|
Post by RealNoob on May 5, 2022 9:17:57 GMT -6
I was recently told by a manufacturer that USB devices on Mac no longer can control the latency involved. Apple did something that creates inherent latency on USB.
I have the Revolution 2X2 which I think sounds better than my Antelope Orion Studio. Mixing on its DA is great but recording has too much latency to be useful. On the other hand, the Antelope device is easy to recording and to create live mixes for streaming which is what I do the most with it these days.
Anyone else experience the USB lag on Mac or can it be solved? so many interfaces are going USB but with the reported issue, it makes the unusable for me.
Hoping someone has some insight.
|
|
|
Post by raddistribution on May 5, 2022 11:38:52 GMT -6
That is correct. Starting with big sur apple closed access to the kernel for driver developers, new drivers can only work in userspace. Apple is pushing developers towards using core audio and only core audio. Several manufacturers have temporary solutions for installing new drivers at a kernel level BUT (theres a big but here) Apple has commented that they are working on clamping down on this workaround, to get all all audio via usb in line with apples requirements. Pretty much use core audio for usb and nothing else. Any drivers written at a user level (the only option going forward for usb) are on top of core audio and can only add latency not subtract from stock core audio levels.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 5, 2022 13:30:13 GMT -6
That is correct. Starting with big sur apple closed access to the kernel for driver developers, new drivers can only work in userspace. Apple is pushing developers towards using core audio and only core audio. Several manufacturers have temporary solutions for installing new drivers at a kernel level BUT (theres a big but here) Apple has commented that they are working on clamping down on this workaround, to get all all audio via usb in line with apples requirements. Pretty much use core audio for usb and nothing else. Any drivers written at a user level (the only option going forward for usb) are on top of core audio and can only add latency not subtract from stock core audio levels. So there is no more integer mode? SPL Crimson sounded better with the HAL drivers than the CoreAudio
|
|
|
Post by RealNoob on May 5, 2022 14:31:14 GMT -6
So, I appreciate you, Seth and love mixing on my R2X2 as mentioned.
I am asking and should have framed my question better to see what's happening with other devices like the Apogee Symphony Desktop, etc. I didn't want to hack that thread. Are you guys finding work arounds for your USB interfaces on Mac?
|
|
|
Post by raddistribution on May 5, 2022 14:38:51 GMT -6
So, I appreciate you, Seth and love mixing on my R2X2 as mentioned. I am asking and should have framed my question better to see what's happening with other devices like the Apogee Symphony Desktop, etc. I didn't want to hack that thread. Are you guys finding work arounds for your USB interfaces on Mac? That interface will be exactly the same as they both use core audio. Be aware that windows is a whole other story. Each company is free to develop and use their own drivers at a much deeper level than is now allowed through macOS.
|
|
ericn
Temp
Balance Engineer
Posts: 14,817
|
Post by ericn on May 5, 2022 14:45:26 GMT -6
Wonder if this means we are going to see Apple develop another “ exclusive like they did with apogee with the original ensemble”?
I mean they not let outside developers play with the Kernel, but there own guys are another matter. Plus they could really use a nice interface they can sell themselves.
|
|
|
Post by nudwig on May 5, 2022 14:58:03 GMT -6
There's a thread at the purple place called 'Does Apple's Core Audio resample AD/DA signal'. I found it an interesting read, definitely convinced me to stay on my HDX card.
|
|
|
Post by svart on May 5, 2022 15:44:11 GMT -6
|
|
|
Post by kcatthedog on May 6, 2022 4:02:57 GMT -6
Haven’t noticed any problems with my m1 and the desktop. It comes with the usb cable and there are the appropriate ports on the m1.
|
|
|
Post by sirthought on May 6, 2022 4:44:19 GMT -6
Is the monitoring ITB when tracking?
Or is it like UAD or others that have you monitor through their interface application so latency is lower?
I haven't really researched the recent USB options for latency, nor the BLA stuff.
|
|
|
Post by kcatthedog on May 6, 2022 4:51:47 GMT -6
Still reviewing the manual, but I believe you can do both as you can have the apogee plugs running on it or in your daw.
I don’t use plugs for tracking though so am not concerned: just use my ob and hit the converters.
|
|
|
Post by sirthought on May 6, 2022 6:54:50 GMT -6
I thought most Apogee stuff was near zero latency, like UAD.
My question was actually directed at the OP and his USB interface from BLA.
I guess another question for Seth or anyone else who might happen to know...Why would a manufacturer write Mac USB drivers that aren't using Core Audio? I assumed all drivers for Mac would be, but maybe prior to this change there were ways of getting better performance?
|
|
|
Post by raddistribution on May 6, 2022 7:27:42 GMT -6
I thought most Apogee stuff was near zero latency, like UAD. My question was actually directed at the OP and his USB interface from BLA. I guess another question for Seth or anyone else who might happen to know...Why would a manufacturer write Mac USB drivers that aren't using Core Audio? I assumed all drivers for Mac would be, but maybe prior to this change there were ways of getting better performance? prior to this you potentially could get better performance by writing your own drivers at the kernel level. I believe this change actually affects any usb driver that potentially tries to interact with your Mac. My bet, Apple's intention was to increase stability through standardization. if similar devices always communicate with the OS in the exact same way, system performance technically can be controlled (for the better or worse, depending who you are). Usb audio interfaces is probably a TINY drop in the bucket when it comes to usb devices used worldwide and sacrificing slight performance for standardized control and performance worldwide was once of the tradeoffs. This closed system concept has kind of been Apple's thing for almost half a century now. That being said, for I'd say 99% of people and uses, core audio is a kickass solution. We have drivers from quite literally every single device made, installed on our test systems from doing mods ( and we have seen some companies with windows drivers with worse performance than stock Mac core audio. All I can really say is I'm glad that we were already using core audio... I would hate to be a developer with the though in the back of my mind that sooner or later Apple is going to patch a backdoor that would affect my current system's performance.
|
|
|
Post by sirthought on May 6, 2022 8:22:31 GMT -6
Thanks for the response Seth. I was actually wondering about this interface if I ever needed a non-racked option for throwing in my bag.
Do you think the change has negatively impacted the 2x2's tracking latency? Or is it in the same ballpark as when the product launched?
|
|
|
Post by raddistribution on May 6, 2022 10:15:29 GMT -6
It will have zero affect whatsoever on our interface or any other interface that was already using core audio. It only affect Mac users of usb interfaces where the manufacturer was writing their own Mac drivers for usb. There have been workarounds but apple is clamping down on these as they bypass things that apple believes are there for a reason.
Pretty much there are four ways forward for interface manufacturers. 1. use core audio which is the best option for the vast majority of users and manufacturers. 2. phase out your own drivers in favor of core audio (technically not a bad option but it most likely will affect performance in one way or another). 3. keep writing your own drivers and have them operate on top of core audio (not the best option, as that can only have equal or negative performance to core audio) 4. stop using usb and switch over to an option that accesses the kernel similar to how pcie does. (this option is great except it is shockingly expensive. Most users in theory love this until they see the development costs reflecting in the price of the device.)
|
|
|
Post by sirthought on May 6, 2022 14:57:15 GMT -6
Thanks. We're up to USB 4 now and most devices are still USB 2. I wonder if latency can improve with them holding out with that?
RME has bragged about it, and their reputation sticks for latency. I wonder if they use Core Audio or will need to change their drivers?
|
|
|
Post by RealNoob on May 7, 2022 17:46:25 GMT -6
Mentioning again just for clarity, despite the USB issue for me with recording, I love the DA on the Revolution and use it exclusively for mixing.
|
|
|
Post by sirthought on May 7, 2022 19:10:42 GMT -6
Mentioning again just for clarity, despite the USB issue for me with recording, I love the DA on the Revolution and use it exclusively for mixing. What are you tracking that feels unacceptable?
|
|
|
Post by the other mark williams on May 8, 2022 22:47:27 GMT -6
Mentioning again just for clarity, despite the USB issue for me with recording, I love the DA on the Revolution and use it exclusively for mixing. Rob, apologies if I’m missing something obvious here, but I’m having trouble understanding your latency issue. It looks like the BLA Revolution eschews the “software DSP-run onboard mixer” approach in favor of the “playback/input mixer knob” approach. With either approach, latency of live audio inputs shouldn’t be a problem, though. Or is your problem arising when playing software instruments in the DAW?
|
|
|
Post by sirthought on May 9, 2022 2:41:29 GMT -6
It's not just Rob experiencing frustration with latency on this. According to Julian Krause's review, the latency on this is thing is surprisingly bad. Not out to bash anyone. Julian thinks it is odd given the quality components they've chosen.
|
|
|
Post by svart on May 9, 2022 7:57:57 GMT -6
Makes me wonder if when Apple went ARM they didn't account for something in the USB section of the ARM that causes issues with latency and they've simply hidden it "for your safety" like they've done in the past with other fuckups?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 9, 2022 12:11:57 GMT -6
It's not just Rob experiencing frustration with latency on this. According to Julian Krause's review, the latency on this is thing is surprisingly bad. Not out to bash anyone. Julian thinks it is odd given the quality components they've chosen. Interface manufacturers often prioritize latency over sound quality. Even in the ones that sound bad. Focusrite and RME did. They chose worse sounding chips with less latency and used to use (probably still do) minimum phase anti-alias filters on the analog outputs. They only spec out better than many older two channel converters because they use more modern converter chips so a Babyface Pro specs out better on the back of the box than a Dangerous Convert 2 but of course one sounds like constrained tinny shit and the other’s awesome. I don’t know what’s in the Black Lion Revolution but I doubt it’s junk.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 9, 2022 12:19:13 GMT -6
Makes me wonder if when Apple went ARM they didn't account for something in the USB section of the ARM that causes issues with latency and they've simply hidden it "for your safety" like they've done in the past with other fuckups? An internal Usb safety buffer to compensate for bad pc builds with poor component drivers (including the audio interface ! ) is real. Usually it’s motherboard, gpu, or networking. Often macs needs the Wi-Fi disabled to get really good stable, low rtl over usb. PCs without Intel wifi cards or Intel chips on the mobo tend to be horrible for recording ime. MOTU did this on Windows over USB in a driver revision to accommodate off the shelf laptops and gamer desktops. So my latency literally went up overnight and I couldn’t even run more plugs in a bigger buffer. But then I could plug it into random laptops from Best Buy it wouldn’t work on before at all. My Apogee Symphony Desktop let’s me disable the safety buffer for recording.
|
|
|
Post by svart on May 9, 2022 12:47:40 GMT -6
Makes me wonder if when Apple went ARM they didn't account for something in the USB section of the ARM that causes issues with latency and they've simply hidden it "for your safety" like they've done in the past with other fuckups? An internal Usb safety buffer to compensate for bad pc builds with poor component drivers (including the audio interface ! ) is real. Usually it’s motherboard, gpu, or networking. Often macs needs the Wi-Fi disabled to get really good stable, low rtl over usb. PCs without Intel wifi cards or Intel chips on the mobo tend to be horrible for recording ime. MOTU did this on Windows over USB in a driver revision to accommodate off the shelf laptops and gamer desktops. So my latency literally went up overnight and I couldn’t even run more plugs in a bigger buffer. But then I could plug it into random laptops from Best Buy it wouldn’t work on before at all. My Apogee Symphony Desktop let’s me disable the safety buffer for recording. I meant more at chip die level. The "IC designers" are generally teams that work on sections of the chips. For ARM, it's even more so since you can almost cut-n-paste licensable cores into your chip. USB being asynchronous as well as non-real-time is particularly susceptible to being seen as the red-headed stepchild. It's doubtful that the *USB core* designers understand DPC/roundtrip latency as important metrics since the most likely most-important metric is total transfer speed. Sure they can burst transfer huge amounts of data quickly, but it might take quite a bit of time to buffer the data into RAM or FIFO and since the USB cores are using this newfangled common RAM bus technology, I'm sure the main CPU cores have priority over the bus.
|
|
|
Post by RealNoob on May 27, 2022 19:28:27 GMT -6
OK, this is solved. For clarity, on my 2018 MBP which was a fast machine, there must have been some kind of driver or other issue causing problems. No matter what settings I used, if I tried to monitor inputs after hitting the DAW, it was significantly delayed and impossible to keep time with - like playing with a really slow drummer, when you match, he slows again - lol. This is important because when multi tracking, you have to have some playback to record with.
Also, it had been said that Apple did something that was the root cause with this and to me, that would have meant every machine and every interface so I couldn't understand how people where getting on OK with the Symphony Desktop.
I just got my M1 Max, 64G, 2T, 10 core machine and got everything loaded. I took a sec this evening and it is problem free with the USB interface. My DAW was defaulted to 512 and it was fine. I knocked it down to 64 and it was seamless. dialing from input to return on the interface was imperceivable.
... all this said, it could have been a problem isolated to my 2018 machine. I am relieved and happy with my BLA Revolution 2X2 once again.
|
|