The PARIS Forums


Home » The PARIS Forums » PARIS: Main » Check my math?
- Check my math? [message #105053] Sun, 14 March 2010 21:10 Go to next message
kerryg is currently offline  kerryg   CANADA
Messages: 1529
Registered: February 2009
Senior Member
Administrator
Hi folks - latency's a hot issue for many PARISians, so I put a little latency explainer up on the Wiki. If any folks with relevant expertise want to run their eye over it and make sure I'm not totally out to lunch on what I've written (particularly the math) I'd be grateful for any corrections.

"... being bitter is like swallowing poison and waiting for the other guy to die..." - anon

[Updated on: Sun, 14 March 2010 21:11]

Report message to a moderator

- Re: Check my math? [message #105055 is a reply to message #105053] Mon, 15 March 2010 06:52 Go to previous messageGo to next message
drfrankencopter is currently offline  drfrankencopter   CANADA
Messages: 137
Registered: July 2009
Senior Member
Well, personally I think it makes more sense to describe latency in terms of time, and distance (speed of sound), than MIDI ticks, but I guess some folks would appreciate the reference in a musical timing sense.

By my math, here's a sample delay chart (for 44.1 kHz). Remember that sound travels about 1 foot per millisecond, which can also give you an equivalent sonic reference point (e.g. would you worry about the timing of a track that you mic'd an extra 2 feet back from the performer?). Anyways, here's the chart (please note that these values rounded off to 2 decimal places):

#of samples | delay
1 | 22 micro seconds
64 | 1.45 ms
128 | 2.9 ms
256 | 5.8 ms
512 | 11.61 ms
1024 | 23.22 ms
2048 | 46.44 ms
4096 | 92.88 ms

Cheers

Kris
- Re: Check my math? [message #105056 is a reply to message #105055] Mon, 15 March 2010 09:31 Go to previous messageGo to next message
thesandbox1 is currently offline  thesandbox1   GERMANY
Messages: 167
Registered: July 2009
Location: Carmel, IN
Senior Member
http://www.svn.net/artguy/paris.notes/2-6-03
"That's as good as it gets in my experience. Analog input to channel with compressor and EQ to
stereo bus to monitor out in 1.5ms at 44.1KHz is incredible. And I do think it matters. All
over the internet, I hear Native DAW owners yakking about "8ms latency is just the same as
being 8 feet from the speakers", etc. No, it's not. Not if you're wearing headphones as in the
majority of recording, certainly singing, not to mention drumming. Not that timing really
matters for drumming or anything.

I talk to pro drummers who absolutely DREAD going to project studios with "low latency" Native
rigs. No fun at all. Of course, the together types will monitor through a digital console
(still with more latency than Paris) and resolve the issue that way. But how much $$$ have you
saved and complexity have you gained by having to own a console just to monitor properly for
ODs?

The comb filter effect introduced into headphones by larger latencies is a big factor in the
way a singer feels about the sound of their own voice in the cans, IMO. Higher latencies create
enough lag between bone conduction and what's hitting the ears from the cue feed to create
issues for many singers, including pitch. Many variables here, but latency does not help
anything."



drfrankencopter wrote on Mon, 15 March 2010 06:52


By my math, here's a sample delay chart (for 44.1 kHz). Remember that sound travels about 1 foot per millisecond, which can also give you an equivalent sonic reference point (e.g. would you worry about the timing of a track that you mic'd an extra 2 feet back from the performer?). Anyways, here's the chart (please note that these values rounded off to 2 decimal places):





- Re: Check my math? [message #105057 is a reply to message #105056] Mon, 15 March 2010 15:14 Go to previous messageGo to next message
kerryg is currently offline  kerryg   CANADA
Messages: 1529
Registered: February 2009
Senior Member
Administrator
Latency is such a hot-button issue - how much latency is "acceptable" is going to vary from player to player, and often style to style. Small amounts bug me during tracking.

I want to get "hard numbers" for the rountrip latency of ProTools HD in there - anyone got exact data?


"... being bitter is like swallowing poison and waiting for the other guy to die..." - anon

[Updated on: Mon, 15 March 2010 15:15]

Report message to a moderator

- Re: Check my math? [message #105058 is a reply to message #105053] Tue, 16 March 2010 07:53 Go to previous messageGo to next message
Ted Gerber is currently offline  Ted Gerber   CANADA
Messages: 705
Registered: January 2009
Senior Member
Hi Kerry -

Thanks for the great work as per usual. Two things:

1. I believe the UAD latency is specific to the host computer, and the new (vs 5 & up) UAD control panel displays that particular system's latency.

2. Q: I'm planning on going the Vertex route, but how is audio latency across submixes accounted for? That is plugins used on submix 1 resulting in a delay of the rest of the audio on that submix, but what about the audio on subix 2?
Or is Senderella used for this?

Ted

- Re: Check my math? [message #105059 is a reply to message #105058] Tue, 16 March 2010 10:08 Go to previous messageGo to next message
dnafe is currently offline  dnafe   CANADA
Messages: 390
Registered: February 2009
Senior Member
Hey Ted

I used faderworks on my last project across two cards and subgroups and I didn't notice any comb filtering or anything odd sounding for that matter.

Now if Senderella can cross submixes and you have a common FX across several tracks split between two submixes I would think there is a real possibility of hearing some odd latency generated artifacts.

- Re: Check my math? [message #105060 is a reply to message #105059] Tue, 16 March 2010 11:12 Go to previous messageGo to next message
Ted Gerber is currently offline  Ted Gerber   CANADA
Messages: 705
Registered: January 2009
Senior Member
thanks Don
- Re: Check my math? [message #105061 is a reply to message #105059] Tue, 16 March 2010 14:04 Go to previous messageGo to next message
drfrankencopter is currently offline  drfrankencopter   CANADA
Messages: 137
Registered: July 2009
Senior Member
dnafe wrote on Tue, 16 March 2010 10:08


Now if Senderella can cross submixes and you have a common FX across several tracks split between two submixes I would think there is a real possibility of hearing some odd latency generated artifacts.




This will require further testing, because it depends on where the source of the latency across submixes lies. If the latency originates from the way the editor window feeds tracks to the mixer, then there WILL be potential for latency artifacts on the Senderella sends, but if the latency is generated in the EDS mixer itself, then the Senderella sends will not be affected as they occur in the native plug-in chain before the EDS mixer.

I hope that makes sense.

It's something that's easy enough to check by a null test across 2 submixes.

Cheers

Kris

PS: For most aux type effects this latency would be no big deal...but for parallel compression it could be a deal breaker!


Yet another PS: I've managed to get the Senderella source code to compile under the VST 2.3 SDK. Unfortunately the source that I have is not the same as the most recent version, but the core components are all there. I'm now going through the process of documenting the code, and hope to begin modifying it to better suit Paris this week(March break).

[Updated on: Tue, 16 March 2010 14:15]

Report message to a moderator

- Re: Check my math? [message #105062 is a reply to message #105061] Tue, 16 March 2010 14:51 Go to previous message
dnafe is currently offline  dnafe   CANADA
Messages: 390
Registered: February 2009
Senior Member
drfrankencopter wrote on Tue, 16 March 2010 17:04
dnafe wrote on Tue, 16 March 2010 10:08


Now if Senderella can cross submixes and you have a common FX across several tracks split between two submixes I would think there is a real possibility of hearing some odd latency generated artifacts.




This will require further testing, because it depends on where the source of the latency across submixes lies. If the latency originates from the way the editor window feeds tracks to the mixer, then there WILL be potential for latency artifacts on the Senderella sends, but if the latency is generated in the EDS mixer itself, then the Senderella sends will not be affected as they occur in the native plug-in chain before the EDS mixer.

I hope that makes sense.




Makes perfect sense
Previous Topic: Interconnections for 2 EDS cards
Next Topic: Nice macro keypad on ebay
Goto Forum:
  


Current Time: Sat Dec 21 04:05:35 PST 2024

Total time taken to generate the page: 0.00856 seconds