Audio / playback glitches with Mac

Hello,

I’ve found that when I play back Dorico on my mac’s built-in speakers I get these odd and faint glitch sounds, almost like skipping or tapping sounds - there is a weird slightly swung rhythm to them. They occur randomly, but there seems to be some correlation to the loudness of the audio.

I’ve been investigating this problem for the past few weeks, as it also occurs on Max 8, but not other programmes I have like Ableton 10, nor on other apps like Chrome, Spotify, Music, or Finder Audio Preview. However, the problem sometimes affects audio playing from the usually-not-affected-apps once I open Dorico e.g. I was listening to Spotify and as I opened Dorico the glitches appeared.

I’m using Dorico 3.5 on a MBP 2019 on OSX 10.15.7.

Apple support say I need to take it for investigations as they can’t determine the issue remotely. I have updated the OS and other software but do not want to move to bigsur given the substantial teething problems it has with other software.

There is no mismatch between any of my software and the audio driver settings. Nothing I do in Dorico affects the glitches (e.g. switching off the VST instrument or deleting). I suspect it is either some kind of problem with third party apps or even a fault with the speaker drivers on my machine.

Has anyone else had this problem with Dorico, or does anyone know why the issue might be?

I can try to send a sample but only by recording on my phone as trying to record on my mac gets rid of the problem, somehow.

Any thoughts would be much appreciated - it’s a difficult time to part with your computer, so I’d rather fix this myself if possible rather than hand it in for repairs!

Thanks,
Nick

Could be I’m “special” when it comes to correctly dealing with it. But sure, its a routine thing to have to fuss with at times and it isn’t a computer problem (Unless you count its needing a bigger CPU, etc. to be a computer problem) Anytime your computer can’t catch up with what is being demanded of it, you can find yourself in this position.

The buffer settings in your ASIO drive for Max or Ableton might not be great for Dorico, if you’ve set them for lower latency when recording live. So you might have to make them bigger for certain work in Dorico. And Dorico is working a lot harder than Spotify or whatever since it is also rendering the score, running VSTi, etc.

One test; if something sounds Glitchy in Dorico but the export of the score sounds correct - to me its a clear indication that the computer itself is fine. I had what sounds like the same MPB and vintage - or pretty close. It took some tweaking.

1 Like

As gdball mentions, try increasing the buffer size. To do that go to Edit > Device Setup and then click on the Device Control Panel button. There you can change it then.
And if the problem persists, yes, please make a recording, even just via smartphone. That will give us an impression and we may be able to draw conclusions from the noise we hear.

1 Like

Thanks both!

Thank you gdball for sharing your experience so extensively, and for the suggestion on changing the buffer size. Increasing it to 512 or above seems to get rid of the problem.

Re computer performance - this is a disappointment to me, as I bought this MBP fairly recently with some upgrades specifically to create a robust live performance machine, capable of doing things like live ambisonics plus electronics etc at low latency (hence low buffer size).

Even if this drifts away from Dorico I’d be interested to know people’s perspective on whether you’d expect these kind of glitches on the machine with the below specs. I really wasn’t expecting audio drop outs from playing back solo piano stuff in Dorico at low buffer sizes. Fortunately low latency isn’t a requirement for my Dorico work so I can continue with this fix fine for now, though I’m more bothered about the implications for other programmes like Max. But also, the problem only occurs on the built-in speakers, not with the headphone port, nor external audio interfaces - weird! Good thing I don’t do gigs with my laptop speakers :sweat_smile:

Model Name: MacBook Pro
Model Identifier: MacBookPro16,1
Processor Name: 8-Core Intel Core i9
Processor Speed: 2.3 GHz
Number of Processors: 1
Total Number of Cores: 8
L2 Cache (per Core): 256 KB
L3 Cache: 16 MB
Hyper-Threading Technology: Enabled
Memory: 32 GB

Here is a glorious recording from my phone of playback of a random transcription of Hildegard von Bingen. The problem doesn’t occur on exported audio. Playback template is standard HSSE+HSO (Pro) with one VST instance.

I guess overall I’m just puzzled (presumptuous?) over why this kind of things happens at all - it was never an issue in Sibelius for such small projects on less powerful machines. Shouldn’t it work ‘out of the box’ as is, automatically fitting the system, or is my computer more at fault?

I don’t know if it would help with your issue but have you considered upgrading to Big Sur? I found OS11 to be much more stable than 10.15x and I’m on a 2015 MBPro with only 2 cores and and 8 gigs of ram.

For the moment sadly not, as I’ve seen there are several issues with bigsur and other programmes that I use.

I think it was Paul Walmsley who once told me that its not about load per se, but whether it can do every thing it needs to in the time before it has to render the next audio sample.

At a minimum rate of 44100 samples per second, it has slightly less than 0.023 milliseconds to get it all done between samples with no buffer. Way too fast. With various buffer sizes it will have:

64 (64/44100) = 1.45 ms
128 = 2.9 ms
256 = 5.8 ms
512 = 11.6 ms

At the smaller buffer size, the computer has to work 8 times faster than it does at your setting of 512 - It makes sense to me at these speeds that any small differences in the amount of work required add up quickly. Some (most?) DAW’s actually have multiple buffers of whatever size you picked, so they might effectively have a larger buffer (and latency). In Studio one the number of buffers is configurable, so that might be a difference when you compare Ableton to Max with the same buffer size.

Using an article from Sound on Sound (so that it’s not all my opinion) 12 ms is about what a guitar player is used to given the speed of sound and the distance from their amp, and they say for keyboard players:

“Even on acoustic pianos there’s a delay between your hitting a key and the corresponding hammer hitting the string, so a smallish latency of 6ms ought to be perfectly acceptable to even the fussiest pianists. Famously, Donald Fagen and Walter Becker of Steely Dan claimed to be able to spot 5ms discrepancies in their performances, but the vast majority of musicians are unlikely to worry about 10ms, and many should find a latency of 23ms or more perfectly acceptable with most sounds, especially pads with longer attacks.”

Its the vocalists who can’t tolerate even 3ms sometimes - especially if you are talking a round trip from their singing into the box, processing, and then back out of the box to them again. (latency times 2) Though my interface is supposedly very low latency, its been more trouble than its worth and painful to attempt.

What I do is I run a monitor output from my audio interface to a small live console WITHOUT the vocalist(s). Vocal mics go into the console first, and what they hear is an analog mix of direct vocals and computer, with some rack RX (comfort verb) that only they hear if we’re recording. I have an analog pre with Autotune that I don’t always mention. :slight_smile: And that’s only if its my equipment. At a venue, you’d be using the house system in a similar way anyway…

Keep or sweep man, just my thoughts. I routinely crushed Sibelius too, back in the day. I hear what you are saying about a simple project, but I 've found to my embarrassment that many things can be different from one project to another that are outside the control of Dorico.