Ok, so if my buffers are at maximum already, then the ASIO Guard is probably of no help to me? That’s kind of what I was already guessing without having taken the time to investigate it. I’m not really fussed about latency most of the time as a very high proportion of my work involves mixing.
Without derailing this thread can you explain why so many tracks are used? Some huge orchestral arrangement? I’m just curious. Maybe I’m old fashioned, but I rarely exceed 150.
Take all of the 1/2 dozen or so libraries (strings, brass, winds, choir) you score with, tally up all the individual patches you could easily have upwards 500 - create a MIDI track and an expression map (possibly) for each patch. if you’re insane like me and you like to have individual outputs for each patch, add 500 more for tracks. Or, if you’re more practical 500/16 outs.
I also have templates with > 1k tracks. Hence the reason i figured I would explain that the “improvement” in performance is really not what you think.
The new system is really slick, and it is a really cool feature. So,don’t take my posts as negative. But, it’s important to note where the new system DOESN’T help. And that is when you ASIO buffers are already set high to support a large project that is already maxing your system. The actual “TOTAL” performance (for me), is actually slightly less than C6 and C7. But, general performance of a project that is not maxed out is much smoother and easier to work with. And, the not having to switch buffer settings so often is a great thing.
Isn’t the idea that you can now monitor at a lower latency than previously possible because ASIO Guard will bump up the latency on your non-live monitored tracks?
Yes, as I said earlier, Low, Medium, High determine how much more buffers/latency the project runs at than what is actually set on your interface. So the “perceived” performance can seem huge … yet, if you just went to your interface and set the buffers to 2048, you would get the same performance. There’s probably a little more to it, but as a user this is the general outcome.
The cool trick is that when you monitor a track, it switches to the base latency you have set. It does that very well, no clicks, pops, skips, jumps, stutters etc… so far. So you set your card to a low latency, but the project is actually running at a high latency.
Yes, exactly that.
Impatient to try it, I just had a go. Win 7 64 bit, i7 4930, 64 gb RAM. In a project with a large VE Pro orchestral template (about 30gb worth of samples in VE Pro, 500 tracks or so but of course only a fraction in use at once, various soft synths, VIs and effects), running at a 48 buffer of a 4ms round trip (the RME minimum) and asioguard to full had an average load of 50% and peaks around 20%.
I think some are looking at this wrong. Sure, if you were working at massive latencies and continue to want to do so, then there’s nothing to shout about (though from my briefest of tests, it’s hard to see me maxing out the CPU with pretty much anything I own in any quantity). However, most of us would plug in a mic or play an instrument, so will need some real time performance too. And that’s the huge gain. There’s simply appears to no longer be any practical need to trade off. That’s a pretty huge deal.
My only gripe is that the new font means that the thinnest track height available is now considerably wider. This definitely needs addressing if it’s the case (could be wrong), I’ll be experimenting more later.
I think the point is, you get the best of both situations, but not an actual performance increase. Also, there is no point in having your buffers set high anymore (with some exceptions). The system will do that for you … which is nice. And as you noted for those of us who want to record into large templates, it’s cool that I don’t have to keep resetting my buffers and temporarily disabling stuff. So, I’m not be disparaging the feature at all, just trying to help people understand what’s going on so they can make informed optimization choices. For me, Medium at 5ms seems to be the sweet spot for my large templates. I don’t get any performance benefit from High or if I do it is minute. However, as I’ve said once I start loading on UAD plugs and reverbs and such, my ACTUAL total load before death is slightly lower than C6/C7.
I’ve read that also, but can’t get my head around it …
If one is recording at low latency so they can monitor reasonably well without disconcerting delays … but the backing track is at high latency … I have a hard time picturing how one can record well, unless you can perform to a backing track that plays long after you hit your notes …?
(Isn’t that the reason that “Constrain Delay Compensation” exists … to avoid added latency of some tracks that have VSTi’s with long latency (so it’s not crazy trying to record)?
Help! 
This is frequently a DPI scaling problem. If you only have this issue in a single application, you can often fix it by locating the EXE in Windows explorer, looking at the file’s properties, and checking “Disable display scaling” on the Compatibility tab. If it seems to be happening on multiple programs, you can usually fix it by going to the DPI Scaling control panel (easiest way to get to it is open a command prompt and type: dpiscaling) and tweaking your settings there. You might have to restart Windows after using the control panel in order to see the effects.
Sometimes the fuzzy type has to do with ClearType text, which you can also adjust from that Dpiscaling control panel (click “Adjust ClearType text”).
Yes, that used to be kind of what the Delay Compensation did … it was a way to record at low latency in a project that had high latency plugins.
So, 20 years later …
There is enough memory and CPU available for the project to play ahead. Imagine an invisible cursor that has already played your song and loaded that into memory. When you click play, it looks like it is playing what is under the cursor, when in reality it is streaming stuff it wrote into memory awhile ago. To you it looks like it is just now playing ( the project cursor ), but to the project it is already playing 20ms or whatever ahead of you… you just don’t hear it yet.
So, you click the monitor button … it simply subtracts the extra buffers and adds to the stream at the correct location. I’m oversimplifying a bit I’m sure, but it isn’t too far off of this. It also does take system resources to do the “look ahead” processing.
Instead of Constrain Delay removing the plugins from the path, the system now just reads the output into memory before YOUR cursor gets to the point you need to hear it.
Thanks, Jalcide, I’ll have to get my head around this. Does remind me of the movie Interstellar somehow 
The new font is excessively large and bold. The font style itself looks like something from a comic book rather than what you would find in professional software.
Yeah, it’s a nasty font alright. I was wrong about the track height though, it’s just an optical illusion caused by that gaudy font. I remember they played around with the fonts a lot with 7.0.x til they got a good balance, hope this one gets revisted quick as from my early playing around, it’s pretty much the only thing I can fault. Absolutely terrific otherwise, been rock solid for me in early testing. I like the new mixer, I can get a lot more useful info on the screen at once now and it’s clearer too (I don’t use the strip, mind).
When you say Constrain Delay do you mean Constrain Delay Compensation (the orange button?)
Thanks for that explanation, I think it’s possible I might be able to understand it, I’m not sure. I’ll try again later when I’m not so tired 
It does sound like standing in a laundry basket and picking yourself up with the handles…
From what JMCecil has excellently explained, I take it that if the playback tracks are not loaded with a lot of CPU processing – as a result of manually optimising for low record latencies – I would not expect much benefit either, as there is nothing to compensate for?
Our stuff is fairly simple, and when adding tracks, those being played back don’t need any processing (as opposed to having to turn off processing for the lower latency), so I would expect NO advantage from the new AsioGuard.
If the process is as JMCecil explains, I suspect that AsioGuard (AG) will give some benefit to those using samples at higher project sample rates, as CPU is used to do the real-time sample rate conversion from the native 44.1k standard of most libraries. If so, that means those recording (audio or VSTis) to samples playing back will be able to use lower latencies. However, the original AG may have already catered for this processing, so C8’s update may provide no benefit or maybe adds some extra.
That’s a real interesting question Patanjali. I hadn’t thought about the repercussions of the process of read ahead sample loading. Obviously, in the original ASIO guard sample engines were not covered. Now they are. Definitely worth exploring as I use a crap load of VSL libraries. I hadn’t considered the repercussion of the on the fly conversion that has to be done ahead of time. That is a different area than what I am referring to in relation to the buffers settings affecting the “perceived” project load.
The first thing I do in any Cubase install since the introduction of ASIO guard is turning off ASIO guard. I do not use any software monitoring so keeping my buffers at 1024 samples is fine for me. I do this because the very first version of ASIO guard worked very pooly on my system for some reason and I haven’t tried to give it another chance since (if it aint broke don’t fix it).
But yeah, when I first ran Cubase 8 (and noticed that the UI was not crashing but was indeed the “improved” version) I noticed that not only did it boot faster, but projects loaded faster and ran smoother. The performance in Cubase 8 that I’m getting so far is noticeably better than 7.5. I have yet to run into a non UI related issue, though I’m sure I will, as I would with any new software release.
It is not done ahead of time, but rather delay compensated to align with the rest of the project. That is if you’re speaking of recorded tracks
Almost every issue I was experiencing after updating from 7.5 to Cubase Pro 8 disappeared by turning ON ASIO Guard.
Using Windows 7 64bit Ultimate w/Aero
The only issue that remains is some weird GUI behavior with Native Instruments’ MASSIVE.
Though the function of the synth is not affected.