If you monitor your inputs analog, then there is no concern at all.  Any DAW worth its weight will re-align your performance based off the Driver's reported latencies (will offset the take accordingly and automatically).  You could have the latency set to 1024 samples and still get it to line up perfectly w/o lifting a finger!  Should be pretty close to "Sample accurate" which means the feel and timing will be completely un-altered compared to what you hear when you are performing the take against the backing track.  Them DAW's are pretty smart :-P

FWIW - I prefer to monitor "wet" w/effects and compression through the DAW with around 5ms RTL (or: 5' @ the speed of sound), and anyone I've tracked (even myself - a drummer of almost 30 years now) can never detect any latency when monitoring themselves.  Even vocalists that can have the "head voice vs headphones" phase issues seem perfectly happy with this method (literally not a single complaint over here!).  RME's ultra-low latency drivers and a fast PC make this possible w/o breaking a sweat.  Either method works fine (monitor yourself analog "dry" w/o any latency, or "wet" through the DAW with around 5ms RTL), but I always monitor wet for the fun factor (sounds so good with a full drum mix going in your cans as you lay them down!).  The total recall factor this affords (all mixing and monitoring is done and saved in the DAW itself) is also a nice convenience IMNSHO.

The one "gotcha" can be if you use external AD/DA converters that differ from what the RME's Driver reports.  Practically all RME boxes report the latency of their ANALOG I/O - so if you use an external AD/DA that has more or less conversion latency than the RME converters, these reported offsets can become misaligned (can cause phasing and comb filtering when combining similar signals - this won't really affect the "feel" of a take as this is generally less than a ms or so).  There are many threads on the issues this can cause, but in the end it is a relatively small concern in most scenarios.

You got it!  I do lean towards PCIe (actually, exclusively on PCI/PCIe), but I'm kind of old-school like that sometimes.

Although the range of 5ms still falls well within the Hass effect and even within the player-to-player physical distance separation vs speed of sound, I can understand why you want to go with analog monitoring for such tracking tasks.  The good thing about DAW latency - if you are setup correctly - it's a consistent latency and can be "played into" much like playing with a band in a large room where distance creates the natural analog delay between musicians.  But I simply can't feel 5ms on my own drums - either that, or I just cope automatically at these low latencies.  Get me to 12-16ms RTL when tracking drums and I can't stand it, but can still "play into it" reluctantly :-)  Regardless, the tracks I play would still be lined up to the backing tracks I heard with the same relationship upon playback.  When it's consistent, it's easy (well, easier) to manipulate...

You're gonna love having an RME setup for sure, and I bet it will knock your socks off if you ever try its low latency capabilities with a proper "wet" monitoring setup

For me, I'm monitoring 24-40 inputs in realtime with all kinds of VST processing and crap going on - so using the DAW for my monitoring is the bees knees - hands down.

I do suppliment with tasty discrete analog preamps and compressors (mostly DIY stuff) - but not for latency concerns, but for the SOUND!  I still monitor all of the analog front-end goodness through the DAW.

An analog setup that would match my "ITB Monitoring Solution" would be well out of my budget (20K or likely well more), and would actually add compexity and TONS more wiring.  I'm just not able to hear/feel the ~5ms RTL in my system, so why would I bother going with an alanog monitoring setup in my case?  Total Recall of a session = Time Saver galore...

If you are doing a stereo pair on a Tabla or something, analog monitoring is much easier to handle - but I will always use ITB "WET" monitoring with awesome low-latency RME interfaces and a "proper" DAW session setup even when recording a single track.  Pure heaven!!! :-)  The talent is always impressed with the "full mix" type sound of their takes, and I have not recieved a single latency complaint!  The key is knowing your DAW software, and how to avoid additional latency from within the project itself (PDC, etc).  Of course, having a computer that will handle your requirements at these extremely low ASIO latencies might be something else (I'm running a DIY 4.3GHz i7-930 to cope on my end!).

Once you get the RME, give it a shot.  It's not like it is a permanent thing.  I think you will be surprised that ~5ms RTL really isn't detectable to most people on earth (there are bound to be a few exceptions - I have yet to have any in my studio ).  The Haas effect and the speed of sound tend to agree :-D

Hi Jorge7,
as far as I understood, you need as much as possible low latency of you "original" sound of you instrument that you what to dub to your existing tracks. I don't know Audition, so I don’t know if it supports ASIO DIRECT MONITORING (ADM). If so, you should be able to do the overdubbing with zero latency. ADM switch the input signal of FF UC right in the moment when you punch in to your record to your selected output. In my case with Samplitude it works very well.

so long
Holger

Hi Randyman, you are totally right that there is no possibility to use the internal DAW effects while working with ADM during the process of recording/dubbing on the recorded channels. But in the case how RME uses the ADM there is no latency as long as you use analog in- and outputs of the unit, here the ADM signal is just an impulse for totalmix . The signal that is monitored that way does not go thru the converters on its way to the monitor. You can see it if you watch Totalmix. In the moment when ADM get active the inputfader of the channels that you want to dub goes up. That means that for example if you just want to dub something from your !analog! input 1 of your FF, the fader 1 goes up. So the analog input 1 will be "switched" directly on the FF to the ! analog ! output you that you are monitoring and parallel goes the signal, via the converters, to the hard disk. The DAW, in my case Samplitude, makes sure that the relation in time of the newly recorded material to the already recorded stuff will be right.
   
Holger

Hi Vinark,
it is true. You could read about in the manual of the FF on page 78 "Zero Latency Monitoring" and you could hear it when you try it.
As I wrote, if you use, I my case Samplitude, I switch there to the mode that is called "Hardware Monitoring". Then I will be able to hear the signal I want to record without latency, but as said before no chance in that mode to have internal DAW effect on the recorded material during the process of recording.
So long
Holger

After a another look to some tech notes I must admit you are right, the signal goes thru the converters, but not thru to whole DAW. RME talks about 68 microseconds at 44.1 kHz that’s nearly zero and much better as the best latency you could ever reach if you go do software monitoring.

Holger

Can someone please enlighten me as to the difference and/or the advantage in the case Asio Direct Monitoring versus monitoring through the UC's Total Mix (hence not through the DAW)?

I'm using Sonar X1 and I think it does not support ADM.

What I could figure out from the manual, was that I can link the faders in my DAW to control those in Total Mix. Is that all?

Randyman, the delay compensation in the software should eliminate the input latency issue as you said, I just have to learn how to use it in Audition, although I did not see it mentioned in the Audition 3 manual. Thanks for that tip.

Regarding the detectability of a 5 ms delay, even non-super-humans like us can hear that. Do a little experiment. Play and record a few flams, from completely coincident in time to barely separable to clearly separable. Don't use a snare drum, use something with minimal sustain, like 2 hardwood drumsticks hitting a smooth heavy hardwood object like a floor or a heavy table, so you get real short click impulses. Use a mic with really good impulse response, like an Earthworks omni, and no EQ. Find a flam on the waveform that you can just barely distinguish as 2 hits when you play it back, measure the separation in time on the recorded waveform, and tell us how short a separation you can perceive as 2 separate sounds. You might even be in the range you previously considered super-human. The Haas effect tells us about inability to distinguish locations in space of 2 sound sources when the sounds are less than 20 or 30 ms apart in time, but does not tell us we can't distinguish as separate in time 2 sounds that occur closer in time than that. I don't think Haas was a drummer...

Hi Mathias.
First, my question about the input latency affecting the recording has been well answered by Randyman in his first post. There is no problem. The question came up because I could not find in the Adobe Audition 3 manual any discussion of delay compensation or realigning of the incoming signal to account for the conversion and software latency delays. It does seem to do this, but it is not well documented, or at least I could not find it by looking in the index or skimming different sections of the manual.

The discussion which arose after that, of whether the ~5ms RTL produced by "wet" monitoring through the interface and DAW is perceptible or not while recording, involves both musicianship and psychoacoustics. For multitrack recording with multiple musicians playing simultaneously and "wet" monitoring, in which you and Randyman have lots of experience and I have none yet, I take you both at your word that no one is disturbed by a 5 to 10 ms delay. For single track at a time overdubbing of very short sustain percussion parts, which is an important part of what I do, I will just have to try ultralow latency "wet" monitoring, as Randyman suggested, and see if I prefer it to direct analog dry monitoring.

@ffucowner, none of this thread has been talking about delays created by plug-ins. I may be using the wrong terminology, but as I described it in the post just prior to yours, I am talking about what Randyman and I described in our first posts. To describe it more precisely, what I mean is time realignment of the newly overdubbed track with the existing tracks to account for the latency created by the DA and AD conversions and the DAW software. It seems most DAW software does this automatically, although in Adobe Audition I don't know what they call it. I was not able to find it documented in the Audition manual.

Lots to respond to:

WRT Haas effect - Agreed it also has to do with spatial positioning and amplitude relationships, but it basically states that 2 similar sounds of similar amplitude spaced anywhere from 5-30ms (broad window there!) can be heard as "one sound" but will affect perceived origin or "width" of said sound.  I liken this to the sound of the acoustic drums in the air (also incurs a small time delay from drumhead to ear - say 3ms for 3' @ speed of sound) compared to the sound of the drum in the headphones.  I'm not likely to hear 2 distinct sounds until I get around 20ms RTL or so, and I should be able to play into that groove if I'm worth my weight ...  Believe me - I would not accept my system if I was always on the verge of "Flamming" in my heapdhones - I'd still be on Analog!!!

The "impulse response offset" of hitting a dead clave (or whatever) into an Earthworks mic and trying to get the smallest duration "Flam" I can detect just isn't something I'd ever have a desire to record as "Music" - so I don't really find that relevant for my workflow.  I base my statements on real-world application of Wet ITB monitoring, and how the overall feel holds up - not a hypothetical thought experiment that would never happen "naturally" in my studio.  No complaints from the talent - and I'm the biggest critic as the Drummer I record most often!  If I'm happy, then I know most others will be ecstatic with the setup!!!

RE RTL - I tested the MF-I's ASIO Analog-to-Analog RTL (@ 32-Samples) with the Centrance Latency Tester and got the 3.1x ms RTL.  My HDSPe MADI/SSL setup is about the same, but I need to manually specify the delay of the AD/DA offset as the MADI Driver doesn't know which converters I'm using (see point below about PDC vs Record Offset Placement).  I haven't done a true "Loopback" on a full session, but I can tell the latency is still very low, and I got to my ~5ms RTL figures by adding the few latent plug-ins I run plus a small bump.

AFAIK, Simply inserting a "0 Offset" plug-in doesn't add any delay to speak of.  It might add one sample, but most actually report "0".  Bussing might add another sample.  So even in a complex project, that *might* add up to 20-30 samples total?  I believe it will be far less.  But you are correct that I need to actually setup a test to confirm what I get "Through the DAW" in my larger projects.  I can definitely hear the distance-based-latency on room mics compared to the direct acoustic sound (distance comes into play, approx 15ms due to 15' distance plus my RTL in this case), but I can't hear any latency from direct sound to tight mics or overheads in the headphones.

To verify, I can record an analog loopback as a test, or I can actually record a common Mic to 2 different DAWs, plus record the Main DAW's Mix to the second DAW and get an RTL.  I'll have to mind my P's and Q's to make sure I'm not omitting any AD/DA conversion latencies or otherwise getting fooled by the record offset placement (using 2 separate DAW's should eliminate that variable).  No time this week, but maybe this weekend.

PDC - Refers to "Plugin Delay Compensation", and will increase latency of any live inputs to "line up" the plugin delays across tracks.  Even if you insert a latent plug-in on a track you aren't currently arming, it will STILL offset the armed track to keep everything lined up (at least in Cubase - Logic might be different as it does some wierd stuff to cope with low latency!).  So - don't use any latent plug-ins while tracking!  It will KILL the low RTL you are working so hard to maintain :-)

"Record Offset Placement" is what most DAW's do automatically based off the ASIO's reported AD/DA and ASIO offsets.  This is what allows your performance to line up EXACTLY with the backing track as you heard it during the performance.  This should not affect RTL when monitoring (AFAIK), but will ensure what gets played back is the same thing you heard during the performance WRT relative placement of the backing tracks against your performance.

All I can tell ya is it works every time in my setup.  I sold my old mixer, and rely strictly on the DAW for all of my monitoring/bussing/DSP needs, and supplement with a "very analog" front-end for the interesting/tasty flavors Digital can't produce on its own.  I don't even think I'd want to get a nice console at this point! :eek: