SAC active crossover

[Jan Kijonka]

Hello,
I would like to use the SAC VST plugin (Linquitz-Riley Filters) to crossover the mixed signal into the several bands (SUB, WOOFER, MID, TWEETER) which I would route to the four outputs of the soundcard and then lead to the four separate power amps and speakers.

My question is: The latency caused by the related Filter will only affects the output channel, where the filter is used ? It means, the output channels which would be used for PA speakers with VST plugin filters enabled will have different delay with respect to the other outputs (e.g. Aux) where the filter has not been used ? Or, the output delay is synchronized (the signal processor process all samples from all channels and send it to the outputs at the same time) ?

I thing, it's good idea to use the digital filtering of the signal processed by the CPU, but there are two drawbacks:

1) Different delay for each speaker.
2) Pops/clicks or some error on the outputs could destroy tweeter or mid.

Maybe, there is some way to bypass these drawbacks.


Thank you,
Jan Kijonka

[cgrafx]

Hello,
I would like to use the SAC VST plugin (Linquitz-Riley Filters) to crossover the mixed signal into the several bands (SUB, WOOFER, MID, TWEETER) which I would route to the four outputs of the soundcard and then lead to the four separate power amps and speakers.

My question is: The latency caused by the related Filter will only affects the output channel, where the filter is used ? It means, the output channels which would be used for PA speakers with VST plugin filters enabled will have different delay with respect to the other outputs (e.g. Aux) where the filter has not been used ? Or, the output delay is synchronized (the signal processor process all samples from all channels and send it to the outputs at the same time) ?

I thing, it's good idea to use the digital filtering of the signal processed by the CPU, but there are two drawbacks:

1) Different delay for each speaker.
2) Pops/clicks or some error on the outputs could destroy tweeter or mid.

Maybe, there is some way to bypass these drawbacks.


Thank you,
Jan Kijonka

1. SAC only supports zero latency plugins, including the Linquitz-Riley Filter. There is no delay between each speaker.

2. A stable system will not have any pops or clicks.

[Jan Kijonka]

1. SAC only supports zero latency plugins, including the Linquitz-Riley Filter. There is no delay between each speaker.

2. A stable system will not have any pops or clicks.

I don't understand. There is always some delay from input to the output. In my situation, I measured about 5-6 ms with a scope. In addition, VST plugins can deliver next-negligible delays.

Furthermore, the digital FIR filters (with finite impulse response) cause a signal delay of the duration of half length of the filter essentially, so about 2 ms for the length of the filter 100. (I work in Matlab and design digital filters).

When you use bandpass for MID and lowpass for SUB, you get two filters for bass, but only one filter for sub and you get a different delay for each filter.

[Bob L]

If SAC detects that a plugin is withholding samples (because it needs to cache samples before it can return processed data) then SAC bypasses the plugin, because it will not allow certain signal paths to be late compared to other signal paths... attempting to preserve the phase and signal quality from input to output.

But... it cannot control plugins that cleverly hide any introduced latency by returning zeros for the first few buffers until real data is returned. So it is up to you to decide whether you want to use the plugin or not... does it sound ok... does it introduce some strange phasing that you do not like... etc.

Signal processing does just that... it affects the signal and processes it along the way... you have to decide if the results are to your liking or not.

Bob L

[Jan Kijonka]

If SAC detects that a plugin is withholding samples (because it needs to cache samples before it can return processed data) then SAC bypasses the plugin, because it will not allow certain signal paths to be late compared to other signal paths... attempting to preserve the phase and signal quality from input to output.

But... it cannot control plugins that cleverly hide any introduced latency by returning zeros for the first few buffers until real data is returned. So it is up to you to decide whether you want to use the plugin or not... does it sound ok... does it introduce some strange phasing that you do not like... etc.

Signal processing does just that... it affects the signal and processes it along the way... you have to decide if the results are to your liking or not.

Bob L

Could the plugin be bypassed during the live concert because of detection to loss of some sampes (not processed signal in time) ? How can I recognize it in SAC ? Simply, that some plugin is visually off ?

I have tried the ASIO4ALL drivers for SAC, but the latency was about 10-15 ms, but only 5-6 ms with the original ASIO drivers. There were also a lot of pop/clicks during the playback, sometimes even for the original ASIO drivers. Yes I know, I must set the length and number of buffers to have the signal clear of artefacts, but where is the right threshold to have the signal clear in every situation ? If I understand, SAC can bypass any VST plugin to have the signal clear, but when I turn on a lot of inputs and a lot of internal EQs/limiters etc., and the signal stops to be processed in time, it will starts to create any artefacts caused by miss any samples anyway ?

[Jeff Scott]

I use an RME Raydat Card for my live SAC Rig. Running at 1x64 with the Dual LR plugin for my crossover. Works great.....no Pops or Clicks...no audible delay or phasing. Is there delay from the time the audio signal enters the rig, is processed and leaves the speakers? Yes...but no more than 5 to 7 ms from my measurements That's the equivalent of standing 6 or 7 feet from a signal source.

That's the nature of digital audio..until someone gets the A/D/A to perform faster....this is what we have to live with.

If you want 0ms delay...you'll have to go analog...or spend $$$ on a hardware console that will most likely cost 10x as much and may do somewhat less than the SAC rig.

[Butch Bos]

What audio interface are you using ?
ASIO4ALL will cause to much latency as you found
Link Riley filters work great with no measurable delay

Butch

[Jan Kijonka]

What audio interface are you using ?
ASIO4ALL will cause to much latency as you found
Link Riley filters work great with no measurable delay

Butch

M-Audio 2626 with FireWire interface, windows 7 optimized according to "Windows 7 Tweaks" recommendations. HW: Intel Core2Duo E8400, 4GB RAM. I have tried the latest ASIO4ALL at first, but SAC crashed with buffer below 96 samples. There were many artifacts on the outputs.

Then I measured the latency with scope (CH1 scope input: 10 Hz signal to the input of the soundcard, CH2 scope input: any output of the soundcard). I measured latency between the input/output about 15 ms or more.

Then I found out the original ASIO driver delays the signal much less and I can set 1x buffer 64. However, it appeared occasionally some a pop / click on he output independently on the VST R-L plugin enabled/disabled.

I didn't test "Force Single CPU", I don't know if it could be better in my HW configuration.

[mr_es335]

Hello,

I have tried the ASIO4ALL drivers for SAC...

...Jan must be using the onboard audio, as most current, if not all audio interface comes with ASIO drivers. He may be like me, who test things out on a non-rig system for testing purposes. But then, I expect not to obtain great results. So, Jan, I guess a hardware update is in order.

PS: Just saw your posting...what FireWire card/device are you using...this is very, very important. It must only be TI.

[Jan Kijonka]

Hello,
...Jan must be using the onboard audio, as most current, if not all audio interface comes with ASIO drivers. He may be like me, who test things out on a non-rig system for testing purposes. But then, I expect not to obtain great results. So, Jan, I guess a hardware update is in order.

PS: Just saw your posting...what FireWire card/device are you using...this is very, very important. It must only be TI.

Hi,




I didn't know that each sound card has an original ASIO drivers. I use the older M-Audio ProFire 2626. I assumed that the ASIO drivers that I found in the ASIO protocol list box in SAC are some default ASIO drivers included with SAC.

I use the Sound Blaster Audigy2 FireWire onboard port for interconection.