I’m asking your pro advice here for this difficult mater.
I’m running C3 perfectly, it’s directly attached to my mk3 soundcard and it’s the only usb connection on my W10 PC. This works great for my rock bands. So my question is beyond C3.
Now, for my full show bands (multitrack in Reaper DAW, with extra video, DMX lights over usb) I keep on getting problems.
And I am getting somehow desperate to get it fixed, as some important big gigs are coming up.
I recently suffered from windows defragmentation shedule messing the disk IO up (luckely on a rehearsal but still heart attack). But yesterday evening on the gig again some ghosts turned up.
suddenly my midi keyboard was +1 tuned. After resetting some panels and wiggling some controllers it was gone. First time I got this and dunno why.
USB connections drop sometimes. I’ll probably replace some cables but I think it might be a W10 problem, after updating, the initial problem was fixed, but it’s still unreliable.
I keep on getting a problem with the mk3 soundcard (first runned on usb port 1, then that kinda crashed and now only usb3 on the left works for it) Sometimes the card is only half recognised. I’ll try switching to Roland octacapture maybe.
I have the cpu spikes under control, but still glitches happen with Avenger synth, I’m not sure if it’s disk IO or cpu related (cpu doesn’t go over 12%)
video works ok, low quality mp4 and doesn’t give me high cpu spikes or disk IO.
So yesterday evening was again a concert with lots of stress, scary looking at the CPU meter gauge and hoping it all keeps on working. Brrrrrrrrrrrrrrrrrrr
Funny thing is you can prevent and test as much as you want at home, still on stage it seems the computer senses the live gig stress and funny things happen (or is it me )
So I was wondering:
should I switch to a Mac system for more performance and reliability? We all know W10 (I even imagined reinstalling W8.1 if that would be possible lol
I have a recent new laptop W10 latest update (updates and background processes locked with UWT4 or manually disabled) SSD drive 2.7GHz quad (octa threading) core with 16 Gig ram
I see Mac users also suffer software probs, so that won’t probably justify the budget increase… ? Besides switching to a mac would complicate my workflow, I now create stuff on my home pc and sync with my laptop pc. A Mac system would be an investment also, but if it solves my headaches worth to do.
I should probably invest in a RME soundcard, but that’s way over my budget. Is MOTU mk3 or Roland Octacapture really not secure for this situations?
should I add an external powered usb hub to compensatie voltage drops?
would a backup power connector secure external voltage drops (critical for usb)?
I know there are guys over here with years experiences for big live gigs.
What are your best practices to be 90% sure everything is stable on stage during gigs?
(OK, only kidding! I am also on the MOTU Digital Performer forum (Motunation) and the majority of users there are Mac folks, and Apple keeps throwing them all curve-balls as well. That investment will only introduce new problems you will need to learn to compensate for. And being a PC user for some time, you likely will start to feel hobbled by the closed system Apple forces you into.)
All you really can do is simplify the rig as much as possible, and I would also separate the video from the audio to two computers. Video (and video cards) are notorious for messing with audio timing.
Ever since I went into the Device Manager and disabled the “Motu Audio Wave for 64 bit” feeble attempt at making the Mk3 “Windows Audio capable” I’ve very seldom had any issues with the Mk3. It does not like brownouts, so a voltage regulator is likely a prudent investment. Tripplite sells one for about 200 USD that is a switching-tap transformer type - makes a mechanical clicking noise when switching but that would not matter on stage. (Haven’t used that - a friend has one and likes it.) Otherwise, spend a few times that amount for a Furman.
As you see, I only have one device to disable. This is a MOTU Ultralite mk3 hybrid I have here, but the same is true for my MOTU 828 mk3 hybrid on the main studio machine. (This is my video streaming and mastering machine.)
But the gist of it is that yes, you do not want MOTU and Windows Audio intermingling, as the drivers change the INTERNAL sync setting of the interface whenever it detects a sample rate change in the content of the media being played back, which “grinds gears” and is a very clunky approach. They would have been better off leaving Windows Audio to employ a sample-rate-conversion go-between, but they either got lazy or thought that pros would WANT the internal sample rate changed automatically to match the content. In real-life, this can wreak havoc as even two songs in a playlist can often be one at 44100 and the next at 48000, kicking in that awkward change to the interface and sometimes conflicting with the very next audio that comes into it. Definitely have Windows Audio’s “Sounds” turned off anyway!
Well, some of that could be from attempting too low a buffer setting also.
I’m happy with 512 for live performance, but I use in-ear monitors so there is no “air-time” lag as would exist if I monitored over speakers.
Also there’s a problem with avenger plugin somehow.
On my motu and octa soundcard it gives cracks from latency.
But not on the wasapi!
Very frustrating and thinking to buy falcon to switch and start all over.
I think the key thing is to ditch everything that doesn’t need to be running. Do not have defragging on a schedule, and see tip below: Move to an SSD and it is not necessary. Disable virus checkers and firewalls during a gig (when you are not connected to the the internet!) and have the computer in flight mode when you do not need to be accessing the internet.
Go through Brad’s excellent Glitch Free optimisation guide.
Get as big an SSD and as much RAM as you can afford to keep load times short, and load as much as possible into RAM.
Investigate optimisers such as Bitstream
Play with sample buffer size. Make it as large as possible before you notice latency and then back off a little.
As I mentioned in previous thread, consider several computers if you wish to run video as well. I would be nervous putting a video task on the main live host. I gave a use case where I had SCS11 running on a separate laptop, and both laptops connected via RTPMIDI and a single CAT5 lead.
If you haven’t made the switch to SSDs, they are much more affordable than ever. And they don’t need to be defragged (if fact, they should NEVER be defragged). Technical reason: the NAND chips that make up the SSDs have a finite number of writes they can perform. Since there is no seek (positioning the read/write heads over the correct cylinder), and no rotational latency (waiting for the right sector to spin under the read/write heads), there is no difference in performance over a fragmented file and a nicely sequential file.