Easy : just buy some ear plugs and take 3 steps back !
Just going along with the joke
2 Likes
I used to be involved in telecom engineering. The human brain can detect audio latency at a minimum threshold of 20-30ms. Some people can âfeelâ latency of 15ms but we never engineered for this because its rare. Summary - the latency you are talking about here seems to be well below known levels that humans can detect. Which explains why my QC feels âright thereâ no matter how much processing is being used. If you want to test yourself you can set up a signal echo with varying delay time, have someone else increase delays, and see when you can actually detect a delay.
5 Likes
Got to be careful with how you present this. There is a difference between being able to âdetectâ latency, versus being able to âdetect latency while playingâ. Humans can easily detect latencies under 1ms. For example, two identical tracks played in a DAW while slowly increasing the delay of one, you will hear flamming and/or phasing even at sub-1ms. I know I can. But detecting latency while playing an instrument is what we are concerned about here. And typically humans can tolerate quite a bit more latency in this case (~15ms before it starts to get annoying)? The brain usually compensates (we play a bit ahead of the beat without even realizing it), but there does reach a point where itâs intolerable and negatively affects the feel of your playing. Iâve only noticed too much latency in the QC (playing, not measuring, as I described) when using the most busy presets and before the recent update that improved latency. IMO itâs all good now, whatever you are doing with the QC, totally acceptable.
3 Likes
Only if the original and delayed signal are played at the same time, but thatâs a physics thing, not human perception.
If you are playing a note on a keyboard, for example, and someone slowly introduces latency without your knowledge, you will only start to notice somewhere between 10-20ms.
2 Likes
Good to hear that the last updates improved the situation for you and maybe others who had issues with the QC latency so far. It was quite a heated discussion a while ago.
I myself never had the issue since I only use relatively simple presents and no send/return effects.
2 Likes
I think exactly this is the âproblemâ why a lot of users think they notice the latency and are unhappy: They play at low living room volumes and hear both, the guitars acoustic sound and the output of the QC. The current latency really is no problem when playing loud enough / through headphones so that one isnât able to hear the guitar / pick noise/ string noise itself.
Also a super important thing when dialing in tones you intend to use in a live / rehearsal / recording setting: Make sure you canât hear the guitar directly, because it greatly influences the perceived tone you get (e.g. hearing the pick hitting the strings lures you into believing your tone has a lot of clearly defined attack where in reality it could be a flubby undefined mess)
10 Likes
Hello David :-). With all my respect, thatâs why I sold my QC, too much latency. I bought a Boss GX 100 and problem solved for me. Hope the QC get better in the future, it is the most user friendly music hardware in the market. I really loved it but âŚ.
1 Like
This isnât really the issue at all.
The issue is actually that in a professional live setting, say your preset is at around 8ms of latency - which one of my big ones is, using the latest firmware - if you start adding in a digital desk, digital IEM transmitters and receivers and whatever else is in the path, youâre gonna soon be looking at 10-12, maybe worse, milliseconds of latency. Thatâs when it starts to become noticeable. Iâm disappointed that it doesnât appear to have been properly fixed in the update.
Sure. But there is a limit how low of a latency you can achieve with digital gear. Axe fx for example can also have presets with 10ms and beyond. High quality A/D and D/A conversion will always introduce a fixed latency and for audio processing often a lot of buffering is needed. The QC is currently on par with other comparably powerful modelers. Maybe digital stuff isnât for you if you are concerned about that. But then you have pretty high latency also if you are standing at some distance to your cab / speaker.
4 Likes
Is it on-par though? It seems a little behind to me, which is a shame as it sounds great and the form factor and usability are fantastic.
Can anybody tell me how to set up the Usb audio properly in Logic to try the ping with that? That should just do the internal latency, right?
Reposting in case somebodyâs interested :
- USB in the QC.
- Load your preset on the QC and turn off all blocks (otherwise the ping is altered and cannot give you a consistent result)
- Set your input on the row to USB X (I use USB 5)
- Set your output row to USB X (I use USB 7)
- Load an new Logic Project (48k) and put an I/O Plugin on the track
- Set your output to USB X (USB 5 for me) and input to USB X (USB 7 for me) and ping.
- the samples you get back need to be converted to ms with this formula :
1000 / (sample rate / samples). 48000 for the sample rate (the QC clocks at 48k)
BTW, if you use the FX1 or/and FX2 Send/returns, you should either test with just a cable between the Send(s) and Return(s) or turn off (or set 100% dry) your external pedals/gear.
1 Like
Although some handle it better than others, youâre right about all current modelers battling with latency introduced by the signal chain and types of blocks and routing employed, as well as the inherent latency in todayâs A/D and D/A conversion technology.
Although it can help put it in perspective, the latency introduced by the distance from your cab is not a justification for signal latency on a modeler, but instead, all the more reason to have an ongoing focus on reducing it. The additional latency introduced by a modeler can further negatively impact the feel and responsiveness caused by the distance from your amp.
Unless you go to IEMâs that cab to ears latency is basic physics and something we will never get around, so reducing the latency on the modeler is our only available and best option.
Going fully digital just screams for an IEM / silent stage / direct to FOH setup. Maybe thatsâs just me, but thatâs the reason I always take the sound travelling speed out of the equation when thinking about on stage latency.
1 Like
Why when I do this do I not get anything but just 0 samples? It doesnât seem to move?
WeirdâŚ
The QC is set as your audio device ?
Do the test on an empty QC preset with just a volume block, you should get a positive result (tiny but positive)
Just noticed I have the software monitoring on (not the auto, just the green one)
Canât verify cause the QC is in the carâŚ
Ok I think I got it working. Patch 1A Brit 2203 is showing as 214 samples which is 4.8ms. Is that what everyone else gets?
One of my main presets is 357 samples / 8.1ms on the main rhythm tone and 548/12.4ms (!!!) on my lead sound. Thatâs with a capture on row 1, capture, IR and volume on Row 3 and a delay, reverb and gain block on Row 4. Nothing mental.
EDIT: I should add, the preset as a whole is using 19 blocks; wah, compressor, four captures and a trem on Row 1. Row 3 has 3 captures, an IR and a volume pedal and row 4 has two delays, three reverbs and a gain block. I didnât think they had much impact if they were switched off however?
Are you on the latest CorOS ?
Nothing in the FXLoops ?
Yup, 2.2.2. Nothing in the loops either. Edited the above post a little.
Yes, just saw that.
Every block adds latency (on or off).