low latency monitoring doesn't work on 002r latest drivers

[jntracks]

i'm doing a record, first full session with PT9. just got to vocals and found that low latency monitoring isn't working. checking it or unchecking it has no effect.

singer is complaining, it's very embarrassing... please help
on 002r, latest driver. 9.0.2

SiSoftware Sandra

Computer
Model : System manufacturer P5E WS Pro
Workgroup : WORKGROUP
Host Name : workstation-PC
User : workstation

Processor
Model : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
Speed : 2.4GHz
Cores per Processor : 4 Unit(s)
Threads per Core : 1 Unit(s)
Type : Quad-Core
Integrated Data Cache : 4x 32kB, Synchronous, Write-Thru, 8-way, 64 byte line size
L2 On-board Cache : 2x 4MB, ECC, Synchronous, ATC, 16-way, 64 byte line size, 2 threads sharing

Computer
Mainboard : Asus P5E WS Pro
BIOS : AMI 0803 06/16/2008
Bus(es) : ISA X-Bus PCI PCIe IMB USB FireWire/1394 i2c/SMBus
Multi-Processor (MP) Support : No
Multi-Processor Advanced PIC (APIC) : Yes
Total Memory : 4GB DIMM DDR2

Chipset
Model : ASUS X38 Processor to I/O Controller
Front Side Bus Speed : 4x 267MHz (1GHz)
Total Memory : 4GB DIMM DDR2
Channels : 2
Memory Bus Speed : 2x 534MHz (1GHz)

Memory Module(s)
Memory Module : G.Skill F2-8500CL5-2GBPK 2GB DIMM DDR2 PC2-8500U DDR2-1066 (5-5-5-15 3-24-6-3)
Memory Module : G.Skill F2-8500CL5-2GBPK 2GB DIMM DDR2 PC2-8500U DDR2-1066 (5-5-5-15 3-24-6-3)

Video System
Video Adapter : NVIDIA GeForce 8600 GT (32 SM4.0 1.2GHz, 256MB DDR3 1.51GHz 128-bit, PCIe 1.00 x16)
Video Adapter : NVIDIA GeForce 8600 GT (32 SM4.0 1.2GHz, 256MB DDR3 1.51GHz 128-bit, PCIe 1.00 x16)

Graphics Processor
OpenCL Processor : NVIDIA GeForce 8600 GT (32SP 4C 1.3GHz, 256MB DDR3 1.51GHz 128-bit)
Compute Shader Processor : NVIDIA GeForce 8600 GT (32SP 1.2GHz, 256MB DDR3 1.51GHz 128-bit)
CUDA Processor : NVIDIA GeForce 8600 GT (32SP 4C 1.3GHz, 256MB DDR3 1.51GHz 128-bit)

Storage Devices
WDC WD800JB-00JJC0 80GB (ATA100, 8MB Cache) : 75GB (D:)
Seagate ST3250410AS (250GB, SATA300, 3.5", 16MB Cache) : 233GB (C:)
Hitachi HDS721010CLA332 (1TB, SATA300, 3.5", 7200rpm, 29MB Cache) : 932GB (E:)
_NEC CD-RW NR-9100A (ATA33, CD-RW, 2MB Cache) : N/A (F:)

Logical Storage Devices
project (D:) : 75GB (NTFS) @ WDC WD800JB-00JJC0 80GB (ATA100, 8MB Cache)
Hard Disk (C:) : 233GB (NTFS) @ Seagate ST3250410AS (250GB, SATA300, 3.5", 16MB Cache)
jntracks (E:) : 932GB (NTFS) @ Hitachi HDS721010CLA332 (1TB, SATA300, 3.5", 7200rpm, 29MB Cache)
3.5" 1.44MB (A:) : N/A
Optical Drive (F:) : N/A @ _NEC CD-RW NR-9100A (ATA33, CD-RW, 2MB Cache)

Peripherals
LPC Hub Controller 1 : ASUS (ICH9) LPC Interface Controller
LPC Legacy Controller 1 : Winbond W83627DHG
Audio Device : ASUS (ICH9) HD Audio Controller
Audio Codec : Realtek Semiconductor 888 High Definition Audio
Serial Port(s) : 1
Disk Controller : ASUS (ICH9) 4 port Serial ATA Storage Controller 1
Disk Controller : ASUS (ICH9) 2 port Serial ATA Storage Controller 2
Disk Controller : ASUS 6145 SATA2 Controller
USB Controller 1 : ASUS (ICH9) USB Universal Host Controller
USB Controller 2 : ASUS (ICH9) USB Universal Host Controller
USB Controller 3 : ASUS (ICH9) USB Universal Host Controller
USB Controller 4 : ASUS (ICH9) USB2 Enhanced Host Controller
USB Controller 5 : ASUS (ICH9) USB Universal Host Controller
USB Controller 6 : ASUS (ICH9) USB Universal Host Controller
USB Controller 7 : ASUS (ICH9) USB Universal Host Controller
USB Controller 8 : ASUS (ICH9) USB2 Enhanced Host Controller
FireWire/1394 Controller 1 : ADS Technologies TSB43AB23 1394a-2000 OHCI PHY/link-layer Controller
FireWire/1394 Controller 2 : ASUS VIA VT6308 1394 OHCI Controller
SMBus/i2c Controller 1 : Intel ICH SMBus

Printers and Faxes
Printer : Microsoft XPS Document Writer (600x600, Colour)
Fax : Microsoft Shared Fax Driver (200x200)

Network Services
Network Adapter : Marvell Yukon 88E8056 PCI-E Gigabit Ethernet Controller #2 (Ethernet, 100Mbps)

Power Management
Mains (AC) Line Status : On-Line

Operating System
Windows System : Microsoft Windows 7 Home 6.01.7600
Platform Compliance : x64

Windows Experience Index
Current System : 5.9

Performance Enhancing Tips
Tip 2546 : Large memory modules should be ECC/Parity.
Tip 2 : Double-click tip or press Enter while a tip is selected for more information about the tip.

[daeron80]

It's a known issue. The workaround is to strip your mix down to the point that you can set the hardware buffer to 64. Print your VIs or whatever. There aren't more than a handful of singers in the whole world who will complain about 5 ms delay. Especially since you'll be able to give them a bit of verb in the cans, unlike with LLM.

[jntracks]

i ran last night's session with 64 samples. that's when he complained about the "metallic" sound...
at least it's not impacting timing with the music, i guess.

but 64 samples at 48k is detectable to me and, unfortunately, to him also.

waiting patiently for a bug fix...

[mykhal c]

for my own education, how do you cats come up with the delay for 64 samples? if i take 24/48 session...hilite 64 samples in the timeline...this amounts to 1.25ms of delay accordin' to the timeline from sample 1 to sample 64. curious...

[jntracks]

it actually works out to 1.33 and 1/3 ms. 64/48000 will give it to you in seconds.

it's a funny tin can kind of sound you hear in the headphones when singing into the mike. i don't know, like it's phasing a bit or something. with the voice in your head maybe?

i'm so tired right now i can hardly think straight. it doesn't make sense that you'd be able to hear that short a delay unless it's causing some kind of comb filtering effect with, i guess, the sound of your own voice in your head.

but to make sure it was the monitoring latency i put the h/w buffer up to 1024 and listened to the much bigger difference, switching LLM on or off had no effect.

....now wait this doesn't make sense because at 1024 the time delay from playing a note to hearing it sounds more like a quarter of a second... but from my math that should be like .00213 seconds at 48k.
so i'm doing this wrong. i must be.
time to sleep i guess.

[BobbyDazzler]

Sounds like you may be monitoring input and return?

[mykhal c]

Originally Posted by jntracks

it actually works out to 1.33 and 1/3 ms. 64/48000 will give it to you in seconds.

thanks for that.

i'm so tired right now i can hardly think straight. it doesn't make sense that you'd be able to hear that short a delay....

exactly ....point i was gettin' at

[daeron80]

The confusion here is between hardware buffer settings and latency times, both of which are measured in samples but which are related by some strange unknown forumula. A HWB of 64 samples produces a roundtrip recording latency of 311 samples on an 002R. My page with details. If you use a different interface, YMMV.

[mykhal c]

Originally Posted by daeron80

The confusion here is between hardware buffer settings and latency times, both of which are measured in samples but which are related by some strange unknown forumula. A HWB of 64 samples produces a roundtrip recording latency of 311 samples on an 002R. My page with details. If you use a different interface, YMMV.

too 'challenged' to follow you on this one but appreciate the link. i guess i'm missin' the reason why roundtrip is involved here. i've recorded myself against a click many times and i know for a fact there were times when my note placement was much less than 301 (64@24/48) samples offset delay to the printed click....not all the time tho yep...confused on the roundtrip application. but no need for futher explanation...i'll re-read your blog post

[jntracks]

Originally Posted by daeron80

The confusion here is between hardware buffer settings and latency times, both of which are measured in samples but which are related by some strange unknown forumula. A HWB of 64 samples produces a roundtrip recording latency of 311 samples on an 002R. My page with details. If you use a different interface, YMMV.

ok, this post (and some sleep) makes me feel better. so i'm not crazy after all.

and i'm not using a hardware mixer so it's not monitoring in two places. (that would also have been noticeable at higher buffer sizes)

i think the "metallic" sound has got to be some kind of slight comb filtering with the sound in your own head, right? cause we don't hear it in the control room, but we both notice it when we put the headphones on and talk into the mike. AND it wasn't a problem for any other instruments, as this recording has all been PT9. so all signs pointing to phase with the voice on the other side of the ear drum.