Cantabile pre-load rocks :)

Hi all,
yesterday I have starting using the pre-load feature for my setlist - about 30 songs with large Kontakt libraries etc: Cantabile loaded everything without breaking a sweat, topping at 16Gbytes memory steady. I played it a couple of hours with no issues, no glitches, and of course instant switching :grinning:

So great work :grinning: and kudos to Brad.
Next step: linked racks, to remove some Gbytes (hopefully!)
Luigi

3 Likes

I also just converted to pre-load for live gigs, and it’s made my life much easier.

@brad One question I have: To speed up things, I’ve been running my VSTs out of a RAM drive. Is this even necessary when using pre-load? Isn’t everything already loaded in RAM at that point? I ask this because my ultimate goal is to move everything into one monster set list, containing a corresponding Cantabile map to the programs in my Nord Stage. That will probably amount to 6-7 banks, and while many of the songs will be blank (or nearly so), I still may end up needing 20+ GB of RAM for the pre-load.

Anybody else out there doing this?

Regards,
-BW

Do you mean using a RAM Drive? I do use pre-loading to memory using a monster set list of about 325 Songs. I have a machine with 32 GB and one with 24 GB so lot’s of room but on plugins that have large sample libraries they still have to stream from disk so in that case I guess the SSDs I use in my rigs for samples are like RAM drives but they are not set up that way. If I may ask, what VSTs do you need to load that add up to 20 GB? I run around 12 to 13 GB but I may have a smaller set of instruments or smaller memory requirements. Any MIDI PG changes you map into your monster set list for the Nord Stage would take almost no space to store.

Cheers,

Dave

@dave_dore Let me clarify my earlier comments. I’ve been running my VSTs out of a RAM drive for some time now. That drive requires about 10GB for VSTs, as well as media files that I play. Beyond that, I’m starting to use pre-load to keep from having to put 5-10 sec. delays at the beginning of songs for performance. (I like to keep the flow going in a live set.) I’ve got roughly 50-60% of my songs converted from VSTs in Cakewalk to VSTs in Cantabile, and the initial pre-load is already at ~11GB. Doing the math, I’m figuring that it will easily top 20GB when I’m finished. So, I’m revisiting every part of the system. The big question is, should I be giving up 10GB of general RAM to load the VSTs on a RAM drive, if they are all going to be preloaded, anyway. I have 32GB total to play with on the Cantabile PC.

Regards,
-BW

1 Like

In my opinion, no, according to your calculations you should have plenty of system RAM without the need of the RAM drive. Do you use linked racks for your VSTs? FWIW that practice can reduce the amount of memory needed by a lot and speed up changes as well.

1 Like

I’m looking at linked racks. It will certainly help for synths. For guitar/amp VSTs? Maybe not as much. Most of my AmpliTube 5 presets are just an amp and speakers; but they could be almost any amp head tied to any speaker cab, in a variety of rooms, and I don’t repeat myself all that often. [My rationale: most (IRL) amp/pedalboard setups use overdrive, EQ, etc. to make one amp sound like many rigs; or as a way to get a certain tone at more manageable volume levels. When you have a palette like AT or Guitar Rig, why not just go for the ‘real’ sound, and then scale the volume accordingly?] The downside is that most of my presets are unique. (ex: out of 59 guitar setups, I’m using 22 different amp heads.)

Regards,
-BW

1 Like

I’ve been gigging with Cantabile for years and have faced the same question with my amp VSTs. My experience is that using linked racks for the amps is a huge benefit, and definitely lowers the CPU hit quite a bit because you’re only having one instance of the VST. Not using linked means you have multiple instances. This is crucial.

I now have one linked rack for each amp vst… 1 for S-Gear, 1 for Neural DSP Cory Wong, 1 for TH-U, etc. Individual racks use negligible CPU resources, it’s the plugins that use more resources. And having individual racks is easier to organize/program for me.

Each rack has numerous rack states with various tones. I set each tone’s volume and ensure that EQ falls within my personal parameters (i.e. HP EQ at 100Hz, LP at 5K or so, to mimic speaker cab), and then lock it so any changes won’t save unless you intentionally unlock it first. So they’re all “normalized” to the same volume and are gig-ready. Very important for live work, and easier to program new songs… just grab the rack states you want, make some song states and you’re done.

However, and this is key, you have the option to use Exported State saved at song level, which simply means that you can take a rack state that’s close, tweak it, and it’s saved in the Song State, not the rack state (unless you manually save the rack state too). That’s essentially what you’re doing with un-linked racks. Click once on a rack, at the top left select “States”. Here’s Exported State checked for the Rack:

And here, inside that rack, you can determine if the plugin’s snapshot/“Entire Bank” is saved only in the rack state, or also at the Song State. This is currently set for both:


The check on the left means Song State will take precedence over Rack state. (When both are checked you just need to verify, once in a while I get some odd behavior from certain plugins.)

However, there’s also a philosophical and workflow issue here, and that’s up to personal taste. Like you, because Cantabile offers such flexibility, I would try to nail the tone for each song and song part as perfectly as possible. And I had over 100 tones. But keeping them all normalized was a PITA, and endless tweaking was counter-productive. But now I’ve been trimming it down to about 20 tones. And my soundguy and I spent a lot of time tweaking EQ, drive, stereo separation, and effects for each one. That focus helped a lot, and each one sits perfectly in the mix now. Plus it makes the soundguy’s job easier, and our overall sound is better for it. However, sometimes on a new song, I’ll take one of those 20, and make a minor change, like adding reverb or whatever. Plus, I’m spending less time on tones and more time on playing. But hey, as I said, the answer will be different for everyone.
Hope this helps! Enjoy Cantabile, it’s incredibly powerful.
Tom

3 Likes

@twaw Hi Tom, thank you for the very detailed explanation. It got me thinking in another direction, and while I don’t think I’m going to be using exactly the same approach as you, I’m definitely going to use linked racks! :grinning:

I started with the guitar stuff, because most of my keyboard VSTs are modeled (vs. sampled), while the guitars use almost 3GB of samples. Just FYI, this is what I’m going to do for now:

  1. I created four linked racks, two for LP guitar, one for Strat, one for Rick. Each one gets its own copy of AmpliTube. The Strat and one of the LPs also gets a copy of Stardust 201 Tape Echo.
  2. For each of the amps (and tape echo), I dedicated up to 64 presets, each one representing a head+cab setup, plus a room sim, and any pre- or post-amp effects.
  3. For each of the guitars, I dedicated PC65-PC128, mapped down to PC1-PC64. Then I saved off guitars like ‘Les Paul Bridge Hot’ or ‘Strat Pos 2 Vintage’.
  4. For every song, I send two program changes to the linked rack at the start of a song, one for the guitar, and the other for the amp+efx.

Now, I just have to go in and substitute the proper guitar rack for the plugins I’ve already saved in a song … :roll_eyes: But the end result should cut down on both memory use and load time.

Regards,
-BW

Cool! Not sure I completely understand your goals, but on your #4, maybe you could use song states for that and avoid having to program 2 PC commands. A song state basically saves the various rack states and slider volumes… And more if you want it to. It’s essentially a snapshot of all the racks in your song, which rack states, the volumes of each rack, the routing, etc. If you click once on a song state and then at the top of the sidebar click on states. Then in the bottom left you will see the state behavior. Check off what you want to be saved with each song state. Enjoy!

I’ll probably end up doing that at some point. Right now, my goal is to finish porting my ~110 songs from 100% under Cakewalk to the Cantabile model. As such, there are times where I’ll stay with the less-efficient-but-more-familiar path, simply because I’m running out of time to learn new tricks. Cantabile has a lot of very cool features, but I’ve been using Cakewalk since v2.03 for DOS, and I can do some things without thinking about it. (ex: If the sliders/drawbars are active on my Numa Compact 2x, Studiologic insists on broadcasting their values on every PC. I can’t filter this out externally, or I lose the sliders as MIDI controls. So, sending Numa PCs and then Stage PCs – sometimes, even a repeat PC for the Stage – at the start of a section is kind of second nature to me.)

When I have time, I’ll work on simplifying things. I really wish that I would have moved to Cantabile three years ago – it would have made things much easier!

Regards,
-BW