Has anyone found that audio interfaces seem to de-sync their hands slightly? It’s something I’ve always wondered about but I can’t tell if it’s mental or not.
For me it’s more generally about tone. If the interface is working properly, usually the lag is negligible to my ears - or perhaps I adapt to it quickly. But if I’m playing thorugh a tone I don’t like, everything about my technique goes wrong!
Yes, there’s a delay, when going through your computer, and that’s what’s causing this. Depending on your setup it can be noticeable.
I have my axefx connected straight to my studio monitors, so I can track while recording over USB, for this reason.
1 Like
I’m using a pretty affordable interface (Focusrite Scarlett 2i2 1st Gen) and there is no lag at all. It feels the same as playing through an amp in that area.
Most of the time is a matter of Computing power that results in latency. Nowadays the audio interfaces can do the job fairly well.
The 2i2 (both versions AFAICT) have a ‘Direct monitor’ switch. In the ON position, this routes the signal both to your computer and straight to the connected monitors. It was added exactly to solve the problem with the round-trip delay.
I’m guessing that this is set to ON for you 
Yeah, I have an Apogee Duet and then the small JAM interface they make, too. I think for the JAM it’s 18ms delay and the Duet is 7 or so.
Right now I’ve been sticking with my THR10X for practice. There’s no delay on a solid state amp so it makes me feel better at least lol.
I was going to make a new post about this issue but then I found this thread – I hope you can tolerate its revival. I’ve been noticing that moving from a few different latency-plagued rigs I have to a traditional tube amp that there is a striking difference in my left and right hand sync. One set-up is the cheap and hence forgivable iRig device that I bought to give my family a break via headphones. But, also running the Focusrite Scarlett Solo into my MacBook Air leaves me feeling all out of sync as well. I realize there’s a direct monitor feature but that leaves me with a clean, dry signal. I guess the latency is in the MacBook and its plugins. Move over to the Marshall and I’m in perfect sync. Maybe more advanced players can adjust for this lag without thinking of it but being in the learning process makes me want to not use anything with lag for practice.
My sentiments exactly, I’m very sensitive to latency as well.
The last software rig I was using incorporated a Babyface Pro that had a latency way below 5ms, was still no match for my tube amp based rig in tone and feel.
I refuse to put up with latency for now but I’m not a gigging musician yet, maybe later this year with the band. I hear playing live is going to induce a lot of latency issues due to monitoring, stage size, layout etc. I hope I can pull off old school type gigs with big iron.
I miss the 80s where HIFi meant 12" woofers etc. Rock sounded good dammit, there’s no substitute for moving air. Sorry I digress.
2 Likes
To some extent, it seems to me that latency can be reduced at a driver/software level by reducing the number of samples.
One user around here, I think it was @eric_divers, mentioned that the typical delay of a live rig is ~6ms, which should be attainable with good interface + good computer + good plugins.
Also, am I wrong is saying that the Yamaha THRs are actually amp simulators rather than “real” amps?
There’re different tricks to reduce the latency. For windows OS: switch to ASIO or WASAPI. Reducing the buffer (number of samples) helps too, though it can’t be made too small because software starts “choking”. 256-512 samples are usual choice. Increasing the bitrate helps too, so it worth to use maximum bitrate possible (96k,192k).
On my notebook I get latency about 6ms which is equal to standing 2m from loudspeaker. It’s an old notebook, with mic input only (I had to make little schematic to get high input impedance).
Yamaha THR is a simulating amp indeed, though it has nice sound. From it’s description I may suppose that they chose the approach close to that used in Revalver app: simulating amps by simulating all stages (valves+caps+resistors etc)
Ok, I’d seen some chinese USB audio interface and I decided to give it a try.
It was better than I expected. Intermediate interface (USB) don’t influence latency noticably despite my worse expectations. Actually I believe that now I have lower latency than before. It has ASIO support and 192k bitrate which is essential for low latency. They even claimed high input impedance though I didn’t check it yet. As for sound - it kind of lack some high frequencies though it could be my pickups… Settings are a bit tricky and I didn’t find a way to use my internal soundcard for output (if you use this USB interface you have to use it as input and as output simultaniously).
Despite the presence of clipping LEDs sound becomes a bit distorted before they starts to flash, so I have my gain regulator at halfway to max. And, yes, it has gain regulator, volume regulator and blend regulator. Latter one allows you to hear your clean guitar sound before sending it to a PC (nice thing to check the latency).
Anyway it worth the money (~50$) and is a good solution for me since I use notebook, which, obviously, don’t have PCI slots, firewire or whatever.
If 50ft is 44ms for sound to cross, why so much concern (50ft / (343m/s) is Google search)? And what about pipe organs?!
It seems the main trick would be in-ear monitors.
In fact, the most interesting way to think about latency might be in terms of the closest note of that duration. (200bpm 16ths is 75ms or so, etc.)
It’s actually 13.(3) ms. (200*4/60)
You are right that is is 60s/800, but run it through your calculator again, you likely had a typo. 
Stupid me… It’s 13.(3) picks per second (pps?). That’s what you get when you spend your time playing Tomb Raider till 2am…
P.S. I didn’t use calculator.
I absolutely notice this. Another way to look at it is a 22ms latency is roughly equivalent to playing with your amp 25 feet away! (based on the speed of sound being around 1116 feet per second.) If you would ideally be more like 6-12 feet from your amp then that’s 5-11 ms of latency. Lower than most interfaces will go.
22ms is a way too large 
6ms is a typical delay that could be achieved on almost any trashy win PC. I get it on my halfbroken notebook with internal soundcard with no linear input, no ASIO support. Though I had to use typical tweaks (like ASIO4ALL and stuff).
…
It’s an interesting topic. If you are used to play without headphones in a studio then you don’t need a latency lower than 3-4ms, because it means that your ears are closer than 1-1.4m to loudspeaker. But if you usually mike a cab and listen through headphones then the latency could be uncomfortable.
Anyway, I found that my hands make ‘autocorrection’ when I deal with a latency up to 15ms. It takes about 2-5 minutes to get used to a new latency. I guess that’s why pro artists play well whether they are on stage or in the studio. Our brain is a smart thing )
All good points- but to be clear are you talking about 6ms while playing through a guitar plugin? That’s what I meant- the plugin will add some additional latency. That’s what I experience in protools. Cheers
Yes, 6ms after all processing, which is kinda cool when you know what’s really going on behind the scene (like driver stack stuff, win multitasking, ADC, buffering, memory blocks transfers, processing (which is a hell of work), DAC, etc)
I measured it with my mic input turned on. So I had direct sound and the sound after being processed. Then I recorded it on my smartphone, got the waveform, measured - and voila.
You can actually calculate pure processing time more or less precisely. With the block size of 512 samples and 192k bitrate software (host+plugin) must process the block faster than 512/192000 = 2.6ms (or else it wouldn’t have time to get and process next block).