foobar> ASIO in > ASIO out to speaker.


Best regards
Dominic Ko

[OzOnE_2k3]

Hi Dominic,

I had similar problems on my HTPC, but mainly when recording the log sweep for the DRC measurements though ASIO....

The problem is most likely down to the buffer settings for ASIO - due to the fact that Console needs to basically record, process, and playback at the same time, you might need to increase the buffer size to get good playback without the crackles and pops.

Try increasing the buffer size (or latency) setting on the Audiotrak driver (under the Config menu I think). Sometimes, the 512 or 1024 settings might be enough to stop the problem.

If that still doesn't work, you should check your IRQ settings - right-click My Computer (or go to System under Control Panel), then go to Hardware > Device Manager, then choose View > "Resources by type", then expand the IRQ branch and look at which IRQ the Audiotrak card is on......

You will often find that a PCI card will "share" it's IRQ with the graphics card or another device which has heavy usage. Although PCI is "supposed" to be able to share IRQs without a problem, I've often found that this is not the case.

If your Audiotrak card is sharing the same IRQ as a device like your graphics card, the Audiotrak might experience glitches due to the large amount of data transfer while the PC is trying to share the IRQ between two or more devices.

The easiest way to change this might be to physically move the Audiotrak card to a different PCI slot, then let Windows re-detect the card....

(Often, a specific PCI slot in the PC will try to share it's IRQ with the graphics card - moving the PCI card will usually stop it sharing with the graphics card. The Audiotrak card might still share an IRQ if you move it, but hopefully the device it ends up sharing with won't affect it Audiotrak so much.)

If you do try moving the sound card, give the IRQ settings another check to see if the IRQ has changed before testing the card again. You might need to set the Audiotrak settings again though (buffer / latency etc.)

Hope that makes sense.

OzOnE.

[ralf]

Hi, who knows about to measure, if DRC did a good Job?

How can one measure over all the Vst Plugins (Convolver)?
Which Equipment do you use?

My measurements (with audiocontroler AC) say, that in the bass region DRC does a good job, but in higher areas it is less smooth. - far away form the Pictures- in the DRC manual.

Ralf

[ralf]

when i use the Rush method, i have the same problem: when saving as 0.24 the impulse is there, when saving 16.8 -no impulse.
Whats wrong?

[OzOnE_2k3]

Hi, Ralf,

I'm also still not sure of the best program to use to see the results of DRC. It would be great if we could do those waterfall plots of the room response, but I can't seem to find a decent program to do them.

ACXO Player has a function for verifying the DRC result, but I've always had trouble getting ACXO to calibrate properly etc. It should be possible to use ACXO Player with your own DRC config files and you could then test the result of the DRC filter quite easily.

I use Smaartlive quite a lot for testing the average frequency response via the microphone, but you really need something like Praxis to give you the move extensive graphs etc. - The short answer is that I still don't really know the best way of verifying the DRC results.

I don't know why the impulse peak doesn't show up with Cool Edit when 16.8 format files are opened? I'm assuming the impulse should look similar if the file was saved as 0.24 then opened as 0.24 again? Obviously the default file type for DRC is 16.8, so problem is strange to me?

I've finally been paid for my new job, so I've also finally bought an ECM8000 which should be here tomorrow. I'll let you all know the results with the ECM8000, but I'm hoping it will be a lot better than with my DIY attempts at building a mic!

OzOnE.

[OzOnE_2k3]

Hi guys 'n' gals,

Just an update - I got my new ECM8000 a few days ago and it worked perfectly first time!

The resulting filters are much clearer and smoother than any of the filters I made with my DIY WM-61A mics. The measurements are now very consistant, and I've made filters for all seven speakers (plus the sub)!

With the DIY mic, the rear bipolar speakers (Mordaunt MSB-20's) seem to confuse the measurements, but now the rear speakers sound amazing. They've got tons of warmth and top end clarity now, and mix extremely well with the sound of the front speakers. As a result, surround pans and effects move around the room properly (even too the sides of the room as they're supposed to).

All the effects have proper weight and snap to them (Star Wars and T2 have never sounded better), and I'm hearing things I never heard before.

One slightly confusing thing is which DRC filter to use for the sub (LFE) channel? When using the Bass Management plugin, the output to the sub is taken from the other 5 / 7 channels, but there is also the dedicated LFE channel (from DVD's) which needs a DRC filter.

I tried measuring the sub on it's own with DRC (rec_imp), but the resulting filter is of a very low level - this is due to DRC trying to "boost" all the missing top end frequencies.

So, to get a filter for the LFE channel, I just use one of the filters from the front channels to see which one sounds best. Since my sub is placed between the front left and center speakers, I tried the filter for the front left speaker on the LFE channel. For some reason though, the filter for the front right speaker seems to sound best?

To measure all the main speakers, I simply leave the Bass Management and speaker delays in place when I use rec_imp. This means that the sub operates at the same time as each main speaker being tested, and it has the proper main / sub crossover filters applied etc.

The only real problem I faced is that the sub output was clipping quite badly. I had to reduce the LFE level (chan 6) on Pristine Space to avoid clipping the output from Bass Management, then turn up the sub's own volume control to acheive the proper output.

One reason for this goes back to the fact that I can't seem to generate a DRC filter for the LFE channel directly - The Bass Management plugin boosts the LFE output by 10dB, so when I use a DRC filter from one of the front speakers on the LFE channel, the 10dB boost clips the output.

I used some test tracks from the "Pod Race" scene in Episode I and looped back the sub output into Cool Edit to see when it clipped. I then reduced the LFE channel level in Pristine Space until it didn't clip any more.

I'm still trying to get my head around the fact that the LFE channel is "recorded" 10dB lower than the other channels, so when it is played back, the sub output has 10dB more headroom and it's peak output (at reference levels) is 10dB louder than any one of the main channels.

(ie. if the loudest output from a main speaker channel is 105dB at reference level, the subwoofer is able to output 115dB.)

I'm still not sure how "proper" bass management is supposed to work, but my setup sounds fantastic so far. Ideally, it would be nice to boost the LFE output externally (before the analog cables connect to the Denon). It wouldn't be a good idea to reduce the levels of the main channels instead, as the volume on the Denon needs to be set very high as it is once the DRC filters are in place!

Once I've measured each speaker then generated DRC filters for each, I load PS back into Console, then I use Realtime Analyser RAD to set the speaker levels very accurately with the ECM8000 in the listening position. I'm sure there are better programs out there for setting the levels, but RAD seems to do a great job, and is easy to use.

I can then set the relative speaker delays using RAD's "Impulse Response" function to get the impulse peak of each speaker to within a millisecond of each other. (I use Voxengo Delay plugins on each output channel in Console to set the actual delays.)

So, in summary, if anyone's thinking of getting into DRC, do yourself a huge favour and buy a decent preamp and measurement mic right from the start. It's probably worked out cheaper for me than trying to build my own mic and getting crappy results. My Phonic mixer and ECM8000 perform consistantly well now, so I just have to sort out the minor issues with speaker levels / target curves / clipping / bass management / mic position etc.

I thought that DRC was supposed to compensate for the relative speaker levels? Does it also work for the delays too? I've tried the full DRC filters as well, but there doesn't seem to be much difference in the relative delays or speaker levels to the MS filters?

If anyone has any suggestions about the speaker levels / bass management when using DRC, I'd be very grateful. I'm by no means an expert on this stuff, but I've learnt a lot so far, and I'm always keen to learn more.

Again, sorry for the long post, but there are many different issues to get across in case other people are looking for help.

OzOnE.

[houstonian]

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)

has anyone tried this?

I have multiple FIR filters on 8 channels and am using about 40% of the CPU on a Smithfield 2.8 GHz dual core.

[OzOnE_2k3]

Hi, houstonian,

I don't think I'll get chance to try that sort of thing any time soon. I also don't use linux, and I'm useless at programming, so it would be a bit of a challenge.

I've just started work as an engineer in a PC shop, so I might be able to get hold of component a bit easier now.

I don't have any problems running Pristine Space on my old 2800+ Athlon XP though. It has eight DRC filters, and only takes about 10% CPU power at the most. Is PS not the same as FIR?

I still wish there was a cheap DSP method out there for doing DRC on. I'd gladly spend out for building the PCBs if such a solution was available. The only thing I've found so far which will do DRC is something like the Sony DRE-777 reverb units, but they are hideously expensive.....

http://vintageking.com/New-Brands/Sony/Sony-DRE-S777-2

Oh well, a PC it is then for now.

OzOnE.

[manisandher]

Hi.

Has anyone had an opportunity to compare a PC XO (Frequency Allocator or Acourate) with the DEQX PDC?

How do they compare?

I've read this whole thread (one of the most interesting, BTW) and am aware that ShinOBIWAN went from PC XO (Console, Waves & Voxengo) to DEQX and back to Frequency Allocator.

I currently have the following setup:
- Foobar2000 (ASIO driver)
- RME Fireface 800 (SPDIF out, word clock in)
- Esoteric D70 (SPDIF in, word clock out)
- 2 x Pass Labs XVR1 (one per channel to give 3-way XO)

I have been advised by others on this site that I would be making a mistake by replacing my XVR1s with the DEQX.

I have to say that my biggest issue with the DEQX is the quality of the hardware:
- no word clock I/O (which I feel is absolutely crazy)
- no Firewire or USB input for direct audio streaming
- outdated DACs (and opamps I think)

Any views most appreciated (especially from ShinOBIWAN, if you're still checking this thread out once in a while).

Mani.