Noel, we think alike. I just came to the same conclusion (see my edited post) and then noticed your reply on the refresh! hahaha.

So, that is indeed how you do it.

Very interesting stuff.

I am sorry for not being as technical as others here. Out of all of this though it is still not clear to me that given my current system whether or not I can expect more performance with regard to the Audio Engine in X2.
 
It's really just a general question that overall, should the average user expect better performance. If not, then there is really no reason for me to upgrade at this point. Not that I wouldn't want to make use of some of the other new features and functions, but I need audio performance more then anything right now.
 
Thanks
 

Thanks, Noel. Yeah it's maxed. :-)

Since my chain is so close to fitting into a single core, it's possible even a small improvement in X2 could push it into the green. I'm crossing my fingers.

I may even try for another 100 mhz on my overclock, too.

The single core executes it all when I'm driving it with audio. It's just that liiiittle bit extra with a VST, instead, that kills it. So close.

Thanks, Guitarpima. No, I have not messed with the MIDI buffers. Good idea. I'll try it.

Hi, Jeff.

Thanks for the suggestions. But, I've been studying mastering since the late 80's, produced a lot of boy band style pop and did it the more traditional way, for years. So, now, I'm very confident and opinionated about all the things I'm intentionally "doing wrong." :-)

My newer production approach centers around "mixing into the mastering chain" and there is no looking back, for me. I think it's very specific to an EDM style of music and does not apply, as much, to most other styles.

There are some very specific goals I need to reach in terms of dance-floor-ready dynamics. Having an "iterative feedback loop" -- being able to change the stem source material dynamically -- presents me more control in reaching that goal.

In other words, the more a genre of music evolves to become more characterized, defined and inextricably bound to specific mastering attributes, the more natural it becomes for mastering to shift upfront into the creative process. I mean, there is literally a "final db" placement of a kick and snare that almost has to be what it needs to be. I wanna make sure it "gets there," has a space carved out for it, without collateral damage. That it's not a compromise made during the mastering process. Remember, mastering has traditionally been much about compromises made because the mastering engineer doesn't have the ability to take a stem or stereo file apart. Mastering engineers go to great lengths, using sophisticated mid-side processing, etc. in a sometimes futile attempt to change attributes of a misbehaving element in the mix. If those mastering engineers could only have access to the darn instrument to fix it on the spot, I promise you they'd do it (when no one was looking). Well, now we can, as we're all mastering engineers and we have the tools to do it non-destructively and in real-time. It's a new age.

Actually, I disagree with you on the only "3 or 4 processes," for my EDM genre at least. Just the opposite; I'm finding it better to make many small, subtle changes in dynamic shaping, rather than big changes at the end.

For example, I'll agonize over 1db of gain reduction that I might yield from a non-linear summer's overdrive modeling so as to get a small amount of "crunch" in the top-end of a kick.

Yes, my latency is already maxed. Thanks for that suggestion, though. I still have a few more things to try (like the MIDI buffers suggestion).

Yes! Love it. I used to rent the hardware version of the 2500 and am thrilled with Waves emulation of it. It has the same seemingly endless amount of forgiveness. It's like you almost can't push it too hard. :-) That said, I'm only putting about 2 or 3 db of gain reduction at any stage, now -- with my newer mastering approach. 

The API2500 is amazing for keeping the integrity of kicks and bass in a full mix. That is, it's a great compressor for the mastering stage, a stereo mix.

Get it. You'll love it. Also, you can't have too many compressors. :-) Each compressor is unique in its subtle characteristics. As many mastering engineers say, "compression IS eq, too." Compressor nerds are a whole subculture in the Gearslutz forums. I've seen forum threads go back years over one compressor.

Cheers.

attalus
  
I am interested in the Low-Latency performance improvements of X2 as well. I wish someone from Cakewalk would talk about this! I just bought X1 to get X2 for free in large part because of Low-Latency improvements, i hope i won't be disappointed. This is important to me and Audio engine improvements are amongs the few things i care about. I hope i won't be disappointed.
 
Thank you. That was the point of the original post but although some good discussion here it should have been in another post and I am still not warm and gushy on the X2 upgrade. MAybe I am missing something but although I've seen references to low-latency improvements I've seen nothing to what expectation that may raise.
 
 

I'm no expert here, and all this talk of cores and stuff goes right over my head, but ever since I've turned off "core parking" I've found that, in general, all of my cores except for the first one are all used pretty much equally.

I'll sneak back behind my rock now...

yeah, it's good to bring up core parking, disabling that is a good practice for DAWs in general (yup, mine's disabled).

the cliff notes version of this thread is that Sonar is optimized for low-latency performance (a good thing) over something called pre-buffering (read thread for more on this). basically, this means that when Sonar deals with a serial single path (like you get with a mastering chain) it must crunch it all on single core. So, if you have a good-sized chain that won't fit on a single core, the engine will drop-out. Most users won't run into this as Sonar will nicely utilize those extra cores for other tracks, effects chains on those tracks, and other app-related tasks. Overall, Sonar is very good at multi-core, just not in my non-typical use-case.

it's only a problem when you're doing a chain that can't be broken up into parallel (a.k.a. "track") signal paths. the chain i have hovers around 20% and occasionally peaks into the 30% range, which means on my 4 core system 25% of the total cpu output = 1 core (25% x 4 = 100%). so during those peaks i go over my 1 core allotment (for that particular chain) by about 5% and get the drop-outs.

I'm posting this for the archives, in case it helps anyone, in future.

I finally found a really solid solution to my desire for a real-time, ITB, mix-into mastering chain, with Sonar X2. Quick recap: The mastering chain is too demanding -- Audio drops outs like crazy. My computer can't be faster, it's a liquid-cooled Ivy Bridge. I can't use a separate computer as I can't afford buying a second ilok licence for all my waves plugins, etc. The computer is fast enough, but Sonar won't break up that mastering chain across more than one core.

This solution was inspired after trying the Studio One V2 demo (nice DAW, btw) and noticing that it, too, was putting the entire serial, stereo bus effects on only one core (and getting audio drop-outs) exactly like X2. Sonar seems to even be a little better performing than Studio One, in this regard.

My solution: I use Sonar's "external fx" out to go into Reaper (running in live, record-enabled monitoring mode only). Reaper just acts as a dumb VST rack. Then, out of Reaper back into Sonar -- for just the mastering chain portion.

Bam. That's it.

Reaper has settings to limit which cores it uses, so I set it to use the last four (Sonar can then hog the first four all it wants). Reaper's audio engine will spread the mastering across its four allocated cores, so there is plenty of horse-power there for the "external" effect.

To avoid buying another audio interface, I found a way to "hack" the one I have by creating an all-digital loopback on it. I just took two cables for my MOTU 896HD (Virtual Audio Cables were not working well) -- an optical cable, looped from the ADAT 1/2 out to the ADAT 1/2 in (a closed loop). Then I did the same on the AES/EBU 1/2 digital out/in. Pretty sneaky, ey? So, it's Sonar ADAT port out, to Reaper's ADAT port in, then Reaper AES/EBU out, to Sonar AES/EBU in -- the Sonar external effect module completes the connection.

It works so slick. Nothing important is lost via the 24 bit digital connection, so I can do a commercial-grade Ozone dither at the end of the chain (in Reaper) and what comes back into Sonar it's not altered in any way (no clipping overs, etc.)

If you didn't have an audio interface that you could do this sort of loopback workaround, you could just buy another audio interface with a digital connection (like ADAT lightpipe that most have) that's compatible with the digital connection of your primary interface.

Basically, this is a way unlock the full potential of today's really fast computers and keep it all in one box. I think mastering is the main use-case for this, but perhaps there are other uses.

Anyway, I'm very happy with this solution. Everything is working more smoothly now, as it's taking that burden of my main DAW.

Cheers.