How to connect external hardware FX?

astroman · Post by **astroman** » Thu Aug 04, 2005 12:42 am

On 2005-08-03 23:48, funkysam wrote:
In my opinion, one of scope's main selling point has allways been the flexible no latency routing with external gear...

of course we all agree on this, even if the 2nd part is wrong

Of course there is latency with external gear and it's easily enough to spoil any mix or drive you mad if you're unaware of it's existence.

The point is not if the thing is faster than hell or if it has any influence on someone's live performance.
A (processed) signal that comes in a few samples late or earlier regarding it's source will heavily interfere if both signals are mixed together.

As Johann mentioned you cannot notice that in a delay fx (say Husker's Space Echo) because it's late anyway, but an eq or a compressor would be almost out of control as it's frequency depending. It can even beef up certain sounds significantly...
I bet my *rse there are thousands of mixes in which the operator was totally unaware of this relation between parallel signals.
It's possibly even the main source if someone complains about not being able to get that transaprent and punchy sound...

I've had almost the same setup as Husker for some time and it took me several months before I noticed something didn't match.
By a mixture of unawareness and lazyness a (parallel) part of the signal (which should have been muted) made it's way to the monitors.
In such a situation your audible result is completely different from the recorded signal. If you tweak it as good as you can get it, burn a CD and proudly present it next day - you'll definetely be screwed

Of course it's not a problem at all if you monitor and record at the appropriate position of the processing chain - but first you have to be aware of it and todays typical setups are often pretty complex.

that's why such a stupid rule of thumb is really great - always keep your processing chain serial and double check the signal path where creativity drives you to leave the rule

cheers, Tom

funkysam · Post by **funkysam** » Thu Aug 04, 2005 2:57 am

On 2005-08-04 01:42, astroman wrote:

of course we all agree on this, even if the 2nd part is wrong

Well, in fact, I know all this, but when I wrote this, I had delay and reverb effects in mind, where we don't have to worry about those samples latencies.

But these are not scope related problems only, because each digital driven studio with outboard gear has to deal with them, especially when mixing.

at0m · Post by **at0m** » Thu Aug 04, 2005 3:28 am

I make use of the Source/Dest/External Effects modules. They allow me to save the connection within.

When I change external connections, I just resave the modules and don't have to adjust my projects for it - the modules will be loaded corrected in the project when they're saved with the updated connections

Post by **garyb** » Thu Aug 04, 2005 8:45 am

all gear, analog or digital, has SOME latency...

at0m · Post by **at0m** » Thu Aug 04, 2005 1:06 pm

http://www.planetz.com/Pulsar/PerformanceLatency.html

husker · Post by **husker** » Thu Aug 04, 2005 2:24 pm

It's a very good point that Tom makes...of course there is *some* latency, which will really screw things up if you do any parallel mixing of dry/wet signals outside of the effect. But this is the same of ANY mixing situation (unless its a purely analog effect).

I can hear very nasty phase cancellation effects through my VM88 monitoring setup if I monitor dry/wet signals together, so I am always acutely aware of what I record.

The only latency I really care about is any delay between my input actions VS what I hear. Scope and/or hardware based processing is always going to be faster than a ASIO/VST based signal chain. The fact that scope is totally flexible AND very fast makes it unmatchable.

Cheers, Wayne.

astroman · Post by **astroman** » Thu Aug 04, 2005 2:49 pm

On 2005-08-04 15:24, husker wrote:
...But this is the same of ANY mixing situation (unless its a purely analog effect)...

exactly, and it's even valid in pure analog situations - at least if stages are coupled by caps (a widespread method) there should be a measurable (though very small) delay. Dunno about coils in this context, they might be faster...

cheers, Tom

husker · Post by **husker** » Thu Aug 04, 2005 3:23 pm

Yikes! If I started worrying about delay through capacitors i'd never get any music done

darkrezin · Post by **darkrezin** » Thu Aug 04, 2005 3:34 pm

blazesboylan · Post by **blazesboylan** » Fri Aug 05, 2005 8:16 pm

Sorry if it sounded like I'm down on SCOPE! I am definitely a BIG advocate of CWA products.

But yeah I agree with everything above.

I've looked at the difference between outboard (analog) gear latencies. The differences between about 5 or 6 different compressors (some with transformers, some all IC based, etc) was about max. 8 samples at 44.1 kHz, IIRC.

The latency doing the old "in-out, in-out" D-A and A-D is 2 orders of magnitude greater than the discrepency between the "fastest" and "slowest" of the compressors I tried.

Interestingly, too, 2 of the same model of compressors from the same manufacturer with the same length of cables (also from the same manufacturer, same thickness, connectors, etc too) resulted in latencies that were 2-4 samples out.

Compared to 2-4 ms, though, 2-4 samples is unnoticeable.

Keeping everything serial is nice in theory, but for mixing it's also nice to go parallel!

Cheers,

Johann

at0m · Post by **at0m** » Sat Aug 06, 2005 2:55 am

At 44kHz, 4 Samples, that's the resolution of a 11kHz signal - and everyone can hear that tone. I wonder, what samples would represent the 11kHz. Would a triangle or sine be a loop of -1,0,+1,0 values, or rather a square waveform of -0.5,-0.5,+0,5,+0,5? Anything near or above that would be pulse width modulated. Both would sound much different, in amplitude and in harmonics. Mixing in a 4 samples delayed signal with a dry version can add an inverse to a 5.5kHz waveform (consisting of 8 samples for the loop) and mute it. Of course one hears what's coming through or recorded, and if one's not happy with it, it won't get recorded.
But you understand now that the side effects of a couple of samples delay, if not compensated for, can get quite dramatic for mixing wet/dry versions of hihats, pads etc that live in those higher regions.

astroman · Post by **astroman** » Sat Aug 06, 2005 9:13 am

very right, the understanding is the main point.
and of course it's perfectly ok to record a couple of signals completely out of 'timing' if the resulting sound is simply cool

I consider my 'personal' example fairly representative, as I monitored the main mix properly, but simply overlooked a signal from the 'patchbay' that should have been muted - things like that just happen...

As Atomic describes, this behaviour is frequency dependent - 4 samples are easily detected on a hi-hat sound.
In my case it was about 20 samples (0.5 ms) on a (typical mid) pad sound, which faked a filter when moving the channel fader up and down.
I remember it that well because I searched like an idiot for that filter in the routing... must be somewhere...

a doubled voice or guitar sound would have been screwed equally, as would any layered sound, misaligned attack phases etc - it's far from exotic

regarding a (pure) 11 khz source: it doesn't matter if it's a perfect sine, a triangle or some pulses. Regular folks (without esotheric talents) can't distinguish it.
Afaik at least one harmonic is necessary to determine a sound's character (or origin), and at 11 khz the first harmonic is at 22khz.

if you try it at home you may be able to distinguish it, tho - but probaly not due to you listening capabilities but due to errors the speaker generates.
Those are likely to be different for these 3 sources

cheers, tom

<font size=-1>[ This Message was edited by: astroman on 2005-08-06 10:14 ]</font>