Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[NoWo]

Hi Pieter,

congratulations to your EQs. I like them a lot, seems like you built in a certain healing factor too -:)

One thing is bothering me:
The LinEQ is eating up all my power. Okay, it is an older machine (Dual Athlon 1800), but is it really that much for just two bands? I get only a stuttering sound with my Fireface 400 and 1048 samples. Trying to lower the quality leads to no other result.

NoWo

[Matt_G]

NoWo wrote on Mon, 30 June 2008 00:26

Hi Pieter, congratulations to your EQs. I like them a lot, seems like you built in a certain healing factor too -:) One thing is bothering me: The LinEQ is eating up all my power. Okay, it is an older machine (Dual Athlon 1800), but is it really that much for just two bands? I get only a stuttering sound with my Fireface 400 and 1048 samples. Trying to lower the quality leads to no other result. NoWo

It's probably the plug-in latency that's causing the stuttering. Try increasing the buffer size to at least 2048. I'm reporting a latency of over 66000 samples using the highest quality setting.

Matt

[NoWo]

Hi Matthew,

thanks for trying to help.

Unfortunately the Fireface 400 driver does not allow higher values than 1048 samples, God knows why.

NoWo

[cass anawaty]

NoWo wrote on Sun, 29 June 2008 18:10

Hi Matthew, thanks for trying to help. Unfortunately the Fireface 400 driver does not allow higher values than 1048 samples, God knows why. NoWo

Wow.  That's good to know--I occasionally have people ask me about certain interfaces, and I never thought to take that into consideration.

[NoWo]

Cass Anawaty wrote on Sun, 29 June 2008 19:16

NoWo wrote on Sun, 29 June 2008 18:10 Hi Matthew, thanks for trying to help. Unfortunately the Fireface 400 driver does not allow higher values than 1048 samples, God knows why. NoWo Wow.  That's good to know--I occasionally have people ask me about certain interfaces, and I never thought to take that into consideration.

I wished someone had told me earlier -:(

RME seems to be so fixed to have interfaces with the smallest sample values that they completely forgot that there will always be situations where you need the high values.

I just downloaded the latest driver but it remains the same. Even my cheapest soundcard ( and I have a lot ) can at least do 2048, but not the RME.

NoWo

[Tomas Danko]

In all honesty, could you still call a 66000 sample latency "real-time"?

Barely. Maybe one could use a lower quality setting to tweak the sound and then get an offline rendering to achieve the highest quality without bothering with buffers.

FWIW I can set larger buffer sizes than 1024 on my RME Hammerfall DSP PCI so perhaps it's just the Firewire units that can't do it.

[cass anawaty]

Tomas Danko wrote on Mon, 30 June 2008 10:43

Barely. Maybe one could use a lower quality setting to tweak the sound and then get an offline rendering to achieve the highest quality without bothering with buffers.

I just glanced through the manual once, but I'm pretty sure that's how it works.

[Tomas Danko]

Cass Anawaty wrote on Mon, 30 June 2008 16:06

Tomas Danko wrote on Mon, 30 June 2008 10:43 Barely. Maybe one could use a lower quality setting to tweak the sound and then get an offline rendering to achieve the highest quality without bothering with buffers. I just glanced through the manual once, but I'm pretty sure that's how it works.

But then why would a buffer size larger than 1024 bytes be needed then? Offline is offline.

[cass anawaty]

Tomas Danko wrote on Mon, 30 June 2008 16:50

Cass Anawaty wrote on Mon, 30 June 2008 16:06 Tomas Danko wrote on Mon, 30 June 2008 10:43 Barely. Maybe one could use a lower quality setting to tweak the sound and then get an offline rendering to achieve the highest quality without bothering with buffers. I just glanced through the manual once, but I'm pretty sure that's how it works. But then why would a buffer size larger than 1024 bytes be needed then? Offline is offline.

I meant the "lower quality" setting works that way (I think it's called "auto").  You monitor the lower setting, but it renders at the higher one.

If I'm going to bounce at the higher res, I want to monitor the same--but I could see where it might help someone w/ a sluggish machine.

[NoWo]

Hi,

as I already said in my case also the "low quality trick" does not work, maybe I need a quadcore.
For now I save myself with tweaking the "normal" Sonoris eq and then put the values into the LinEq.
After some renderings I have to say the results of the LinEq are realy better than the normal eq, they seem to be clearer, have more depth and more punch.

NoWo

[TotalSonic]

NoWo wrote on Mon, 30 June 2008 16:46

Hi, as I already said in my case also the "low quality trick" does not work, maybe I need a quadcore.

fwiw - one of my DAW's is an older 2.8GHz P4 and it runs the Sonoris eq's in "HQ" (high quality upsampling mode) without any problem.

I think most likely you just need to be beef up your cpu a little to run the Sonoris eq's without any problems.

Best regards,
Steve Berson

[Andrew Hamilton]

Fwiw, my P4 is only 2.0 GHz, but when I activate the 2048 sample buffer in PTLE7 (Digi002), I can run one instance of Sonoris in HQ mode and one instance of PSP Xeon in the (aux) channel fader.  There is a delay (hysteresis) to the change in sound wrt changes in settings.  If one moves the mouse very slowly, one can achieve results through successive approximations (left and right).   How slow to respond to settings adjustments is the Weiss hardware LinPhase EQ?


Here's a real hoot:  G4 as a host for a vintage quantum DAW?  Last night I was doing a null test in DP3 (of all versions) between the DP3 "edl" and a "merged soundbites" version (contiguous file) of the same edited program as in the edl.  It nulled perfectly, but when I added a master fader and invoked SpectraFoo (version 3 radf17) as a plug in, in order to verify the amount of nullification, the synchronization of all tracks got so out of whack that it didn't null anymore!  So, if your DAW doesn't have auto latency-management (which I guess DP3 didn't (read: doesn't) have), a measurement of the goings on will affect the goings on... Just as in a quantum experiment.  


Andrew

[TotalSonic]

Andrew Hamilton wrote on Mon, 30 June 2008 21:26

Fwiw, my P4 is only 2.0 GHz, but when I activate the 2048 sample buffer in PTLE7 (Digi002), I can run one instance of Sonoris in HQ mode and one instance of PSP Xeon in the (aux) channel fader.  There is a delay (hysteresis) to the change in sound wrt changes in settings.  If one moves the mouse very slowly, one can achieve results through successive approximations (left and right).

The lack of efficiency could possibly be host or RTAS wrapper related as I don't have any real problems with the Sonoris eq's efficiency using the native SAWStudio version.   Also I think it helps efficiency greatly if the cpu is capable of handling SSE2 optimizations - not sure if your older P4 handles the SSE2 instruction set.

Best regards,
Steve Berson  

[bruno putzeys]

If I understand well, the high quality mode does not take more processor power, it just has more latency (because it allows more overlap for the filters to settle).

[Andrew Hamilton]

Maybe it's just extra latency for HQ-mode, but the single instance of Sonoris uses processing cycles like a stack of Waves IR's, if PT's CPU meter is any indicator.