Processing power?

  • Thread starter Thread starter ecktronic
  • Start date Start date
TravisinFlorida said:
have you tried freezing some of the plugins on your tracks? that should free up some cpu usage.

Not sure what you mean by freezing.

I want all plug-ins to work in real time as I will be automating the plug-in variables in the mix.

Thanks,
Eck
 
xstatic said:
My system runs just fine with 48 tracks and 100 plugins. Its not even dual core....
Fek!! What is your cpu? Do you have the UDA?
 
Robert D said:
Before I upgraded my system I was running 48 tracks on a P-III 700 with a TC Powercore card doing the verbs and other high demand plugs. Your P4 is probably up to the task if you do a fresh windows install and optimize it for audio, and then work smart. Working smart means not using multiple plug instances where 1 plug on an aux buss will serve multiple tracks, cleaning up extranious edits and such, etc. Take a couple of hours to make a clean test project and methodically find out what processes add what amount of load to your computer. Find out which plugs are CPU hogs, and which are efficient.
You've either got to do this homework to figure out how to make your current rig work, or upgrade.

-Bob

Thanks alot.
I am looking to upgrade somehow but not sure which way yet. cpu upgrade or UDA card.

The Waves R-verb plug-in by uses up alot of power, the most actually.
I amnt looking to find ways around this like creating a reverb aux track. I need complete control of each plug-in as I am going to be atuomating the plug-in controls and using a different verb for every track almost.

Thanks alot for your post.
Eck
 
ecktronic said:
Not sure what you mean by freezing.

I want all plug-ins to work in real time as I will be automating the plug-in variables in the mix.

Thanks,
Eck

Just commit man. Don't be a control freak. :D
 
I would go for the DSP card (UAD), the plugins are really good and will probably go alot further than a CPU upgrade
 
well the primary source of audio goodness aside from soundcard is RAM, lots and lots of juicy RAM. Not too expensive.....
 
ecktronic said:
I am going to be atuomating the plug-in controls and using a different verb for every track almost.

Eck

Ok, this is where you are getting into trouble. If this is something you feel you need to do, then it's time to introduce you to Mr Freeze.

ecktronic said:
Not sure what you mean by freezing.

Track freezing is a feature that a lot of recording software is now supporting. It is essentially a method of creating a bounced version of the track, taking the original track and all of it's effects and other loads offline, hiding the original track, and replacing it with the new bounced track, all at the click of a mouse. Unfreezing is another click away. It's a great tool for people who, for instance, feel the need to have 40 different reverbs! :)
Otherwise, you're going to need some serious MIPS to do all that processing.
 
What sample rate are you using ?

If you're using 96Khz, you will get a big performance boost in changing to 48KHz, with a minimal loss in quality.

The reverb plugs are typically the heaviest CPU users. If you're able to freeze the reverb'd tracks and disable the plugs on those, that will help.

Also if you've got say 6 tracks of drums, you can bounce them down to stereo, and archive the originals.

It all adds up
 
Moving to a nicer CPU (something in the dual category;) ) will be a huge improvement. Adding RAM is a really big boost until you get to about 2 gigs. Then the DSP card is what will be the next greatest boost for your system. The UAD cards DO put a little tax on your system and that taxing seems to increase exponentially when you start filling up the UAD's DSP.

In my opinion here though, the single biggest problem that I can see is that you are using so many individual instances of reverb. Why on earth would you feel you need to slap a different reverb as an insert on most every channel?
 
altitude909 said:
I would go for the DSP card (UAD), the plugins are really good and will probably go alot further than a CPU upgrade

Thanks for that.
Eck
 
Robert D said:
Ok, this is where you are getting into trouble. If this is something you feel you need to do, then it's time to introduce you to Mr Freeze.



Track freezing is a feature that a lot of recording software is now supporting. It is essentially a method of creating a bounced version of the track, taking the original track and all of it's effects and other loads offline, hiding the original track, and replacing it with the new bounced track, all at the click of a mouse. Unfreezing is another click away. It's a great tool for people who, for instance, feel the need to have 40 different reverbs! :)
Otherwise, you're going to need some serious MIPS to do all that processing.


Yeah I am running Nuendo v2.2 does it do tthat?

If not I can always just render the plug-ins to the track to save process power.

Thanks again.
Eck
 
Bulls Hit said:
What sample rate are you using ?

If you're using 96Khz, you will get a big performance boost in changing to 48KHz, with a minimal loss in quality.

The reverb plugs are typically the heaviest CPU users. If you're able to freeze the reverb'd tracks and disable the plugs on those, that will help.

Also if you've got say 6 tracks of drums, you can bounce them down to stereo, and archive the originals.

It all adds up

Im using 48Hz.

Thanks dude.
Eck
 
xstatic said:
Moving to a nicer CPU (something in the dual category;) ) will be a huge improvement. Adding RAM is a really big boost until you get to about 2 gigs. Then the DSP card is what will be the next greatest boost for your system. The UAD cards DO put a little tax on your system and that taxing seems to increase exponentially when you start filling up the UAD's DSP.

In my opinion here though, the single biggest problem that I can see is that you are using so many individual instances of reverb. Why on earth would you feel you need to slap a different reverb as an insert on most every channel?

Thanks alot dude.
I have 2GB RAM so next stepwouldnt be more RAM then i suppose.

I want to have different reverb on most tracks as each track needs a different reverb plus I am automating the reverb tails etc.
I could use 1 reverb for the toms and OHs, 1 reverb for the snare, 1 for the kick i suppose. But all other tracks would need different reverbs like longer tails, different delay times. I feel these are important to making a mix sound better.

Cheers,
Eck
 
The reason I ask is that this is not how people get their mixes to sound the way they do. In fact, you say that these things are important to make a mix sound better, whereas most engineers (quite possibly even all engineers with experience) might say that this is exactly how to make a mix sound cluttered, scattered, and pull all definition out of it. I don't want to sit here and tell you how to do things, but changing to a more conventional for of effects bussing might just sound a whole ton better. Not only that, but it will certainly take care of your maxed out processor problems. Not only are you running WAY more reverbs than I have ever seen anyone do, but you are picking one that is also notoriously processor hungry.
 
xstatic said:
The reason I ask is that this is not how people get their mixes to sound the way they do. In fact, you say that these things are important to make a mix sound better, whereas most engineers (quite possibly even all engineers with experience) might say that this is exactly how to make a mix sound cluttered, scattered, and pull all definition out of it. I don't want to sit here and tell you how to do things, but changing to a more conventional for of effects bussing might just sound a whole ton better. Not only that, but it will certainly take care of your maxed out processor problems. Not only are you running WAY more reverbs than I have ever seen anyone do, but you are picking one that is also notoriously processor hungry.

Yeah the R-verb is pretty hungry! I have True-verb also but im sure its just as hungry.
I have the built in Nuendo reverb which is pretty digital and nasty so I aint using that.

I can see where you are coming from with using to many diferent reverbs cluterring a mix. I supose I am just over exadurating so my cpu wont die, kinda being over cautious if yo know what I mean.
But saying that do you not agree tha for example a vocal reverbs delay would be around 6ms to let the vocal attack come through before the reverb hits it and also for example a guitars reverb delay would be around 20ms for the same reason?

Thanks Xstatic

Eck
 
how are you only squeezing in 15 tracks? i have a crap computer (less than 3 ghz, 1GB Ram, 140 GB HDD) and i can fit 39+ tracks with three plug-ins apiece on each track?
 
ecktronic said:
But saying that do you not agree tha for example a vocal reverbs delay would be around 6ms to let the vocal attack come through before the reverb hits it and also for example a guitars reverb delay would be around 20ms for the same reason?
Eck

Sure, and no one suggesting 1 verb for everything, though that is a viable approach for the 1 band / one room philosophy. But I don't think I've ever used more than 6 verbs in a mix, and more often than not it's 4. As always, it's whatever works for the song, but I agree with xstatic that too much of a good thing often isn't, and the listener will probably lose any kind of sense of a coheisive acoustic space with all those verbs going on.
 
I think that adding seperate verbs for everything is far more damaging to a mix then sya having a single track sharing a reverb that may not have the "perfect predelay" time. I often end up with no more than two verbs in a single mix. In fact, you would be surprised at how often a vocal will actually work well in a good drum verb. I also think you are fooling your own ears if you are worried that much about the predelay being 6 ms instead of 20 ms. Often times vocal presets will actually use much more than 6 ms (like 30-50 ms predelays). It all depends on what you are after. Maybe you should look into some convolutions. Maybe the reason you feel the need to use so many is that you just haven't found the right verb yet;)
 
The reverb in an aux send bus is the answer, having seperate instances of the same reverb in each track isn't sensible and will waste cpu power.

If you want it to sound like the band are all playing in the same place, then one reverb in an auxilary bus is all that's needed to simulate that one room (use the track aux send pans to place the track in the reverbs soundfield).

It's not unusual to have some tracks such as a snare with a different reverb type for a particlar effect, also you may find you can use a more basic lower powered reverb plug-in where you only want some thickening of an individual track and the aux send reverb you're using for everthing else isn't right for it.
 
Back
Top