The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
What is your tolerance for Latency when tracking THROUGH software?
View Poll Results: What is the largest latency you LIKE working at for critical tracks (like vocals)?
No latency - only analog monitoring is acceptable
47 Votes - 26.70%
Needs to be under 2.4ms
24 Votes - 13.64%
Needs to be no more than 3.4ms
18 Votes - 10.23%
Needs to be no more than 5ms
30 Votes - 17.05%
Needs to be no more than 8ms
44 Votes - 25.00%
Needs to be no more than 14ms
10 Votes - 5.68%
I can work with anything
3 Votes - 1.70%
Voters: 176. You may not vote on this poll

Old 29th December 2013
  #1
Lives for gear
What is your tolerance for Latency when tracking THROUGH software?

Just for fun.....
What is your tolerance for Latency when tracking THROUGH software? Mostly concerned with critical tracks like vocals, live solo instruments, drums etc ... wherever "you" need the lowest latency... it is tough to choose relevant breakpoints so just go with what is closest to you max -

For reference I will put the following out there. (all measured analog to analog with function generator and oscilloscope). I do not have measured info for any others but hopefully those who answer can find out about where they are at in terms of ms (could have done this as samples but that seems even worse....

AVID HD Native @ 44.1 AVID 16X16
32 samples ------ 3.35ms
64 samples ------ 4.80ms
128 --------------- 7.68ms
256 --------------- 13.52ms

AVID HD Native @96kHz AVID 16X16
64 samples ------- 1.80ms
128 ---------------- 3.12ms
256 ---------------- 5.8ms
512 ---------------- 11.12ms

ULN8 (a firewire interface through PT11) @ 44.1
32 samples -------- 5.6ms
64 samples --------- 7.04ms
128 ------------------ 9.94
256 ------------------ 15.76ms

ULN8 (a firewire interface through PT11) @ 96k
64 --------------------- 3.86ms
128 -------------------- 5.2ms
256 -------------------- 7.8ms
512 -------------------- 13.8ms

HDx at 96kHz = 0.7ms

HD PCIe = 2.4ms

--------------------
To me anything over 3ms is noticeable. Not bad - but noticeable. 1.8ms sounds better to me than anything higher so pretty much exclusively work there. I could live with up to 5ms but it just doesn't sound anywhere near as good to me as 1.8ms
Old 29th December 2013
  #2
Gear Guru
 
Brent Hahn's Avatar
 

I'm still on HD PCI for the most part. The latency is noticeable but I live with it. I don't mind nudging. Heck, with something like EastWest Symphony nudging (more like dragging) is a given anyway.

I'm not gonna vote because my experience with those longer latencies is so limited. But the longest one in the list is shorter than the PT Short Delay "faux double" default, so I'm guessing I wouldn't find it intolerable.
Old 29th December 2013
  #3
When tracking vocals or acoustic guitar (where there's undelayed sound occurring in the environment), a little latency doesn't bother me, presumably because the acoustic bleed gives me the timing cues I want.

However, when tracking electric guitar D.I., all bets are off... it feels totally unnatural to me. I tried using Guitar Rig 3 (included in the pro version of my DAW) yesterday at 96 kHz w/ a 64 sample buffer (just under 2 ms) and it was totally demoralizing. I was just hatin' playing. I almost put my git away but thought, geez, I need to wash the taste of that out of my mouth, and plugged my guitar straight into my Blues Jr amp (what a concept!) and all of a sudden, I didn't hate my git any more (and ended up having a lot of fun for a couple hours).

FWIW, I also have a POD XT over here that a pal parked here a few years ago. There's a certain appeal to being able to dial up cookie cutter guitar sounds (because sometimes you just need that sound, eh?), and the POD XT physical knob interface does make it reasonably inviting to spin knobs and tweak settings -- but ultimately I usually find myself irritated by the latency and clearly not playing my best... and there's a very noticeable lack of responsiveness to playing dynamic and the 'vowel' formation I chalk up to shifting frequency responses from shifting impedance relationships between guitar and amp. But it's time that's the real joy-killer --- I find it very difficult to play melodic lead even with the tiny latencies involved -- latencies, it should be noted, that are along the same as one would get from having his guitar amp a few feet away. (Shouldn't be a problem, right? Can't tell you why it is. I don't want it to be. Just feels wrong, like I'm playing in molasses... it's sticky.)
Old 31st December 2013
  #4
Lives for gear
TTT ^^^^
Old 1st January 2014
  #5
Gear Nut
 
Randyman...'s Avatar
 

Quote:
Originally Posted by theblue1 View Post
I tried using Guitar Rig 3 (included in the pro version of my DAW) yesterday at 96 kHz w/ a 64 sample buffer (just under 2 ms) and it was totally demoralizing. I was just hatin' playing.
What other plug-ins were running in your session? And what DAW & Interface?

I used to run into latency issues even at 32-Samples of ASIO buffering (I'm @ 44.1K), but it was due to the fact that Nuendo's PDC was affecting the realtime monitoring. If you have any lookahead limiters or Linear Phase VST processing in your project, you can expect them to have objectionable latency, and PDC will offset all tracks to line up with the latent processing (even if the latent plug-ins aren't on armed/monitored tracks!). I had to specifically omit some of my favorite plug-ins to alleviate this drawback of Native Monitoring, but all latency concerns have completely vanished in my setup.

I've been ticked pink with both V-Drums (Battery/BFD) and Guitar (Amplitube/GuitarRig) tasks at 32-Sample buffers through RME interfaces. I even combine an ANALOG Marshall 12" Combo (no AD/DA at all) with a 4x12" 1960B that is driven from Amplitube VSTi (via a Marshall Tube Power Amp), and the comb-filtering from Analog Combo to Amplitube 4x12" is not even an issue. Granted the tones are somewhat different so that might help mask any detrimental comb filtering, but I CERTAINLY can't hear any delay or slapback from the Analog Combo to Amplitube VSTi 4x12" with my ears between the two. The guitar's "Feel" does not change if swapping between the two. Just as you propose - the latency is on par with distance from the amp (speed of sound) - so why SHOULD you expect to detect this latency? I suspect additional latency is cropping up in your system without your knowledge.

I don't have any analog monitoring capabilities in my studio, and ALL of my monitoring happens through the DAW (including ASIO - I monitor "Wet" even in the cans). I just recently (last month) upgraded to a RME MADIface-XT and it has built-in DSP with TotalMix-FX, but the drop in latency from using Nuendo+ASIO @ 32-Samples was not all that noticeable in my low-latency rig.

My vote was for no more than 5ms, but less than 8ms would be tolerable to most IMO.
Old 1st January 2014
  #6
Quote:
Originally Posted by Randyman... View Post
What other plug-ins were running in your session? And what DAW & Interface?

I used to run into latency issues even at 32-Samples of ASIO buffering (I'm @ 44.1K), but it was due to the fact that Nuendo's PDC was affecting the realtime monitoring. If you have any lookahead limiters or Linear Phase VST processing in your project, you can expect them to have objectionable latency, and PDC will offset all tracks to line up with the latent processing (even if the latent plug-ins aren't on armed/monitored tracks!). I had to specifically omit some of my favorite plug-ins to alleviate this drawback of Native Monitoring, but all latency concerns have completely vanished in my setup.

I've been ticked pink with both V-Drums (Battery/BFD) and Guitar (Amplitube/GuitarRig) tasks at 32-Sample buffers through RME interfaces. I even combine an ANALOG Marshall 12" Combo (no AD/DA at all) with a 4x12" 1960B that is driven from Amplitube VSTi (via a Marshall Tube Power Amp), and the comb-filtering from Analog Combo to Amplitube 4x12" is not even an issue. Granted the tones are somewhat different so that might help mask any detrimental comb filtering, but I CERTAINLY can't hear any delay or slapback from the Analog Combo to Amplitube VSTi 4x12" with my ears between the two. The guitar's "Feel" does not change if swapping between the two. Just as you propose - the latency is on par with distance from the amp (speed of sound) - so why SHOULD you expect to detect this latency? I suspect additional latency is cropping up in your system without your knowledge.

I don't have any analog monitoring capabilities in my studio, and ALL of my monitoring happens through the DAW (including ASIO - I monitor "Wet" even in the cans). I just recently (last month) upgraded to a RME MADIface-XT and it has built-in DSP with TotalMix-FX, but the drop in latency from using Nuendo+ASIO @ 32-Samples was not all that noticeable in my low-latency rig.

My vote was for no more than 5ms, but less than 8ms would be tolerable to most IMO.
I was actually running the standalone -- it was one of those 'just for fun things.' I've had it for a while but it was crashy on my old, quite modest single core so I uninstalled it long ago. Of course, the latency on that machine was too high for realtime use, anyhow. But I liked the idea of using it for reamping keyboards and such. But, I thought, now that I have a host capable of working at 96, why not see how it works out for me. I have to say that, latency issues aside, I didn't really give it a fair shake on sound, as I was using the really crap instrument input on my MOTU 828mk2 first and then switched to direct into the hi-z input of my board. I should have used a DB in the latter case and not even bothered in the former.

I have to say that my sensitivity to latency on DI guitar is pretty off-putting to me. I mean, heck, on raw numbers, it doesn't make any sense. Once you're down to a few ms -- I mean, heck, my amp is currently about 3 feet away -- I dunno.

FWIW, as I think I noted, I never had a 'problem' with monitoring my own vocals, and amplified or acoustic guitar, which was mostly what I was doing when I first got the MOTU -- for quite some time, blissfully using the DSP CueMix with no problem. Because I could hear the transient edge of the sound in the room (and in my own head in the case of vocals). But -- some time down the road -- when I finally went DI with the electric, I was really flummoxed by the sense of things being 'out of kilter' when I first plugged the electric into the MOTU's instrument input. Really, it was only the DI electric. But that got me thinking and I eventually hauled out my old analog board so I could monitor realtime through it.

BTW, you'd think I had a good sense of rhythm for how fussy about time I am. Sadly... heh
Old 1st January 2014
  #7
Lives for gear
 

I've worked with signed artists and professional session musicians who need to be under 3ms. They can actually point out that something is wrong. Even when monitoring thru a console there is still converter latency. There is really no such thing as "zero" latency when conversion is involved.
Old 1st January 2014
  #8
Lives for gear
 

Sound travels .343 meters per ms.

If one has 5 ms latency, that is like standing 1.7 meters from your amp. Most people find that acceptable. If you can't play with 5ms of latency, then you can't well play on stage.

I think the reality is that reported latency is typically not real latency. Latency freaks are people who were told the latency is low, but in reality is not, therefore they think they are hyper sensitive to latency, but their system is crap.
Old 1st January 2014
  #9
Quote:
Originally Posted by SabreChris View Post
Sound travels .343 meters per ms.

If one has 5 ms latency, that is like standing 1.7 meters from your amp. Most people find that acceptable. If you can't play with 5ms of latency, then you can't well play on stage.

I think the reality is that reported latency is typically not real latency. Latency freaks are people who were told the latency is low, but in reality is not, therefore they think they are hyper sensitive to latency, but their system is crap.
Well understood.

That said, I had no problem with the latency in my MOTU 828mkII CueMix doing vocals, acoustic or amped guitar. It was only, later, when I tried to do DI guitar that things felt fishy. And that hit me by surprise, I simply wasn't thinking about it, didn't expect it, since I'd been working that way for some time at that point, just not doing guitar DI up to then. (Now, of course, I'll always be expecting it. )

And, beyond the realm of near-zero-latency-relabeled-as-zero (as MOTU did to match competitors' marketing fabulisms going from the mk2 to the mk3), you know, those milliseconds do add up.

Pretty early on after I got the MOTU, I discovered that on my old rig with the kernel-streaming drivers I had to use it had an up-til-then uncompensated 7.8 ms track misalignment (new overdubs were laid in 7.8 ms late on the timeline). I was doing a bongo part. I'm not the world's greatest bongo player but after a couple passes I laid down something worth listening back to. I thought. But when I listened, it wasn't sloppy with some really rough spots -- it was all off, just consistently, wrong.

For a moment, I was stumped and then I thought, what about track misalignment -- that used to be an issue on my DAW but then they instituted what they called 'full latency compensation' which I naively assumed meant just that (but really just meant plug in comp only). So I did a ping-loopback test and, whaddya know, 7.8 ms off the mark. (The equivalent of about 8-1/2 feet of travel time at sea level, where I am. Just so we know we both still have our calculators out.)

I moved the track 7.8 ms forward and all of a sudden the bongos were sloppy with some really rough spots. Like I recorded them in the first place.

A second is really a really long time if you think about it.
Old 1st January 2014
  #10
Gear Maniac
 

AVID HD Native @96kHz will not run with a 32bit sample buffer? what happens, does the interface just crap itself? or is your computer that can't handle it?
Old 1st January 2014
  #11
Lives for gear
 

I can't sing worth a damned with more than maybe 0.5 or 1ms of latency into headphones but if it's live and via a stage monitor then maybe 5ms or so. Then again, I'm not much of a singer so anything throws me off. I always set up zero-latency monitoring (via an analogue console) in my studio, I've found that singers have a much easier time with pitch that way.
Old 1st January 2014
  #12
Lives for gear
 

Quote:
Originally Posted by work2do View Post
I've worked with signed artists and professional session musicians who need to be under 3ms. They can actually point out that something is wrong. Even when monitoring thru a console there is still converter latency. There is really no such thing as "zero" latency when conversion is involved.
Right, unless you have an analogue console or an interface that has an unconverted 'zero latency' analogue path directly from the preamp to the headphone amp. Technically there is a 'latency' that is related to the frequency response of the analogue electronics - in particular, low-pass filters - but that will be less than maybe 30 microseconds.
Old 1st January 2014
  #13
Lives for gear
 

Thales, just covered this in another thread so here's the repeat. HD Native users report a high error rate at 32 samples so 64 samples is the safe lowest setting for tracking (that doubles the latency right there). Most here still record at 44.1k or 48k sample rate and for their converters going any higher is not worth the trade-offs (that doubles the latency again from using 96k). Measuring latency means the whole time it takes to go from A/D converter input to eventually D/A converter output (marketing likes to skip steps to give the illusion of a lower latency number, it's not lying but omision of key data).

ALL digital has latency and you will not have true zero latency using digital regardless of their marketing. Using an analog mixer's cues or using a separate analog cue system is the best you can get. Next after that are interfaces that use DSP to skip the long trip through the DAW software (Avid HD / HDX (not native)). Other interfaces can skip the long trip through the DAW by splitting the input signal and returning it back out to the D/A converter via it's software mixer (direct monitering). Track through the DAW and likely most will not like the latency effecting their cue feedback loop.

It's pointless to have this poll since some artists are very sensitive to latency and others can adjust so what affects you will not be the same in others. Bottom line is if you change the artist's cue feedback loop, you affect their performance most often to the detriment. The exception to that rule is adding a bit of verb to vocal tracking but most likely you have them in a small space close miced and many singers do better if they hear themselves in a bigger room (more natural sound space). In my opinion cue latency is one of the biggest reasons for getting sloppy timing tracks as adjusting to high latency means guessing / reacting in advance with your playing. Try playing a keyboard with high latency in the cue, you have to guess how far in advance you press the key down until when you want to hear that sound in the cans. While you can adjust (some better than others), it's still guessing and you are going to be off on some notes, hence a sloppy timed track.
Old 1st January 2014
  #14
Lives for gear
 

Quote:
Originally Posted by dale116dot7 View Post
Right, unless you have an analogue console or an interface that has an unconverted 'zero latency' analogue path directly from the preamp to the headphone amp. Technically there is a 'latency' that is related to the frequency response of the analogue electronics - in particular, low-pass filters - but that will be less than maybe 30 microseconds.
Even with a console, if you are doing overdubs, the music you are overdubbing to or stacking vocals to has to come out and into Pro Tools through the converters. There is still latency there correct? I monitor with a console and just assumed there is always a small amount of latency regardless of how low I have on HD/HDX.
Old 1st January 2014
  #15
Lives for gear
Quote:
Originally Posted by Bassmankr View Post
It's pointless to have this poll since some artists are very sensitive to latency and others can adjust so what affects you will not be the same in others.
But this is the point of this poll :-) From a related DUC discussion there was strong opinions that anything up to 8 ms was just great and certainly only select few ever need less than 5ms -- I wanted to test that :-)

Quote:
Originally Posted by Bassmankr View Post
HD Native users report a high error rate at 32 samples so 64 samples is the safe lowest setting for tracking (that doubles the latency right there). Most here still record at 44.1k or 48k sample rate and for their converters going any higher is not worth the trade-offs (that doubles the latency again from using 96k). Measuring latency means the whole time it takes to go from A/D converter input to eventually D/A converter output (marketing likes to skip steps to give the illusion of a lower latency number, it's not lying but omision of key data).
1) It is entirely feasible with PT11 and HD Native to use the lowest buffer setting reliably. It takes a FAST computer and a carefully selected set of low footprint, CPU efficient plug ins (AVID, Softube, Toontrack, Exponential and a few others). I record every day at 96k/64buffer with no issues on HD Native.

2) There is always an RTL decrease at higher SR due to the decrease in A/D D/A conversion time. Unfortunately the RTL consist of a) 2X sample buffer (which doubles when you double SR canceling any gains here), b) A/D D/A conversion time (usually reduces by at least 1/2 when doubling SR), c) any additional time to get data on/off the data bus - FW,USB,TB,PCI etc -- in general PCI and TB can be taken to be zero. FW and USB have much larger values here thus the reduction in A/D D/A time may not make much difference.

2) The AVID I/O is the exception in that one gets MORE gain from going to 96kHz than you might expect. It is because the converters operate in different modes at 44.1 and 96k. The A/D, D/A conversion time on the AVID I/Os is: (as stated in first post - all measured with signal generator and oscilloscope analog to analog through PT11)

A/D D/A conversion time is: 44.1kHz = 1.9ms - this is measured using LLM
A/D D/A conversion time is: 96kHz = 0.47ms - this is measured using LLM

Resulting RTL is
44.1kHz/32 = 3.35ms
96kHz/64 = 1.8ms
Old 1st January 2014
  #16
Lives for gear
Quote:
Originally Posted by Thales View Post
AVID HD Native @96kHz will not run with a 32bit sample buffer? what happens, does the interface just crap itself? or is your computer that can't handle it?
Correct - 64 buffer is the lowest HD NAtive offers @ 96kHz --- thanks Matt
Old 1st January 2014
  #17
Quote:
Originally Posted by ProPower View Post
Correct - 64 buffer is the lowest HD NAtive offers
At 96k. At 44.1 you can select 32, of course.

Interesting poll. ;-)
Old 1st January 2014
  #18
Lives for gear
Quote:
Originally Posted by SabreChris View Post
Sound travels .343 meters per ms.

If one has 5 ms latency, that is like standing 1.7 meters from your amp. Most people find that acceptable. If you can't play with 5ms of latency, then you can't well play on stage.

I think the reality is that reported latency is typically not real latency. Latency freaks are people who were told the latency is low, but in reality is not, therefore they think they are hyper sensitive to latency, but their system is crap.
The issue is best illustrated with vocals. When you sing you get a fairly latency free signal from your mouth. Now you put on headphones and get a delayed signal that combines with it. The resulting comb filtering causes anywhere from a slight to a dramatic tonal shift and loss of clarity. Some are sensitive others not. I hear the jump from 1.8ms to 3ms clearly and anything past that sounds nowhere near as good.

I find the same thing with tracking solo fingertyle guitar. Drums are also that way for me since I hear the hit in the room + a delayed signal in my headphones that flams.

Electric guitar is almost the worst. The difference in delays between the cabinet hitting my ears and the monitor signal in the headphones can change the sound in a huge way (clean guitars here are the worst).

In all of these - I have found three types of folk:
those that hear these things and care
those that hear it and don't care
those that don't hear it...

Hence the poll :-)
Old 1st January 2014
  #19
Quote:
Originally Posted by work2do View Post
Even with a console, if you are doing overdubs, the music you are overdubbing to or stacking vocals to has to come out and into Pro Tools through the converters. There is still latency there correct? I monitor with a console and just assumed there is always a small amount of latency regardless of how low I have on HD/HDX.
If your rig is properly set up and adjusted, the DAW should compensate for the 'combined' round trip latency (cue mix out the DAC and new track into/through the ADC) and place the new track at precisely the right spot on the timeline.

In the past (and perhaps continuing) the drivers for many converters did not necessarily get the self-latency reporting right -- and/or some DAWs simply did not 'bother' with this compensation even if they did.

Back around 2004 or so, I think, I realized that my DAW, Sonar, was not compensating -- at all -- for those tracking latencies (when using the then-preferred WDM-KS kernel streaming drivers for it). With 128 sample buffers and 44.1 operation, that worked out to about 7.8 ms of misalignment -- new overdubs would be that time 'late' in the timeline.

I started having other people test their rigs -- various software and hardware -- and found that, with only ONE exception (maybe two, but I was not convinced the party in question fully understood the issues he was reporting), everyone ended up with some timeline misalignment.

And, actually, the one exception was still 2 samples off using his RME interface and drivers. Since he was the chief tech officer for the company that put out my DAW, however, it was interesting when he was confronted with all these people having done ping-loopbacks with what they'd thought were properly set up machines but had varying misalignments, mostly in the range of 4-12 ms.

At first, the guy tried to maintain that it 'wasn't the DAW's job' to adjust for converter latency but I and others kept pointing out to him that A) most folks willing to test showed significant misalignment problems and B) short of rewriting all those device drivers the DAW ought to give users a way to compensate for such cross-domain temporal misalignment (uncompensated conversion latency).

It didn't help his case that Cubase and Mackie's old Tracktion already HAD such adjustable compensation -- as well as automated ping-loopback testing to set it.

It took what seemed like a LOT of go-round, but he finally promised a fix, which was delivered in the next version (and improved in the subsequent).


PS... I've been thinking about my 'extreme sensitivity' to perceived latency in DSP cues. As noted, I fully understand and largely share the perplexion over sensitivity to latencies equivalent to a few feet of free-air sound travel. I can easily understand that some might simply feel my perception was the result of cognitive distortion -- it's certainly a persuasive notion, at least at first blush. Still, as I noted, I wasn't expecting to have any problems, I wasn't even thinking about it when I had the issue with the DI guitar.

The subsequent issue (the bongo track anecdote) was another place where I wasn't expecting problems -- and it took me a second to realize there was a problem (my first inclination was to just write it off to my bad playing -- but I certainly had a very strong sense that what I'd played hadn't sounded anything like what was coming out -- until I realized that there had been no compensation for the round-trip latency and adjusted the placement on the timeline by that amount, at which point, whaddya know, it sounded like what I'd recorded after all. 8 little ms, eh? But it made the difference between something that sounded just plain wrong all the way through and something that sounded as it had when recorded, vis a vis timing and other tracks. I think I'm going to look into setting up a test bed to investigate/explore these latency issue perceptions.
Old 1st January 2014
  #20
Lives for gear
 

Quote:
Originally Posted by ProPower View Post
The issue is best illustrated with vocals. When you sing you get a fairly latency free signal from your mouth. Now you put on headphones and get a delayed signal that combines with it. The resulting comb filtering causes anywhere from a slight to a dramatic tonal shift and loss of clarity. Some are sensitive others not. I hear the jump from 1.8ms to 3ms clearly and anything past that sounds nowhere near as good.

I find the same thing with tracking solo fingertyle guitar. Drums are also that way for me since I hear the hit in the room + a delayed signal in my headphones that flams.

Electric guitar is almost the worst. The difference in delays between the cabinet hitting my ears and the monitor signal in the headphones can change the sound in a huge way (clean guitars here are the worst).

In all of these - I have found three types of folk:
those that hear these things and care
those that hear it and don't care
those that don't hear it...

Hence the poll :-)
You are talking about recording with headphones in same room with miced instrument. For guitar, put the cab in isolation. For vocals use direct monitoring with some wet verb (latency doesnt matter). Drums dont need monitors.

I know about the comb effect, but what you are talking about is hearing two sources at once, not the highest latency required to play accurately. For the question you asked, with the tracking scenario you outlined, direct monitoring (no latency) is the only answer.

The question you asked was tracking THROUGH software (suggesting vstis), not monitoring through software while hearing original source, so it seems you have modified the question to support your view.
Old 1st January 2014
  #21
Lives for gear
Quote:
Originally Posted by ProPower View Post
Just for fun.....
What is your tolerance for Latency when tracking THROUGH software? Mostly concerned with critical tracks like vocals, live solo instruments, drums etc ... wherever "you" need the lowest latency...
From my first sentence...

I can see how it could be confusing from just reading the poll question. Sorry about that. I think the results are interesting anyway even if some are interpreting it like yo did... FWIW I find 1.8ms latency for things like vocals sounds very good -- though I don't (and won't) have analog set up to compare with which may indeed sound even better -
Old 1st January 2014
  #22
Lives for gear
 

Using HD Native's LLM for getting converter latency numbers is one thing however it's real world uses are fairly limited. It's basically just a form of an interface direct monitering trick of spliting the signal and skipping the DAW. You only get one stereo LLM track (in the real world you need more than one cue), and you can not add plugs to it.

As one poster mentioned, overdubbing a soft synth late in the game can be a bitch for latency. In that case you may be better off just using a few guide tracks with everything else turned off in the DAW that would add latency (frozen).

As with most things in the studio you have to pick your poisen, with Digital you have to deal with latency. If you use high sample rates and low buffer settings to lower latency during tracking then you have to be aware of errors, SRC effects, how your converters behave/sound at that sample rate, and how your computer deals with the extra data. If you can swing it some form of an analog cue system is one of the better choices.
Old 1st January 2014
  #23
TNM
Lives for gear
Only a keyboardist here but for vocals anything 10 complete round trip and under is fine. I didn't vote as there was no 10 option. I'm happy playng virtual instruments included drummers at anything 5.5 output latency and lower. Really 6 and under is fine, but I've been playing the keyboard since I was a toddler and was getting used to much worse latencies in the early turn of the century so today's latencies are a godsend for playing vi's.

Learning finally after 41 years to play the guitar (acoustic first) so will see how latency affects me when I finally go to record some. Progress is slow due to my back, I don't get much time to just sit with the guitar and learn.

However I would think with acoustic guitar I could just enable the interfaces near zero latency feature and use the onboard dsp for a bit of verb if I needed it and not have any issue.

My friend who is a decent electric guitarist(and teaching me acoustic) is happy with sub 5 round trip. My interface can't give him that through software fx so I'd have to buy something faster for him to track through a software amp (lowest is 8.3 RTL real value with mine including converter and hidden buffer latency).

So for now I just have him bring his amp over and once again use the interfaces zero latency feature. The dsp reverb is a bit nasty and springy so it actually works out quite well for a bit of monitoring verb.
Old 1st January 2014
  #24
Gear Maniac
 

Convrter latency is enough... no more tolerance after that.
Old 3rd January 2014
  #25
Lives for gear
Any more poll takers?
Old 4th January 2014
  #26
Lives for gear
Analog monitoring for me. Nothing else is good enough if you want a great feeling vocal, sweetly pitched and in the pocket. Same for all real instruments whether mic or DI.

For VI's 32 sample buffer. and for MIDI drums I listen to my VDrums module and record the MIDI stream rather thn trigger SD2 live.

As one poster said over dubbing a VI late in a mix usually means making a little stereo backing track to play to so you can lower the buffet back down to 32 from the 1024 I use for mixing. A PITA but well worth it to preserve feel.
Old 4th January 2014
  #27
Eat
Lives for gear
 

when I was a kid it used to fascinate me to be standing outside a public park watching some kid bounce a basketball and the sound of it would reach me a few instants later
Old 4th January 2014
  #28
Eat
Lives for gear
 

Quote:
Originally Posted by ProPower View Post
Any more poll takers?
why if I wuz gay I'd take offense at that remark.
Old 7th January 2014
  #29
Lives for gear
Monday bump :-)
Old 7th January 2014
  #30
Sky
Lives for gear
 
Sky's Avatar
 

Quote:
Originally Posted by Randyman... View Post
My vote was for no more than 5ms, but less than 8ms would be tolerable to most IMO.
Same here. I generally work at 256 buffer in PTHD (~8ms), and sometimes drop to 128 (~5ms) for critical overdubs.

Sky
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump