|
|
Post by gravesnumber9 on Mar 25, 2024 13:47:21 GMT -6
I'm asking because I'm currently rewiring my monitoring setup in my tracking room and was considering going to 96khz at the same time for better latency. You know, I've got everything pulled out from the wall... why not go all the way?
I decided not to do that and just to go back to monitoring directly out of the board instead but that got me wondering, how much latency bothers you?
I see people say anything less than 7ms is unnoticeable and that's just... well, that's a "wow" for my ears. It's very hard to believe that people don't notice 7ms or even 5ms roundtrip latency. I can pick up on virtually any latency at all, even 1ms.
BUT.
I have an odd way of perceiving a lot of visual and aural phenomena and I don't really know what's normal here. I don't want to get into the current babble about neurodiversity and all that bunk, but I have some unique skills to go along with many many more unique flaws. For example I can visually level things perfectly without a leveler (it's very handy when hanging pictures and I love testing this with a bubble level, it's quite the party trick...)
Stuff like that. Most of these "skills" are very annoying to me and others but some of them (like being the Human Leveler) are helpful.
Anyway, here's my question again... even the teeny tiniest bit of latency drums me absolutely insane. But I know that's not true for everyone. So what is the level that bothers you guys and, separately, what is the level that you consider acceptable for clients?
|
|
|
|
Post by Shadowk on Mar 25, 2024 14:06:18 GMT -6
Vocals and drums about 3ms. For a send to mix bus a couple of samples..
|
|
|
|
Post by bossanova on Mar 25, 2024 14:21:09 GMT -6
First off, non-Neurotypical wave 👋
Second, if it’s some thing, like a pad, or a part with a little more tolerance rhythmically, for whatever reason, I think I can get by with his high as 20 ms and just adjust my timing to compensate.
For something that needs to be rhythmically tight or syncopated, 10 ms tops, and that gets shorter based on the level of groove required. At a certain threshold, I just monitor direct, playing hardware and simultaneously tracking to MIDI if it’s a track that’s going to be replaced by a VST.
|
|
|
|
Post by svart on Mar 25, 2024 15:49:00 GMT -6
it depends.
If it's a wet/dry combo, then 1ms is noticable.
If it's just the wet signal, people don't start to complain until around 20ms up to 50ms or more.
|
|
|
|
Post by gravesnumber9 on Mar 25, 2024 15:57:22 GMT -6
it depends. If it's a wet/dry combo, then 1ms is noticable. If it's just the wet signal, people don't start to complain until around 20ms up to 50ms or more. This is what I'm looking for. Cuz yeah, the phasiness (phasyness? phaseness? whatever...) bothers me even on a wet only signal if it's my voice at literally anything other than direct monitoring. Obviously 1ms is less bothersome than 3ms and so forth, but it's all annoying. Then you go direct and it's like "ahhh... my voice is back, I'm me again." So I guess I'm really just talking for vocals. For other instruments I think our brains are already wired to accept latency because it exists in the real world. I don't expect a guitar string to instantly hit my ears when played in the same manner as I expect it of my own voice. So svart , in your experience, does that 20ms number change if we're talking about vocals? |
|
|
|
Post by Shadowk on Mar 25, 2024 16:11:53 GMT -6
This is what I'm looking for. Cuz yeah, the phasiness (phasyness? phaseness? whatever...) bothers me even on a wet only signal if it's my voice at literally anything other than direct monitoring. Obviously 1ms is less bothersome than 3ms and so forth, but it's all annoying. Then you go direct and it's like "ahhh... my voice is back, I'm me again." And people wonder why we go HDX or Carbon.. |
|
|
|
Post by chessparov on Mar 25, 2024 17:01:33 GMT -6
FWIW I got used to Direct Monitoring. On Scarlett/UA Volt/Mackie Blackjack. Mix Knob on Audiobox/(Blue) MBox 2 Mini. Yes! It's the Elephant Graveyard of Prosumer/USB Boxes!  Chris |
|
|
|
Post by gravesnumber9 on Mar 25, 2024 20:08:48 GMT -6
This is what I'm looking for. Cuz yeah, the phasiness (phasyness? phaseness? whatever...) bothers me even on a wet only signal if it's my voice at literally anything other than direct monitoring. Obviously 1ms is less bothersome than 3ms and so forth, but it's all annoying. Then you go direct and it's like "ahhh... my voice is back, I'm me again." And people wonder why we go HDX or Carbon.. Well I'm always HDX tempted and Carbon curious. Haha. But in my mixing / overdub room I monitor direct through the 500 ADAT, it was the tracking room that was concerning me. I finally realized I just need to use my console the way it was intended to be used (don't get me started, classic overthinking on my part) and monitor through that using the AUX's for what they're made for etc etc. Just finished and, yeah, breath of fresh air. It's almost like the people that designed these consoles had a pretty good idea of how to use them. Go figure. DAW as tape machine, works like a charm. I've decided that if I ever get rid of this console (not likely cuz I love it, but it is a bit of a question mark at times) I'll have to to HDX or Carbon. But for now, old school mixed with accurate latency offset is working perfectly. |
|
|
|
Post by bossanova on Mar 25, 2024 23:47:11 GMT -6
it depends. If it's a wet/dry combo, then 1ms is noticable. If it's just the wet signal, people don't start to complain until around 20ms up to 50ms or more. This is what I'm looking for. Cuz yeah, the phasiness (phasyness? phaseness? whatever...) bothers me even on a wet only signal if it's my voice at literally anything other than direct monitoring. Obviously 1ms is less bothersome than 3ms and so forth, but it's all annoying. Then you go direct and it's like "ahhh... my voice is back, I'm me again." So I guess I'm really just talking for vocals. For other instruments I think our brains are already wired to accept latency because it exists in the real world. I don't expect a guitar string to instantly hit my ears when played in the same manner as I expect it of my own voice. So svart , in your experience, does that 20ms number change if we're talking about vocals? See I disagree on the fully wet part. If I hit a key and the sound doesn't come out corresponding with that key press where I expect it, it's going to play hell with my timing. I can compensate to an extent but at a certain point it's just like swimming through mud. Of course there's some imperceptible latency even when playing a piano or a vintage synth, but for me that's virtually instantaneous, whereas digital round trip latency very quickly starts to sound like something is wrong. |
|
|
|
Post by gravesnumber9 on Mar 26, 2024 2:07:04 GMT -6
This is what I'm looking for. Cuz yeah, the phasiness (phasyness? phaseness? whatever...) bothers me even on a wet only signal if it's my voice at literally anything other than direct monitoring. Obviously 1ms is less bothersome than 3ms and so forth, but it's all annoying. Then you go direct and it's like "ahhh... my voice is back, I'm me again." So I guess I'm really just talking for vocals. For other instruments I think our brains are already wired to accept latency because it exists in the real world. I don't expect a guitar string to instantly hit my ears when played in the same manner as I expect it of my own voice. So svart , in your experience, does that 20ms number change if we're talking about vocals? See I disagree on the fully wet part. If I hit a key and the sound doesn't come out corresponding with that key press where I expect it, it's going to play hell with my timing. I can compensate to an extent but at a certain point it's just like swimming through mud. Of course there's some imperceptible latency even when playing a piano or a vintage synth, but for me that's virtually instantaneous, whereas digital round trip latency very quickly starts to sound like something is wrong. I think I can handle a few ms of latency for instruments personally. |
|
|
|
Post by ontravelling on Mar 26, 2024 2:36:59 GMT -6
While tracking, I feel like I notice and am bothered by any latency at all. Using the Cranborne 500R8 is great for me as I can monitor directly while tracking. Another strange thing is that I can also level things very well by sight. I still need to fine tune a bit when I pull out a level but the bubble is almost always between the lines to start, almost never use a level because of it.
|
|
|
|
Post by thehightenor on Mar 26, 2024 5:46:51 GMT -6
Anything over 1ms affects my groove and pocket.
I use various methods when tracking to avoid going over 1ms.
I track at 48KHz so my strategy to avoid latency is achieved through analog cue mixes and using physical instruments (guitar and miced cabs, bass guitar, real percussion) and if recording VI’s monitoring V-Drums and Nord keyboards whilst recording the MIDI performance.
|
|
|
|
Post by noob on Mar 26, 2024 7:19:50 GMT -6
I am skeptical anyone ever has issues at 1ms latency. For me, issues with rhythm and even pitch/tonality begin to arise between 3-4ms.
I can easily work with 1 or 2ms latency.
|
|
|
|
Post by svart on Mar 26, 2024 7:38:12 GMT -6
it depends. If it's a wet/dry combo, then 1ms is noticable. If it's just the wet signal, people don't start to complain until around 20ms up to 50ms or more. This is what I'm looking for. Cuz yeah, the phasiness (phasyness? phaseness? whatever...) bothers me even on a wet only signal if it's my voice at literally anything other than direct monitoring. Obviously 1ms is less bothersome than 3ms and so forth, but it's all annoying. Then you go direct and it's like "ahhh... my voice is back, I'm me again." So I guess I'm really just talking for vocals. For other instruments I think our brains are already wired to accept latency because it exists in the real world. I don't expect a guitar string to instantly hit my ears when played in the same manner as I expect it of my own voice. So svart , in your experience, does that 20ms number change if we're talking about vocals? It might. I think vocalists and guitarists have been the ones complaining most about latency. I'm not a singer, but I've tried over the years and honestly, I can't hear myself well enough through headphones when singing for it to even be an issue. I'm basically just trying to keep up with the timing and overall pitch. I might even add some reverb to my vocal cue so I can more easily detect the pitch. When I play guitar, I don't notice any perceptible latency until it's more like 20-30ms and it doesn't bother my playing until 50ms or higher. I wonder how people can actually hear less than 1/20th of a second of delay. It seems so small in context of time. However, to play the devil's advocate, I once had a guy tell me that he didn't like amp sims because they had too much latency. I set up a marshall amp sim to see if the tone would work in the mix. While sitting there he complained about the latency. I went into the other room to set up the amp and came back to change a patch and he started playing and said "that amp sounds much better!!" I had never switched it to the real amp. I never even turned it on.. He never knew I never turned the amp on. Only when he KNEW that he was playing through the sim did he mention the latency, so I think there is a strong perception effect, FWIW. |
|
|
|
Post by svart on Mar 26, 2024 7:40:50 GMT -6
I am skeptical anyone ever has issues at 1ms latency. For me, issues with rhythm and even pitch/tonality begin to arise between 3-4ms. I can easily work with 1 or 2ms latency. Yeah, 1/100th of a second seems like a very short time to perceive a change. Human reaction times to external stimuli is about 1 second. We just don't perceive the lag because we're always behind. |
|
|
|
Post by notneeson on Mar 26, 2024 7:42:18 GMT -6
This feels like a loaded question, how much latency can you tolerate: how tight of a player/good is your ear.
|
|
|
|
Post by thehightenor on Mar 26, 2024 7:45:30 GMT -6
I am skeptical anyone ever has issues at 1ms latency. For me, issues with rhythm and even pitch/tonality begin to arise between 3-4ms. I can easily work with 1 or 2ms latency. What! How do you know how latency feels to someone else. I can and have passed a double blind ABX latency test. I can easily detect the difference between 1 and 2ms. How - a lifetime of playing real instruments. If someone else can't - that's more than fair enough - but I'm not going to have anyone tell what I can and cannot feel. Sorry if that sounds sniffy - it's not meant to - I'm just saying. |
|
|
|
Post by doubledog on Mar 26, 2024 7:48:46 GMT -6
I never really cared if it was 1ms or 100ms. So I've never taken time to measure it. Either I heard it and it bothered me (or a client) or not. But tbh, ever since I got the Apollo I've been using direct monitoring through the UA console (and muting the armed channels in Pro Tools) and it's just never an issue (unless I forget to mute in PT). And yes, that means for overdubs I sometimes have to be monitoring the mute button (so a singer doesn't get the robotic effect for example) but since the exact timing doesn't really matter, it's certainly simpler than trying to punch in on tape...
|
|
|
|
Post by Dan on Mar 26, 2024 7:51:06 GMT -6
FWIW I got used to Direct Monitoring. On Scarlett/UA Volt/Mackie Blackjack. Mix Knob on Audiobox/(Blue) MBox 2 Mini. Yes! It's the Elephant Graveyard of Prosumer/USB Boxes! Chris get a used SPL Crimson or Neve 88m |
|
|
|
Post by noob on Mar 26, 2024 7:54:45 GMT -6
I am skeptical anyone ever has issues at 1ms latency. For me, issues with rhythm and even pitch/tonality begin to arise between 3-4ms. I can easily work with 1 or 2ms latency. What! How do you know how latency feels to someone else. I can and have passed a double blind ABX latency test. I can easily detect the difference between 1 and 2ms. How - a lifetime of playing real instruments. If someone else can't - that's more than fair enough - but I'm not going to have anyone tell what I can and cannot feel. Sorry if that sounds sniffy - it's not meant to - I'm just saying. I'm sure you can hear 1ms. Human perception is a wonderous thing. I was just saying that 1ms latency doesn't give me any issues related to rhythm or pitch, and I highly doubt it gives others issues. Although it might! I love recording analog tape because you can get the delay down below 1ms into microseconds. However, I don't ever have issues with recording until I get to 3-4ms. That's all I was really saying. |
|
|
|
Post by Dan on Mar 26, 2024 7:54:53 GMT -6
it depends. If it's a wet/dry combo, then 1ms is noticable. If it's just the wet signal, people don't start to complain until around 20ms up to 50ms or more. hearing it or feeling it? Cannot hear it under 40-50 ms easily because of the Haas effect which will change the sound even with 2-3 ms of latency. Feel is different. |
|
|
|
Post by svart on Mar 26, 2024 8:04:11 GMT -6
it depends. If it's a wet/dry combo, then 1ms is noticable. If it's just the wet signal, people don't start to complain until around 20ms up to 50ms or more. hearing it or feeling it? Cannot hear it under 40-50 ms easily because of the Haas effect which will change the sound even with 2-3 ms of latency. Feel is different. Very different things I think. It's easy to hear ~1ms phase anomalies from Haas if you have a direct wet/dry combination. If you just listen to the wet, then I think it's easily up to >20ms before most folks perceive it as some kind of noticeable delay, but probably can't really "hear" it until much higher. |
|
|
|
Post by Dan on Mar 26, 2024 8:06:24 GMT -6
I can notice almost all, even 2-3 ms, when playing. Going to 88.2 kHz and above on a fast computer to get the sub 2ms latency on the lowest buffers or having the direct monitoring digital mixer or blend knob is super important. I can also sense higher capacitance cabling, transformers and inductors in signal path. Get rid of them all including the quad cables for tighter performances and put them down after the preamp and digital recording if they make it sound better. You are not recording to tape anymore and then on consoles and mixers could be used as inserts to tape outside of the monitoring mix! And I love Daking pres and occasionally anything orange like Neve or browner API for weight but they can be felt versus an interface pre or something like a Grace. Only exception is I love my coiled vox cables for sounding like 60s and 70s and they take edge off of stupid pedals feeding stupid amps.
edit: a lot of cheap analog eqs without anything expensive in them will also distort and perceptibly slow down the audio even with something basic like a slight shelf.
|
|
|
|
Post by Dan on Mar 26, 2024 8:30:48 GMT -6
Mixing, I can notice higher buffers at 44.1 and 48 kHz. The faders lag and the meters don't match up on the dynamics plugins that use oversampling or lookaheads to not sound like crap or on hardware inserts. This can be very important for learning what you're hearing and getting the behavior the want from many dynamics processors. take an 1176, for a few vu off the release is normal and exponential and then becomes more colored. Maybe you do not want that or want it monitored so the meter moves on the transients where you want it to? or an ssl bus comp on auto should be set based to where the attack and release starts to speed up based on your attack and ratio settings for inaudible action or otherwise it can hold the audio down or hold the treble down. going to 88.2 kHz and up greatly reduces latency for many plugins. Weiss and MDW eqs become zero latency at 88.2 kHz, others at 176.4 kHz and up, and others, the anti-alias filters can be shortened, eg Molot GE goes from 187 samples of latency on 44.1 kHz to 31 at 88.2. Of course higher sampling rates waste cpu cycles, storage space, and will have increased imd from plugins with inadequate distortion prevention especially at 192 kHz. This is audible. gearspace.com/board/mastering-forum/968641-some-thoughts-quot-high-resolution-quot-audio-processing.html |
|
|
|
Post by mcirish on Mar 26, 2024 9:01:27 GMT -6
Latency during mixing doesn't bother me much but during tracking it does. Normally, this is only a problem when tracking VST instruments. In that case, I turn off all effects that have latency. In Nuendo, it's one button, so it's simple. Then I set the record buffer to 128 samples and I'm good with that.
But, 90% of what I'm doing is real instruments. I monitor those mics right from the Lynx Aurora (n), so there is zero latency. That's the only way I can ever do drums. Of course, there are no effects on the instruments I monitor via the interface, but that doesn't seem to bother any artists or myself.
|
|
