NOOB needs help

[Andrew McDougall]

Hi,

New to all things audio except from the playing of the guitar.

I am using a Scarlett 2i2 into my PC which has a ryzen 3700x processor (if that helps?)

I turn off direct monitoring on the interface and hit the echo button in cakewalk but there is a delay. I have tried to slide the Buffer size slider to fast but when i hit apply it defaults back to 10.7msec/512 samples. Focusrite device settings are Sample rate -192000, buffer 1024

 

EDIT: as i am typing i went back and set focusrite device to 44100 and matched cakewalk settings to this in the ASIO panel and latency was set for 2ms but when i do this i can no longer hear anything.  So at the moment i have a delay or nothing which aint right.

A quick help with this would be great to get me up and running and then i can delve into the no doubt endless stream of youtube tutorials et al to get further up to speed.

 

Danke tune

[reginaldStjohn]

I don't know if there is any "quick" help but getting low latency while recording requires you to set the buffer size on the Scarlett lower. Use the windows control panel to do this. Cakewalk often can't directly change the buffer setting so you have to use the windows or vendor supplied way of changing the buffer settings. 

Now, the no sound issue is another one. Start a new project and make sure that the audio interface and cakewalk are set to the same sample rate and depth. Then record something or drag in an audio file. Make sure your tracks are routed to the main bus and the main bus routed to your Audio Interface outputs, assuming your audio interface is setup correctly in Cakewalk's settings.

 

[bitflipper]

Welcome to the cult, Andrew. The learning curve will be worth it in the end, trust me.

That echo you hear is the time it takes your computer to process the data, which is referred to as "latency". It's unavoidable, although you can, with a powerful-enough computer, reduce it. Maybe even reduce it enough to not be noticeable (as 2ms should be). But it'll always be there, which is why audio interfaces feature direct (sometimes boldly claimed to be "zero-latency") monitoring.

Three main factors dictate overall latency. The latency introduced by larger buffers may not be the most significant of them, and the latency reported by your interface driver will never be accurate because there are factors it can't know about.

First, there is some latency built in to the interface that you can't do anything about. High-end interfaces can have very low internal latencies (~0.5 ms or less) but those are very expensive. A mid-range interface like your Focusrite employs some engineering tradeoffs that sacrifice latency for cost, so its internal latency will be 1 or 2 milliseconds. But way cheaper than a high-end box. Solution: spend the money on a better interface. Didn't promise it would be a good solution.

Second, there is the latency incurred by the computer system itself, transferring data from the interface into memory, juggling multiple buffers within memory, performing any needed conversions to the data, then experiencing the same delays when spitting the data back out (when echo is enabled). Assuming a faster CPU isn't in your budget, the best solution is to do one or both of the following: decrease buffer sizes or increase the sample rate. However, both will be limited by your hardware (e.g. disk drives' throughput may not be able to keep up with very high sample rates, or they'll fill up too fast). That's most likely why you hear nothing when your buffers get down into the 2ms range (try shooting for 5ms, which ought to be adequate unless you're Joe Satriani).

Third, there is the overhead of additional processing within the DAW. This includes both the data housekeeping that is the DAW's primary duty, as well as FX plugins. The latter is the biggest variable, because depending on the type of effect the time it needs can be huge. Solution: don't use plugins while tracking, only adding them in during the mixing process when their latency is no longer a problem.

Of course, you can avoid all (ok, most) of these complications by using your interface's direct-monitoring feature. The downside is that you'll not be able to use the computer as a stompbox, and no reverb on your vocal while you're recording. Some (again, more-expensive) interfaces feature internal effects that solve this issue. For the rest of us, we can do one of two things: use outboard effects while recording, or (better) learn to record dry and add effects later.

When I say "learn", I mean that literally - it's an acquired skill that takes practice. We're not used to hearing guitars and voices completely dry and it's uncomfortable at first. The payoff is that you'll eventually accumulate a much wider array of virtual effects than you could ever afford in hardware form. Even better, you won't need to commit to an effect until you've heard it in the context of the finished song, leaving you with far more creative options.

[Gary Bradley]

Hi Andrew,

I'm running FW hardware too so my guess is you might be trying to reset the interface while CW is still running? If so, CW is likely overriding FW Control. If so, try this:

Terminate CW if it's running. Set FR Control software to desired settings e.g. 48kH/24bit 512 samples which is commonly used. Restart CW and ensure your project is set to the same in Preferences.

CW and FR control shouldn't depend on Windoze sound control settings but I'm convinced they get confused. Sometimes I get locked out at 96kH and neither CW or FR will work. I have to switch all devices off, restart Windoze, set system audio Playback and Recording options to the same settings I record at and then fire up FR, check the settings and then CW. 

As a rule of thumb, 48kH/24 bit @ 512 will give ~10s latency whereas 128 samples might give  about 5-6ms latency. ~10ms is a good compromise in a decent system for tracking but 5ms will likely glitch when the track count is high or there are two many plugins running, especially virtual instruments (which you can freeze before recording). Of course, this depends on your system config.

Let us know how you are getting on!

PS it helps if you can include your full system specs in the post footer!

Best  

Gary

[Starship Krupa]

First, a sample rate of 192000 is high for a novice recordist. Even 88.2K or 96K are considered by most to be for use in special circumstances.

A system with a Ryzen 3700x is likely to include other components that make it so you will be able to track with low latency.

You don't say anything about the complexity of your project, how many plug-ins, etc., but I'll presume that since you mention "playing guitar" that you're trying to record audio, not MIDI with a virtual instrument.

So the "delay" you describe occurs between when you sing or play a note and when you hear it in your monitors (or, likely, headphones).

Okay, the thing to do is to set Cakewalk up to use ASIO Driver Mode in the Playback and Recording preferences.

In the control panel for the Focusrite, set your recording sample rate to 44,100KHz/24 and 4mS. That should be plenty low enough latency at the interface level.

Then try the following steps without adding any plug-in effects:

Set up a small test project and first make sure you can play back audio in Cakewalk. Just drag an audio file from somewhere on your computer onto the Track View, and see if it plays back okay.

If that works, plug in a mic, or plug your guitar in to the interface, and set up an audio track, and set the audio track to have that channel of the interface as its input. Click on the Monitor button on the track and see if you can hear the signal coming from the interface. At this point, if everything is working okay, you should be hearing your voice or guitar nice and crisp, with imperceptible or barely perceptible delay.

If you can't hear yourself, something went wrong. Your system is powerful enough to run one track at 4mS. Make sure you're setting the correct input channel, phantom power is turned on, etc.

If you're hearing enough delay to call it "delay," then make sure you don't have your interface plugged into an external hub or something.

Whatever else you do, ASIO Driver Mode is the preferred setting for most efficient operation, so configure your system that way if possible.