Been using GP in a live environment without issues for several months now. However I’ve hit my first hurdle this week. For a gig this weekend I’ve had to simplify my rig. Instead of using my normal MOTU Ultralite interface I’m using the inbuilt audio interface on my Yamaha MOXF8. This gives me a very simple setup - 2 keyboards via USB straight into the Mac with all audio from the L/R of the MOX (carried by one of the USB connections).
The gig is pretty straightforward - Hammond, piano, wurli, clavinet, brass etc. with only a three or four songs requiring more complicated programming. I’ve used these racks for months on GP with this band but am now getting serious popping and crackling. This is happening mainly on AcousticSamples B5, which is notoriously CPU heavy, but also on other racks with very basic sampled instruments. This is running with a latency of 128 samples, 44.1kHz (MOX interface is, I believe, only capable of 44.1kHz), same as I run of the MOTU.
First thought was that the Yamaha audio interface in the MOX was not up to the job so I recreated the gig in MainStage using the same instruments and splits etc. Works perfectly at the same latency rate used in GP. No crackles, breakup etc.
I’m going to have to use MainStage for tonight but really need the patch persist functionality of GP.
For reference the plugins I’m using are AcousticSamples B5 running under UVI player, Kontakt (Grandeur, Hein Brass,Wurl-E, Soniccouture Clav) Omnisphere 2, and the Air plugins. MacBook Pro is i7 quad, 16Gb RAM, 1Tb SSD.
Several immediate questions come to mind.
- What CPU value does GP display on the top right?
- Are you certain the noises are due to CPU “struggles” rather than audio that’s too loud and causing distortion? (I’ve been bitten by that one a few times)
- If you change the sample buffer size to 256, do you still have the problem? What is the CPU value at that sample buffer size?
1 - The CPU never seems to go above 40% (and that’s only with the B5) with a normal value of around 15%
2 - The volumes are not high. I’ve duplicated the same volume in MS.
3 - Changing the buffer size to 256 lessens the noise quite a lot but it’s still there on initial loading of the rack and for the first few seconds of playing. CPU sits around 7%.
For buffer size of 128 I have noticed that with repeated playing of the same rack, the crackles lessen or disappear. When changing to a new rack for the first few seconds I get a crackle, so it appears to be with the initial loading of the sounds or in some cases with the streaming (e.g. UVI). I only have patch persist on two racks.
I haven’t got time today but I will try exactly the same gig but with the MOTU interface to see if that makes a significant difference.
Yes, it would be useful to know how the MOTU interface behaves. Thanks
Are you using the “Predictive loading” feature? Also - could you please send us your .gig file? Just open a support ticket here https://www.gigperformer.com/support-ticket-system.html and attach the file to it.
Thanks.
One more question if you don’t mind… You mention that the audio device settings in GP and MainStage are the same. Could you open the Advanced audio Settings in MainStage and let us know if the I/O Safety Buffer option is checked and one the same screen - what is the output latency displayed?
Hi, Lee -
Any possibility that you might be doing sample rate conversion on the fly from 44.1kHz (current audio interface settings) to 48kHz (GP still config’d as such [re original interface])?
Mike
Apologies for not posting sooner, been a bit busy.
Yes the Safety Buffer option is on, and will check the latency today. I did push the latency down from its default, and even used 64 sample buffer rate although that did seem to push the CPU.
Mike, I’ll have a look at that. Will have time to setup both configurations this week and try and solve this.
Although it’s an extra piece of equipment, I’m looking at buying a RME Babyface Pro. It will still get rid of the rack but should give much lower latency than the Yamaha drivers.
The quest for the perfect live rig continues 
I’ve been an RME fan for years, their drivers are great. That said, you might want to look at the new Presonus Quantum. I know nothing about it other than the review I saw in Sound on Sound magazine but supposedly their claim is that the device’s latency is way lower than anyone else can handle.
The “Safety Buffer” option in MainStage is explained as adding an additional buffer of the same size to your output buffer.
This essentially means that you are running at almost twice the buffer size. If you look at the latency (same screen) note the output latency. It will be usually reported as at least twice the latency displayed in GP with same settings.
This essentially means that you should be actually able to use a 512 sample buffer with GP and not feel the difference. You can try that buffer size and see how it goes.
I’ve written an article about the latency, buffer size and sample rate in general. You can read it here https://www.gigperformer.com/audio-latency-buffer-size-and-sample-rate-explained
In general - 1ms of latency adds 1ft of distance between you and your monitor speaker. Most important guide though is your ears and feeling. If you cannot really distinguish between 256/512 buffer size - then you should really use the 512 as it will be less taxing on your CPU