Latency Free Tracking with a DAW?

[Jax]

There was a thread somewhere that explained how to track without latency when using a DAW. I remember it involves an analog console.

Can somebody refresh my memory?

[Geert van den Berg]

You've got a couple of options. The most simple solution nowadays is to buy a soundcard which supports direkt monitoring of the inputs, and the more sophisticated ones also have a hardware mixer controlled by some software to do more complex things like seperate cue-mixes. With these cards you don't need an analog board at all. RME, Lynx and Motu are companies that make cards like this. Older cards can also have the direkt option, but these output the recorded signal at the same output number as the input at which it is recorded, like input 4 goes also direkt to output 4 when recording. When using a card like this you will still need an analog mixer. But the computer functions just like a hardware hd-recorder.

Some software, for example Digital Performer, also has the option that you can disable to listen to tracks which are being recorded. This can used if you use it with a soundcard without direkt monitoring and a console. Before the punch-in you will hear what's recorded+what comes from the console if the performer plays along and when punching in you will only hear the sound from the console. This way you only have to hook up all outputs of a console to the inputs of your soundcard for recording and from the soundcard you only have to hook 2 outputs to the console to listen back what is recorded. Offcourse this is less flexible when making cue-mixes, but it's still possible with some workarounds.

[Jax]

Oops, should have mentioned my rig:

PT Mix - ADAT Bridge - Mytek 8x96's - Ghost - outbd pres etc

[Geert van den Berg]

Haha, yeah would have saved me some typing!

But with a rig like yours, you can already monitor with near zero latency. What is it that you are trying to establish or which you don't understand?

[Jax]

Mix has 5ms on the card. Some say this can affect performance during overdubs.

[Geert van den Berg]

I always thought that the smallest amount of latency possible in a TDM system is about 3ms (1.5ms for the A/D and another 1.5ms for the D/A, I don't know if the mixer also adds extra latency could be 2ms), but you are probably right.

I hope others chime in as well. Personally I think that if you don't use any heavy plugins during tracking that those 3 or 5 ms don't matter. If it would be 10 ms then it would start to be a little different but even then I think it wouldn't hurt the performance. I feel that the performer should not feel there's a delay, but when playing you are also focused on the backing track which is in time, so if you don't hear that your sound is slightly late it works.

You could use your console but that would be a hassle in my opinion. There are several ways you can hook it up, that's more of your own preference. The main problem I see, that when using overdubs you want to let the performer hear his previous performance and when recording you want mute the record track, you can do this but you will need to use 2 tracks. For every track you record. Not such a problem with one mic, but less fun when using multiple mics, allthough grouping in PT makes that a little easier.

[BrianT]

If you want minimum PT Mix monitoring latency, then don't insert plugs on record channels and do not add a master fader.

As straight in/out as you can get and you should be around 2.5-3ms, analog in to analog out. That's about as good as it gets for DAWs, except Paris (sorry to bring it up), which is about 1.4ms analog in-analog out, through an EQ, comp, Master fader and monitor section.

Little known factoid:

PT Mix has better analog in-analog out latency than HD. By about 2ms, at 44.1K, IIRC.

[Ruudman]

Originally posted by BrianT
Little known factoid:

PT Mix has better analog in-analog out latency than HD. By about 2ms, at 44.1K, IIRC.

Ahh, Mix strikes back

ruudman

[C.Lambrechts]

Originally posted by BrianT
PT Mix has better analog in-analog out latency than HD. By about 2ms, at 44.1K, IIRC.

2ms seems an awfull lot Brian. Actually had to look it up :

888/24 : 81 samples delay on the analog I/O

192 : 103 samples

both at 44.1 ... so say an analog insert going DA and back inside AD would give you 1ms more latency.

Anyhow, small price to pay for way better convertors imho.

[Knastratt]

2ms won't do. I had a jazz singer in and he hated the "chorusing" (blend of head tone and delayed sound). I got an outboard mixer for monitoring. And everybody's happy.

I've learned to live with latency, but from this latency-free system - creativity has emerged!

[deft_bonz]

i let the singer hear his/her direct signal from the preamp plus the song over a mixer, where i can also give some reverb. i also use pt mixplus system with a 882/20 convertor. you have to mute the record track. if she wants to hear prior vocal tracks, then you have to work with several vocal tracks, but always mute the recording channel.

don't forget to compensate the delay on the vocal tracks after recording.

regards deft:::

[zimv20]

Originally posted by Knastratt
I got an outboard mixer for monitoring. And everybody's happy.

how do you split the signal from the mic pre? or are you using the board pre? or are you running the pre into the board and taking the direct out?

[BrianT]

Originally posted by C.Lambrechts
2ms seems an awfull lot Brian. Actually had to look it up :

888/24 : 81 samples delay on the analog I/O

192 : 103 samples

both at 44.1 ... so say an analog insert going DA and back inside AD would give you 1ms more latency.

Anyhow, small price to pay for way better convertors imho.

OK, 1ms. My bad. And to be fully fair, in that case the 192 would have less analog latency at 96K (which the 888 won't do) than the 888/24 at has at 44.1K.

[RKrizman]

Originally posted by BrianT
[b]If you want minimum PT Mix monitoring latency, then don't insert plugs on record channels and do not add a master fader.

Are you sure a master fader adds latency?

-R

[MMazurek]

I can see why a singer could maybe hear it, but I've never had a problem with anyone actually being bothered by it.

Racently, I had a two bandmembers ask about it... "hey, won't we have latency in the monitor mix?"...
...to which I replied... "No. The 4-5 msec or so latency is about equal you sitting 4-5 ft back from a guitar amp instead of having your head an inch from the grill."

Isn't this more or less true?

[BrianT]

Yes, unless you are wearing headphones, in which case it's like everybody else is 1 inch from your amp grill, while you are 4 or 5 feet back. Trust me, a decent drummer can feel 5ms while playing if he's in a set of headphones, no problem.

More importantly, when singing, the comb filter effect created by the delay vs bone conduction and acoustic vocal bleed through the headphones gives some singers fits, affecting their ability to nail pitch. It seems to affect some greatly, some moderately, and some not at all. That's one reason we have big online disagreements about such things. With humans making music, one size does not fit all.

Try this as an experiment. If a singer is havng a tough time with pitch, flip the phase on their monitoring channel. It might make a difference to them. Theory demonstrated.

That is one of the underappreciated reasons that latency matters, even at relatively low values. I'm pretty sure that the "10ms is the same as being 10 feet from your amp" line was invented by a guy using all Native hardware, before Direct Monitoring, trying to cover his butt. I mean, if that statement's true, why not just go 25ms and pretend you're 25 feet from the amp. Shouldn't make much difference, if that analogy is true.

[BrianT]

Originally posted by RKrizman
Are you sure a master fader adds latency?

-R

To be honest, I have not measured it personally. But several truly gifted musicians who own PT have told me that they avoid a stereo master fader when overdubbing because theu can feel a slight difference. These guys are inhumanly tight players, so I believe them.

But no, I don't know this firsthand, as a scientifically measured fact. Yet.

[RKrizman]

Originally posted by MMazurek
I can see why a singer could maybe hear it, but I've never had a problem with anyone actually being bothered by it.

Racently, I had a two bandmembers ask about it... "hey, won't we have latency in the monitor mix?"...
...to which I replied... "No. The 4-5 msec or so latency is about equal you sitting 4-5 ft back from a guitar amp instead of having your head an inch from the grill."

Isn't this more or less true?

I think the biggest problem with short latency is that it affects the timbre of the voice in the singer's headphones because of the comb filtering from the time delay. Of course, there will already be comb filtering based on how far away the singer is from the mic. 6 inches is about a half a millisecond, right? In general, I think you're right and would say that latency is an integral part of reality itself, and will be unless we all figure out how to occupy the same space at the same time.

It's a question of tolerances, not absolutes.

-R

[RKrizman]

Originally posted by BrianT

That is one of the underappreciated reasons that latency matters, even at relatively low values. I'm pretty sure that the "10ms is the same as being 10 feet from your amp" line was invented by a guy using all Native hardware, before Direct Monitoring, trying to cover his butt. I mean, if that statement's true, why not just go 25ms and pretend you're 25 feet from the amp. Shouldn't make much difference, if that analogy is true. [/b]

Of course it would make a difference, in a live setting as well. Ever try to play string band music with players 25 feet from each others sound sources?

The thing the amp argument doesn't take into account is that it's not like being 25 feet from your amp. It's like being that plus being right at the amp at the same time, with all the ensuing timbral and rhythmic (or at least groove) confusion.

-R

[Jax]

Ok, speculation and discussion aside, anybody have a concrete answer to my question?

How is latency free tracking achieved using an analog console with a DAW?

[zimv20]

Originally posted by Jax
Ok, speculation and discussion aside, anybody have a concrete answer to my question?

How is latency free tracking achieved using an analog console with a DAW?

one monitors pre-DAW by somehow splitting off the incoming signal

[Jax]

Alright, that's a start.

Anything more in depth? Rome wasn't build on one sentence.

[zimv20]

couple ways i know of:
- make a y-cable coming out of the pre
- tap into a half-normalled patchbay
- run the pre into a board, take the direct out to the DAW
- use a pre w/ two line outs

i am interested in knowing the pros/cons of the first two, if anyone would like to chime in. i'm getting ready to half-normal some of my patchbay and am looking for info on any gotchas.

[studjo]

Use a board and make your cue sends with the aux sends from the recording channels - mind you: you won't do any overdubs with this setup (the muscicians can't hear what's already on tape).

good luck


Jo

[Exmun]

Depending on what pres you're using this shouldn't be a problem. If you're using mic pres with both an XLR and line out on trs, you just send the trs out to the hardware mixer and send the xlr out to PT. Turn the input volume off (out of the mix); send the PT out to the hardware mixer and monitor from the mixer rather than through PT.

[Auxillary]

I dont have much to add other than my experience as a session drummer and studies on the subject

I have done many gigs at places with the mix systems a few years ago as a "kid" (blah, was/is one of those prodigys i guess)... And I always used to complain that my tracks were off after I listen to them back. Whenever I would go into a studio going through a console it would always ALWAYS come out tighter. 2ms or not. This is where my hate of protools first evolved.

Also, engineers take note: Musicians dont use their ears when they play. It's 98% feel, and that is instant. A good musician will always feel the right rythm, right pitch and right timbre and use their ears as a verification device. This goes double for bass, percussion and vocal. When the verification device is impaired they have to:

a. take time to disable it (this is when you hear some guys say "I gotta change gears for a second")

b. Learn to expect it later

c. Some, perfectionists (egoists), cant deal with it. They need instant satisifaction that their performance is spot on. (That's me)

Keep this in mind while the discussion continues. The musical device is human, therefore it is instant. If reflection of the performance is not nearly instant, or at least predictably instant, a problem will occur.

p.s. Predictably instant: The performer can use common sense and basic isntinct to determine when feedback of the performance will occur. An amp 10ft away will obviously take about 9ms to reach the ear, and the various determinations the ear uses to localize sounds are present. Natural and instinctive, no conscious thought necassary.

Performances become "unpredictably instant" when the performer realizes the source of feedback (head phones) SHOULD be instant. If it is not, it becomes a conscious thought, interfering with the primary function; performance. The various forms of localization, timbre recognition and visual feedback are impaired, cause the performer to focus on a less natural form of receiving feedback on their performance.


goodness, I could go on for ages really. twas the subject of a 30 page report I did in college.

[DClarkson2004]

Mix has 5ms on the card. Some say this can affect performance during overdubs.

Hi Jax!
Me, being too lazy reading the entire thread:
I also run my stuff on a MixCore;
I didn't even know that 5ms were on the card, while recording.
C. Lambrechts said that seemed to be "an afwul lot".
So, if you record with a 5ms-latency, one can nudge it back 5ms after the recording,right?
I guess that on an 'Accel-super-dooper-card' you can have 0ms latency, but what's the big deal?
J-lo's recordings were done on Mix-Core with the same latency,
so did sooo many artists when HD or Accel didn't even excist!
You're falling for the same marketing-trap, saying that PT Mix-cards are worthless....why?
People nowadays can record their music at 192Khz....well, that's just great when all cd-players all over the world still run at 44.1;
Get the picture?
Don't be too worried about latency; nobody will hear it

Cheers!

[BrianT]

Auxilliary,

Exactly. You get it, where it matters, trying to generate some emotion while playing or singing.

Fine people in this thread, will you bear with me for a minute? Oh, scratch that. I'm starting a new thread. I hope Aux will join me. There is something important to be learned on this subject, and it deserves it's own thread. I think many of the younger members might have some new food for thought when we're done.

[BrianT]

Originally posted by RKrizman
Of course it would make a difference, in a live setting as well. Ever try to play string band music with players 25 feet from each others sound sources?

No, I haven't. But I guarantee it would be pure joy compared to trying to play *anything* in headphones with 25ms of delay on your monitoring.

'x" milliseconds of monitor delay in your headphones does not equal being "x" feet from your instrument. It's far more of a problem in the cans. That analogy is bogus.

[nikki-k]

Hi!
Aux: Yep, hit it on that post IMO.

I spent alot of time learning, and then doing, FOH (as well as doing moniotrs for live, etc). When I got the urge to play again, I would go wireless, and dealing with the inherent latency of being far from the "source" came into play.
Add to that using weird FX while playing (reverse, delayed notes, etc), and I guess I developed that "relative" sense. I can cope with a greater latency while writing, and even while tracking to some extent. Guess it is a subconscious compensation.

Having tracked with cans, and being in the control room (canless), I have had times using analog that cans were preferable. Turn lights down, catch the vibe, hit Rec. I have kicked everyone out of the control room when frustrated with a take that just wont seem to come.

As a performer, the advantages (to me) of using a DAW (looping during record, instant RTZ/punch-in point, not worrying I am wasting tape, etc) vs the latency thing have simply pushed me to accepting the latency. Improvise, Overcome, Adapt No worries with a 1-5 ms, although the closer to 1ms, the happier I am.

(BTW- Brian- I do miss PARIS still, even three years out now. Kudos for sticking with it!)