What buffer size do you use for recording vocals?

What buffer size do you use for recording vocals?

In particular, for recording vocals using direct monitoring via the audio interface?

Does the buffer size affect the timing of the recording in the daw at all?

Is sound quality affected by using a low or high buffer size?

The buffer size I use depends on how “busy” my project is with plugins. On a relatively fresh, sparse project, I normally am running at 44.1khz with a 256 sample buffer (on an RME Fireface UFX on Windows 10). However, near the end of the mixing stage, when I’ve got massive plugin chains on various tracks and buses, I normally have to push the buffer up to 512 or even 1024 samples in order to keep working without glitches. (I freeze tracks at this stage, too.)

I ONLY direct-monitor through my interface, never with Cubase’s track monitoring. And it’s mainly for this reason: If you monitor through your interface vs with your DAW, then Cubase will correctly align your audio recording with the project. The Cubase track monitor has to do some delay compensation magic, and I believe that does negatively affect your project if you record with it enabled. (It’s been a long time since I experimented with this, though.)

Anyway- your timing is not affected by the buffer size so long as you use direct monitoring. The sample buffer purely exists to ensure there’s always some audio available to stream; it does not and cannot affect the sound quality of the recording other than preventing pops and drop-outs when it’s big enough for the current workload and sample rate.


As long as your audio device driver is reporting the correct latencies then you shouldn’t have any negative impact on timing. But if you do, there’s an offset available. You can calibrate this quite easy too, many guides online for that.

And quality wise, running at a lower buffer could potentially add in pops/crackles on the way in, perhaps. But generally the buffer doesn’t affect quality - it’s just the amount of time you’re giving the audio engine to process everything. - It either falls down, or is the same 0’s and 1’s!

Personally, If I’m direct monitoring the buffer isn’t a concern. If i’m software monitoring then no more than 256 for me - and I also turn on CDC (Constrain Delay Compensation) to prevent any excess latency.

Routing out of Cubase into an FX processor can get more complicated though, I’ve had to nudge items after recording in that scenario.

Latency below 20 ms should be small enough to not be distracting/noticeable. I’ve recorded vocals just fine using software monitoring (and even ran some comfort effects like reverb) with a buffer size that translated to ~20ms latency, with no complaints from the vocalist.

Direct Monitoring is an ASIO feature that can use a possible mixer (sometimes very simple) in the audio hardware.
If Cubase is doing DM then it switches these functions on or off in the hardware.
So it should not matter which software or hardware is controlling DM, regarding the latency.
But many interfaces have advanced DM capabilities with their own software, making it possible to use channel processing in the hardware mixer without too much latency.

If DM is turned off, Cubase does the monitoring via it’s audio engine.
Here come the buffers into the equation.

1 Like

Here are my settings:

So how do I enable direct monitoring in Cubase?

I have usually done monitoring for vox recording via the audio interface.

Monitoring via the interface IS direct monitoring. The other way to monitor is by clicking the little speaker icon in audio tracks. It’s best to monitor on the interface when you can.

When you direct monitor, the latency doesn’t matter. Your recorded audio will always be in time with the track.

The latency varies from interface to interface. Meaning I can get shorter latency using my RME interface than my Steinberg interface at the same buffer size. In case of a microphone both input and output latency add to the overall latency.
If the singer wants to hear effects while singing, Direct Monitoring can become a headache.
Most artist I have worked with are not happy with anything over 10ms, personally I prefer max 6ms latency. All depending on what instrument and personal sensitivity.

Not with the UA Apollos, they provide a mixer with own effects.
There are low latency FX (calculated with the DSP’s) available.

The UA Apollos are not Direct Monitoring capable, since the mixer is always active.
So there is no usable switch for DM available with the Apollos.
Btw, the switch is on the page for the “Universal Audio Thunderbolt” settings.

If you want to print this hardware calculated FX (Amp simulations for example)
you should set the insert effects selector to UAD REC, otherwise it is set to UAD MON that indicates that, the send to the DAW is before the insert effects.
But I’m not sure about the UNISON preamps.

You can use the Console to do your monitoring.

1 Like

I always monitor through the MR 816X Console software. I get reverb without printing and this fine. I use an Apollo on the other Mac and use its Console app with some Reverb, Compression as well when needed.

I NEVER EVER EVER monitor through Cubase using the monitor button on the track

1 Like

I have the modest Steinberg UR28M interface, one of the things i really like about it is that i can direct monitor through it with effects on the headphones that are not printed (reverb which i use, compression which i don’t, and i think some other stuff). “Zero latency”.

1 Like

There is not latency with direct monitoring so if you have problems with clicks or pops then set buffers high. I don’t have problems so I leave mine at 64 samples all the time.

1 Like

Yes, this is what I was trying to establish. Thanks. I’ve always done monitoring using Console with the UAD and as you say, as it’s not a feature of the UA Apollos, the Direct Monitoring option in Cubase is not enabled.

The DIRECT MONITORING option works with Steinberg audio interfaces to my knowledge.
Wit that using a Steinberg box you can turn on the MONITOR button on the track and you are STILL using DIRECT MONITORING just like if you use the Steinberg MR Console software. Only diff is you can do it right inside of Cubase in that instance.

I think that’s right, I recall reading that somewhere as well. Unfortunately, that’s not true for all Steinberg audio interfaces. I use a Yamaha TF1 digital mixing board as my audio interface (and Steinberg is owned by Yamaha), but unfortunately, Direct Monitoring isn’t integrated at all into Cubase. The TF1 is a really powerful digital mixer, and I can only imagine the possibility of it being tightly integrated into Cubase. I hope somebody from Steinberg is listening and making this happen, considering both products are ultimately from the same company :slight_smile:

It works with any interface that supports it so not just steinberg

This is a complete mixer, so there is no need for direct monitoring.

Yes, that’s why I wrote that it’s an ASIO feature.
It’s up to the manufacturer to implement it.
My old FF800 has it, since more than 10 years.

1 Like

Care to list some of them?

Well my rme Babyface pro fs is one that supports it for one, which means just about any modern RME interface will.

Good to know , thanks for that :grinning:

Odd that RME is ALWAYS ahead of the rest

1 Like