Great info on DP 9.13

[nk_e]

Hey.

I came across this on the FB forum. I don't think I'm breaking any protocols by pasting it here, but if I am please delete.


"Hello everyone!

As most of you know, the 9.13 update is now live at motu.com. This update mostly resolved bugs which included issues with the Hardware Insert plug-in, Melodyne, Softube Console 1, RME drivers and VEP MAS to name a few. 9.13 is also the first version of DP available with a German translation (tell your friends in Europe!). We are working on new features and the next update will include a completely new and improved pitch and time stretching engine.

If you've been experiencing problems since updating to 9.1, please read the following. It's likely that this information will help you.

With the DP 9.1 update MOTU introduced two new major updates to the audio engine. Pre-gen is a function that renders audio tracks and virtual instruments ahead of time. Testing has shown this provides up to a tenfold increase in how many virtual instruments and effects can be run within DP.

The second major change was how DP handles buffers and latency with audio patch through and live VI triggering. At the same buffer size DP9.1x will now have half the latency as with previous versions.

For the majority of DP users these changes have provided dramatic increases in efficiency in their work flow. For a small group of DP users there have been problems with these updates.

The first thing to understand is that the pre-gen feature is completely separate from the latency updates. If you are experiencing a problem with one of these features you may not be having a problem with the other. Because MOTU promoted the new pre-gen efficiency, this feature generally gets the blame, even if that is not the actual cause.

The latency update cuts latency in half but that is still at the expense of CPU. A lower buffer setting means lower latency, but also at the expense of higher CPU hit. When upgrading to DP9.1x the DP user may notice no other difference besides faster patch through and triggering. However if the computer was working at close to maximum capacity under a previous version of DP (9.02 or below), upgrading to DP 9.1x will keep the same buffer setting but double the load on the CPU. This can cause performance problems. The solution here is raise the Buffer Size. For example a Buffer Size of 128 in DP8 is a real world latency of 256. In DP9.1 a Buffer Size of 128 is a real world latency of 128. DP9.1 is now as fast as Logic.

Pre-gen works with recorded audio tracks and recorded MIDI parts triggering virtual instruments. Pre-gen can not work on a real time audio signal or record-enabled virtual instrument. Essentially pre-gen is “recording ahead” into RAM to lower CPU load. Some plug-ins work outside DP and therefore can not have the benefit of pre-gen. Examples of this include VEP and UAD plug-ins. Also, because the plug-in is rendering ahead, the graphic response (aka level meters) of a pre-gen plug-in will not match what is happening at the time in the track. For these reasons we have included the ability to disable pre rendering when a plug-in window is open. Choose "Run Open Instances in Real-time" in any plug-in's mini menu to force into real-time. For plug-ins that run outside DP, pre-gen is always disabled.

Suggestions for users having issues with 9.1.

1.) Experiment with various Buffer Size settings. Remember that a higher Buffer Size in 9.1 doesn't necessarily mean slower performance when compared with an earlier version of DP. The newer 9.1 audio engine is faster and therefore can be more taxing on certain computers.

2.) Make sure that Virtual Instruments are not record enabled when they don't need to be. Record enabled MIDI tracks are always patched in real time. It's not possible for pre-gen to work on any record enabled track (there's no audio information for DP to process pre real time). If you've got a heavy plug-in (Omnisphere for example) and it's record enabled during playback, you're not going to get the processing benefits of pre-gen. Also make sure that any plug-in windows set to run in real time are closed when not needed.

3.) Keep open plug-in windows running in pre-gen. If there is no visual meter, there is no reason to run the plug-in in real-time.

4.) Consider loading your template into a new project created in 9.1x. Use the File menu, Load, select the DP project file and then choose the Chunk to load. Doing this has resolved issues for several users.

5.) Several third party plug-ins did not play nice with 9.1 upon initial release. We have resolved issues with the third party developers for all problems that we know of. Make sure you are on the latest updates for your OS version. Also, consider updating to OSX 10.11 or 10.12. DP9.13 is supported from 10.6.8 to the latest update of 10.12. Certain users have reported improved performance when updating to 10.11 or 10.12.

If the 9.1 update has given you problems, give 9.13 a try and let us know how it goes.

Thanks to you all!
Andrew and the DP team"

[terrybritton]

Thanks for giving this its own thread!
Terry

[Michael Canavan]

Good to hear about third party developers.

Makes me think about trying higher buffer settings than the normal 128, to see if play through Amplitube 4 is still possible.

One thing I noticed the other day, (and it's not at all specific to DP, other DAWs I use have the same behavior or worse), seeing all the threads for the 12 core machine here all of them are using maybe 12% at most for everything on the Mac Pro, but in DP the CPU is hammered at 60%. I've always wondered about that. Doing stress tests it seems that most DAWs start pushing at low numbers, (80% etc.) but can grind out sometimes up to twice as much when they're seemingly already redlined. Just to be fair here, some DAWs will max out a single CPU even when plug ins are spread across multiple tracks. Wondering out loud if higher buffer setting changes this CPU sensitive behavior a bit, wouldn't be that surprised if it did.

Anyway, all this makes me think about doing a stress test on DP tomorrow. I'm literally saving pennies for change machines at grocery stores to get a copy of LogicX so I can import some older songs etc. (I'm about $20 away from enough iTunes credit! lol!) It will be interesting to see how they compare CPU wise....

[bayswater]

Anyway, all this makes me think about doing a stress test on DP tomorrow. I'm literally saving pennies for change machines at grocery stores to get a copy of LogicX so I can import some older songs etc. (I'm about $20 away from enough iTunes credit! lol!) It will be interesting to see how they compare CPU wise....

I plan to do this too. I did one using a couple of hundred audio tracks with loads of plugins and a couple of dozen VIs on my ancient coal powered iMac, and DP 8 and Logic 10 were about the same with a slight edge for Logic on CPU use in Activity Monitor.

[David Polich]

My takeaway from this is that I might as well run at my usual buffer settings (256 when triggering
VI's, 512 when recording audio, 1024 when mixing)...setting buffer setting slower does increase the
CPU hit. The only real improvement I've seen is that I get
less latency at 512 now, so I can probably record into VI's at that buffer setting with the same negligible latency as I used to get at 256.

So, hmmmm. The Pre-Gen is more efficient, and decreases latency, but does so at the expense of
increased CPU. That equation cancels out to a null, for me. Maybe I need to try going to Sierra, although Yosemite otherwise is so stable for me that I am loathe to do so.

[cuttime]

Thanks for posting this. My only concern is that MOTU seems to be more engaged at Face Book than they are here. I'm very disappointed that I had to get this info second hand. Those of us who eschew FB and its policies seem at a distinct disadvantage at this point.

[bayswater]

The Pre-Gen is more efficient, and decreases latency, but does so at the expense of increased CPU.

That's not what I read. It says:

"the pre-gen feature is completely separate from the latency updates. The latency update cuts latency in half"

The PreGen lets you run more VIs etc, with the same CPU. It does not reduce latency. The latency update reduces latency by half, but lower latencies still take more CPU, as they always have. The implication is that a latency of e.g. 10 msec now uses the same CPU, all other things equal, as a latency of 20 msec took before the update.

[bayswater]

Thanks for posting this. My only concern is that MOTU seems to be more engaged at Face Book than they are here. I'm very disappointed that I had to get this info second hand. Those of us who eschew FB and its policies seem at a distinct disadvantage at this point.

Ditto on avoiding Facebook, but you don't need an account to look at the MOTU Facebook page. At some point an annoying login nag will cover the screen, but you remove most of it, and then can click on specific items and see them without any interference.

[Gravity Jim]

MOTU spends more time at their Facebook page than they do here because:

1. Facebook is a huge part of everyone's social media strategy;

2. Facebook is more likely to cover a larger number of users and prospects, as a opposed to a forum that hits a solid crowd of maybe 40 or 50 regulars;

3. Because of Facebook, and the dwindling influence of those 40 or 50 regulars, online fora in general are Internet Deadsville, daddy-o.

So, yeah, if you don't use Facebook (I don't anymore) you are at a slight disadvantage, info-wise, but my experience so far has been than if I don't get the information elsewhere, then it's probably info I really don't need - polls, conjecture, wild-eyed rage, trolling, etc. are not helping me do my job.

As for getting it "second hand," that attitude is one of the things that's killing the influence of online fora. You seem to imagine you should be told everything first, directly from MOTU, by virtue of posting here regularly, and that's not the case. If I worked for a company like MOTU, I'd be sick and tired of being made to feel that the forum somehow owns me. Forum membership of any kind (rocks, motorcycles, expensive watches, you name it) does not elevate your status above "Just Another Customer."

[cuttime]

As for getting it "second hand," that attitude is one of the things that's killing the influence of online fora.

Well, excuse me! I didn't mean to impugn this forum or the hard work of a dedicated moderator like James Steele. I always thought that an interested party was a better contact than some arbitrary algorithm. Forgive me of the crime of disappointment, hepcat.

[Tonio]

Thanks for the info nk_e


THIS - sad , but true :

MOTU spends more time at their Facebook page than they do here because:

1. Facebook is a huge part of everyone's social media strategy;

2. Facebook is more likely to cover a larger number of users and prospects, as a opposed to a forum that hits a solid crowd of maybe 40 or 50 regulars;

3. Because of Facebook, and the dwindling influence of those 40 or 50 regulars, online fora in general are Internet Deadsville, daddy-o.

So, yeah, if you don't use Facebook (I don't anymore) you are at a slight disadvantage, info-wise, but my experience so far has been than if I don't get the information elsewhere, then it's probably info I really don't need - polls, conjecture, wild-eyed rage, trolling, etc. are not helping me do my job.

"Just Another Customer."

[Gravity Jim]

As for getting it "second hand," that attitude is one of the things that's killing the influence of online fora.

Well, excuse me! I didn't mean to impugn this forum or the hard work of a dedicated moderator like James Steele. I always thought that an interested party was a better contact than some arbitrary algorithm. Forgive me of the crime of disappointment, hepcat.

Your disappointment is rooted in a bad idea: that your participation in this forum makes you more important to MOTU than other customers. My pointing out that this view is inaccurate doesn't denigrate James's work on this forum or the camaraderie you share with others here: it simply states the truth. Your disappointment would disappear if you didn't expect special treatment.

It's a fact that Facebook, ugly as it is, is Grand Social Station, and that's where the marketing happens. I know MOTU reads and responds to this forum, especially the magnificent MagicD, but posting this detailed explanation by Andrew at FB isn't really an emotional issue.

And yes, I am the heppest. I find that alienates some people, but at 60 years of age, I don't care.

[bayswater]

And yes, I am the heppest. I find that alienates some people, but at 60 years of age, I don't care.

So, aging gracefully is a thing of the past ...

[terrybritton]

And yes, I am the heppest. I find that alienates some people, but at 60 years of age, I don't care.

So, aging gracefully is a thing of the past ...

Once people are my age the graceful stuff starts to kick in...

Terry

[David Polich]

The Pre-Gen is more efficient, and decreases latency, but does so at the expense of increased CPU.

That's not what I read. It says:

"the pre-gen feature is completely separate from the latency updates. The latency update cuts latency in half"

The PreGen lets you run more VIs etc, with the same CPU. It does not reduce latency. The latency update reduces latency by half, but lower latencies still take more CPU, as they always have. The implication is that a latency of e.g. 10 msec now uses the same CPU, all other things equal, as a latency of 20 msec took before the update.

You are correct, Bayswater-san. My bad.

But anyway...so I can run VI's with less latency.
I "did" think my CPU hit would be less, not sure
how I arrived at that conclusion.

Since I only run one VI at a time and then record
its part as audio (I dont run multiple VI's
simultaneously, ever), I guess 9.13 isnt really
changing my life much. Looking forward to the next
version of DP with the improved pitch and
time stretching.