Understanding the State of the Art of Digital Room Correction

[Rufus T. Firefly]

Nice, this was a bit over my head at times but I have a firmer grasp on the process.

Thanks!

 

[dlaloum]

No but I think there is a light version of Dirac with such a limit.

So if Dirac Live does full frequency range Time and Frequency domain correction- then one would assume it would do phase correction throughout the same range?

Not so critical for those with "crossoverless" speakers... but many such speakers have a crossover for the Woofer, even if crossoverless between mid/tweeter.... and one would hope that Dirac would correct that?

 

[Newman]

…the problem with low frequency room corrections is it can result in rather large audio delays and subsequent audio/video lip sync issues. Until we have video processors than can deliberately add delay to the video signal to allow the audio DSP to catch up, its not much good to the HT guys.

Like when you go to your tv settings to adjust lip sync, its always 0-200ms audio delay. Like why not -200 to +200ms, and the minus is adding a delay to the video only? Most times the audio is lagging the video, why would I wanna add more audio delay?

You can expect the video to take longer to process than audio. So audio delay is needed.

 

[dlaloum]

All "surround" gear has had the ability to adjust audio delay since day dot.... (well at least since the 80's!) so I would not have thought that this would be a big deal!?

 

[amirm]

All "surround" gear has had the ability to adjust audio delay since day dot.... (well at least since the 80's!) so I would not have thought that this would be a big deal!?

They can delay but retarding the audio is only a feature of the latest AV products. A FIR filter 65K taps at 48 kHz will have a delay of more than a second. So you need to add a lot of frame delays to video. This is easy in a software player but not sure the hardware capable ones can go this far.

 

[amirm]

To add to above, memory to add delay to audio can be very small. In sharp contrast, storing a bunch of video frames to match delayed audio can take a lot of space.

 

[dlaloum]

They can delay but retarding the audio is only a feature of the latest AV products. A FIR filter 65K taps at 48 kHz will have a delay of more than a second. So you need to add a lot of frame delays to video. This is easy in a software player but not sure the hardware capable ones can go this far.

Ahh, ok, yes, lots kit can add delay to the audio, but we are talking in milliseconds, not seconds.... my old MC1 maxed at 60ms, my more recent Integra DTR 70.4 maxes out at 800ms... so starting to get into the right order of magnitude.

It will be interesting to see what the current generation Integra DRX 3.4 I have on order will do - given it has Dirac Live onboard, and therefore there may be a higher time impost...

although memory is now cheap, we don't hear of large memory buffers being fitted to AVR's... the previous generation ABT2010 processors (as used by Denon and others) - had 256Mb of RAM (DDR2) and the ability to use 2 "memory devices" - so presumably could have (a lot?) more RAM.

That was current 10 years ago:

http://static6.arrow.com/aropdfconversion/392d5574d367159f0172da4b99ecbe0e1e4101ca/abt2010productbriefproductbriefapr2011.pdf

It should not be a big deal, providing an AVR with the ability to delay either audio or video by a couple of seconds...

 

[cany89]

@mitchco I learned quite a lot in 2 hours. Subscribed to your channel as well.

Thanks for the lecture!

 

[McFly]

All "surround" gear has had the ability to adjust audio delay since day dot.... (well at least since the 80's!) so I would not have thought that this would be a big deal!?

Not trying to adjust audio delay. Trying to add video delay, to allow for possible seconds worth of outboard FIR filtering.

For example; Particularly Using TV as the source and streaming services via the TVs apps. If you output your audio to an outboard processor of your choice (say, miniDSP 2x4HD) the TV knows that the audio and video was synced when it sent the audio away out the optical or HDMI ARC outlet to be converted in a DAC, but the DSP processor can add significant delays if you don’t just stick to IIR filtering. The TV only has an option to add more audio delay!

But I want negative audio delay - which is impossible, so, add video delay! TVs have RAM now, surely they could hold 24-48 4K frames in there? 25-50ish mbps is it not? Nothing to a 8gb stick of RAM.

I hope this will be a feature on TVs soon, and I’ll be first in line if it does.

 

[thorvat]

They can delay but retarding the audio is only a feature of the latest AV products. A FIR filter 65K taps at 48 kHz will have a delay of more than a second. So you need to add a lot of frame delays to video. This is easy in a software player but not sure the hardware capable ones can go this far.

FIR filter with 64K taps (65536) at 48kHz would actually have a delay of 683ms (0.683 sec). More commonly used filter at 44.1kHz with same # of taps will have delay of 743ms and filter at 96kHz will have delay of 341ms.

 

[fluid]

Not trying to adjust audio delay. Trying to add video delay, to allow for possible seconds worth of outboard FIR filtering.

Jriver has Lip sync Video delay adjustment, it does have a live input and DSP capabilities and some form of TV Integration but I have never used that to know if would work for you.

https://wiki.jriver.com/index.php/Lip_Sync

FIR filter with 64K taps (65536) at 48kHz would actually have a delay of 683ms (0.683 sec).

With a centred Impulse and rounded up to three decimal places or the nearest millisecond

 

[ta240]

Dirac isn't close to these...watch the video as Mitch will explain it all.

Bummer. That mini-DSP has always been a bit tempting; but without a computer in the audio chain the software options aren't a possibility. But that does explain why his after charts look better than any I've seen before.

I haven't had a chance to watch the video yet.

..... I hate giving presentations as I am more of a sit at the bar and chat kind of guy......

That might make for a really interesting channel. Bar chats about audio. I find a lot of people's videos are much more engaging when they are talking to someone rather than just the camera.

Nice, this was a bit over my head at times but I have a firmer grasp on the process.

Thanks!

For me, that is a big advantage to videos; I can coast through the bits that are too technical easier than when reading.

 

[mdsimon2]

Bummer. That mini-DSP has always been a bit tempting; but without a computer in the audio chain the software options aren't a possibility.

I haven't had a chance to watch the video yet.


That might make for a really interesting channel. Bar chats about audio. I find a lot of people's videos are much more engaging when they are talking to someone rather than just the camera.



For me, that is a big advantage to videos; I can coast through the bits that are too technical easier than when reading.

That computer can be as simple as a $50 RPi and there are very simple ways to get digital audio in to the RPi. In the simplest cases this can all be done in one USB DAC with digital input. Signal path looks something like: digital audio source -> DAC digital audio input (TOSLINK/SPDIF/AES) -> DAC USB to RPi -> processing in RPi -> RPi USB to DAC. Of course this requires just one USB connection between the RPi and DAC as USB is bi-directional.

Michael

 

[Draki]

That computer can be as simple as a $50 RPi and there are very simple ways to get digital audio in to the RPi. In the simplest cases this can all be done in one USB DAC with digital input. Signal path looks something like: digital audio source -> DAC digital audio input (TOSLINK/SPDIF/AES) -> DAC USB to RPi -> processing in RPi -> RPi USB to DAC. Of course this requires just one USB connection between the RPi and DAC as USB is bi-directional.

Michael

Can you point to a specific sample and the detailed connection(s)? Thanks ...

 

[mdsimon2]

Can you point to a specific sample and the detailed connection(s)? Thanks ...

MOTU Ultralite Mk5 and Okto DAC8 pro are examples of DACs that will do this. MOTU has TOSLINK and SPDIF inputs and Okto has AES inputs in addition to bi-directional USB audio.

You can also add a HAT like the Hifiberry DIGI+ I/O or DAC+ DSP which will you give you a TOSLINK input for the RPi and then use a separate USB DAC. I've done this with a MOTU M4 DAC.

Michael

 

[dwkdnvr]

Dave, if you check out the video you will see that DRC-FIR makes the list for SOTA DRC If you decide to go digital XO's one day, then yah you would need to go with one of the commercial versions, which indeed are Windows programs.

Interesting that it still stands up. The best system I've personally owned was build around DRC-FIR + BruteFIR back in 2006 or so. DRC-FIR was as user-unfriendly as anything I've ever used with a myriad of lightly documented parameters to tweak and tune, but in this specific system (where I took care of early reflection and speaker positioning as best I could before correction), it was basically point-and-shoot. Although to be fair I *think* this was before Denis added the output plots which would probably help greatly in understanding the tweaking process.

The one thing that I believe DRC-FIR lacked that Acourate emphasized was inter-channel similarity measure/correction. I think DRC-FIR was a single channel system where each channel was independently corrected to the target; I find it plausible that sacrificing 'absolute' compliance with the target to get better L/R matching could be an improvement.

 

[mitchco]

Interesting that it still stands up. The best system I've personally owned was build around DRC-FIR + BruteFIR back in 2006 or so. DRC-FIR was as user-unfriendly as anything I've ever used with a myriad of lightly documented parameters to tweak and tune, but in this specific system (where I took care of early reflection and speaker positioning as best I could before correction), it was basically point-and-shoot. Although to be fair I *think* this was before Denis added the output plots which would probably help greatly in understanding the tweaking process.

The one thing that I believe DRC-FIR lacked that Acourate emphasized was inter-channel similarity measure/correction. I think DRC-FIR was a single channel system where each channel was independently corrected to the target; I find it plausible that sacrificing 'absolute' compliance with the target to get better L/R matching could be an improvement.

Hi @dwkdnvr, cool! Yes, Denis got the basics of psychoacoustic filtering and frequency dependent windowing down pretty well. To help a bit with the user unfriendliness, gmad on diyAudio has some scripts and insights as well: https://www.diyaudio.com/forums/ful...ectrical-loudspeaker-correction-networks.html

Also on diyAudio, wesayso (Ronald) has some excellent measurements and commentary on using DRC-FIR with his https://www.diyaudio.com/forums/full-range/242171-towers-25-driver-range-line-array.htm

Agreed on interchannel similarity. That was also noted in JJ's presentation and in the conclusions which I included in the video.

There is a widening gap between the commercial DRC applications that are being developed versus DRC-FIR. But if you like to tinker, and as you have already heard with your own ears, DRC-FIR can produce excellent results.

 

[Newman]

It all seems terribly limiting while we still can’t get a PC to seamlessly deliver 11 to 15 channels of audio and accept numerous HDMI inputs from a number of sources, and play the kind of role that an AVR can play in a multichannel home audio system.

 

[JRS]

Hi @dwkdnvr, cool! Yes, Denis got the basics of psychoacoustic filtering and frequency dependent windowing down pretty well. To help a bit with the user unfriendliness, gmad on diyAudio has some scripts and insights as well: https://www.diyaudio.com/forums/ful...ectrical-loudspeaker-correction-networks.html

Also on diyAudio, wesayso (Ronald) has some excellent measurements and commentary on using DRC-FIR with his https://www.diyaudio.com/forums/full-range/242171-towers-25-driver-range-line-array.htm

Agreed on interchannel similarity. That was also noted in JJ's presentation and in the conclusions which I included in the video.

There is a widening gap between the commercial DRC applications that are being developed versus DRC-FIR. But if you like to tinker, and as you have already heard with your own ears, DRC-FIR can produce excellent results.

Hi Mitch--thanks for the presentation. I was wondering why the latencies are so long? I was poking around and found this claim made by Brute-FIR:

With a massive convolution configuration file setting up BruteFIR to run 26 filters, each 131072 taps long, each connected to its own input and output (that is 26 inputs and outputs), meaning a total of 3407872 filter taps, a 1 GHz AMD Athlon with 266 MHz DDR RAM gets about 90% processor load, and can successfully run it in real time. The sample rate was 44.1 kHz, BruteFIR was compiled with 32 bit floating point precision, and the I/O delay was set to 375 ms. The sound card used was an RME Audio Hammerfall.

I guess I don't understand how dedicated 500MHz SHARC's can't seem to manage anything close? Granted my understanding is superficial, but it seems that most DSP speaker correction isn't coming close to this figure.

​

 

[Daverz]

Interesting that it still stands up. The best system I've personally owned was build around DRC-FIR + BruteFIR back in 2006 or so. DRC-FIR was as user-unfriendly as anything I've ever used with a myriad of lightly documented parameters to tweak and tune, but in this specific system (where I took care of early reflection and speaker positioning as best I could before correction), it was basically point-and-shoot. Although to be fair I *think* this was before Denis added the output plots which would probably help greatly in understanding the tweaking process.

The one thing that I believe DRC-FIR lacked that Acourate emphasized was inter-channel similarity measure/correction. I think DRC-FIR was a single channel system where each channel was independently corrected to the target; I find it plausible that sacrificing 'absolute' compliance with the target to get better L/R matching could be an improvement.

Yeah, I just use the DRC-FIR defaults for the "normal" configuration. I simply don't have the patience to tweek a bunch of arcane parameters and listen for subtle changes.

EDIT: It looks like the corrected step response between left and right speakers could use some tightening up.