Questions about two features not in Basic

[Jesse Skeens]

MIXER AND MULTITRACK GROUP PRESETS
REGION / PEAK DATA MEMORY CACHING


Can someone describe these to me? Can't seem to find any info on them.

Thanks

[Yura]

Jesse,

you see, here is all in this table about you question.

http://www.sawstudio.com/SAWStudio/SAWStudio_Compare.htm

[Jesse Skeens]

Yura,

Thanks but what I was wanting were descriptions of these functions. I was looking at the table and just wanted to know what I was missing in Basic.

Jesse

[MMP]

Region & Peak caching puts either of these data types into ram with the intention of speeding up HD access. I always use peak data caching...region caching has always felt a little unstable to me.


Mixer and MT grouping are selection presets. Say you want to affect the volumes of the entire kit, you create a preset on the mixer that selects all those fader channels with one selection. The MT grouping is for selecting multiple tracks at the same time for group editing.

Hope that makes it clearer rather than more murky.

MM

[Jesse Skeens]

Thanks!

[Oz Nimbus]

region caching has always felt a little unstable to me.



MM

It caused me some serious problems last summer. If you're running at very low latencies, I seriously suggest you turn it "off."


-0z-

[earthpig]

along those lines what is the ENHANCED DUAL CPU PERFORMANCE

and

SPECIAL DUAL MONITOR FUNCTIONALITY

?

:)

[Bob L]

The Region memory caching is a great bonus if your machine has large amounts of physical ram... like 1 or 2 gigs.

For very large sessions, that normally take a few seconds each time you start playback, you can cache the entire (or most of) session in ram and remove the disk drives from the loop. I would recommend doing this during the mixing phase... after serious editing and manipulation of the data.

Then starting playback, even on a 72 track solid session is instant... so as you spend hours jumping around adjusting controls, you no longer have the long wait while the data has to be cached from the hard disk each time you start playback.

The enhanced dual features take advantage of dual cpu machines by splitting the threads in a specific way between the two cpus... this can enhance screen redraws and meter displays during loaded sessions that normally run above 50% MT Load.

The Dual monitor enhancements allow you some control over dialog box placements for dual monitors and configurations for vertical or horizontal monitor arrangemnets.

Bob L

[Naturally Digital]

For very large sessions, that normally take a few seconds each time you start playback, you can cache the entire (or most of) session in ram and remove the disk drives from the loop. I would recommend doing this during the mixing phase... after serious editing and manipulation of the data.

Then starting playback, even on a 72 track solid session is instant... so as you spend hours jumping around adjusting controls, you no longer have the long wait while the data has to be cached from the hard disk each time you start playback.
Bob L
I've never heard/read it described quite like this (as affecting playback start time). The name of the function has always been somewhat obvious but...

For much of my mixing/editing/mastering work I have wanted to run smaller buffers. One, for closer to real-time metering but more so, for that 'instantaneous' playback start. I've always liked the feel of a quicker start. I'd never tried region caching, thinking it wasn't necessary for me and recently we've learned that it's better to give the buffers some reasonable size. Now I'm really glad to have some new light shed on this feature.

This is another one of those little details that make SAW special!

Thanks,
Dave.

[Bob L]

David,

You can run any size buffers in SAWStudio for mixdown and still obtain absolute NO Latency Playback Meters...

That is part of the engine design... the meter displays DO NOT depend on buffer latency as they seem to do in other applications.

Try it... set your buffer size to 4096 and look at your kick drum meters... they will still be right on time.

This is why I suggest using higher buffer sizes for dense mixdowns... because the latency will still feel and look realtime... 4 x 1024 will still feel realtime when pushing solos or mutes or changing eq.... and the engine will always be more stable.

You only need low latency when doing realtime live mode monitoring thru the console or when playing live into VST synths... you don't even need low latency for playing back midi tracks through VST synths after the tracks have been recorded... only during the performance.... they will playback in perfect sync no matter what the buffer size.

Bob L

[Yura]

The enhanced dual features take advantage of dual cpu machines by splitting the threads in a specific way between the two cpus... this can enhance screen redraws and meter displays during loaded sessions that normally run above 50% MT Load.

Bob L
Bob,
So you mean this is the difference between Hiperthreading system snd real Two CPU system for SAW? if we have machine with two cpu we have not to Force single cpu in SAW?

[Bob L]

You don't have to force single cpu in hyperthreading machines either... SAWStudio will treat it like two cpus and split the threads... its just that many hyperthreaded machines have stability problems in windows with certain motherboards and drivers, so I always set them back to one cpu.

Hyperthreading does not seem to give any real performance benefits at all in any testing I have ever seen. Dual cpus can definitely improve performance of SAWStudio with heavyily loaded sessions... but again, they seem to be very flakey in many areas... like firewire for one. I choose not to use them.

Bob L

[canipus]

Fact- Hyperthreading has no advantages on the Prescott Die until the clock speed gets up to 3.45Ghz. Enabling it at lower speeds is simply burning up the advantage as heat closes the silicon matrix in the pipelining - and the heat has more of an impact on the arithmetic performance than the extra throughput in the pipeline - result negative performance impact.

Canipus

[earthpig]

Does Saw support addressing more than 2 gig of memory ?
If a machine has more than 2gb say 3-4 gb then you have to turn on a boot option so that windows will allow the application to address this memory. ( /3gb in boot.ini ) However the app has to be written so it can use more than 2gb, this would be great for region caching I think.

http://www.intel.com/cd/ids/developer/asmo-na/eng/189869.htm?page=4

http://msdn.microsoft.com/library/default.asp?url=/library/en-us/ddtools/hh/ddtools/bootini_de16d3ec-c437-4628-805f-8945ea598a92.xml.asp

http://www.panoramafactory.com/discus/messages/17/313.html?1057432320

[Bob L]

Nothing specific is done in the program code to address more memory...

Not sure what would happen... most still have not gotten up to the 2 gig limit.

Bob L

[AudioAstronomer]

I beleive I brought this up near the beginning of the forum... It would be nice as Im hopin to upgrade my machine to 3gb so I can turn on region memory caching for good... in a lot of my large projects it really gets hairy even with 2gb now.

[Bob L]

Region caching may still take advantage of the ram since its not really code address space... but small data regions that Windows may Virtual address and feed to the program when the data is asked for.

When you get your machine up to 3 gigs... let us know what happens.

Bob L

[AudioAstronomer]

Ill be doing it tuesday/wednesday more than likely now...

[earthpig]

From one of those links..

" Note that an application must be linked with the LargeAddressAware flag, in order for it to take advantage of the /3GB switch."

I currently have 3gb in the machine at moment so if there is any testing that could be done for this I am willing to give it a try.
Especially usefull if you start messing about with 96/24 some of those files can be pretty hefty.

[Bob L]

Well, go ahead and fill up the ram with a huge session in Region cahing mode and see if SAWStudio can find all the regions.

Bob L

[SoundSuite]

From what one of those links said...
Don't do the /3gb in XP home, or XP Pro with less than SP2

I've only seen this done in Server builds, never had anyone with a DAW and >2GB RAM.
Additionally, never tried on XP, only done personally on Windows 2000 AS.
It was a database server with Dual Xeon's and 4GB.