96 khz sample rate i/o for better VSTi performance?

[JJ_Jettflow]

First watch https://www.xiph.org/video/vid2.shtml

Doesn't work in my browser.

But is there any difference to stemming out the VST to a higher sample rate than the project or is it necessary to set the project to the higher sample rate first and then stem out the tracks at that higher sample rate?

[zerocrossing]

Your CPU is kind of low to use 96khz. 2.5ghz will be put to the test on any large projects at 96khz.
You may be able to use 88.2khz instead. The benefits are almost the same, but the CPU time cost is only reduced marginally.
Synths are best oversampled, most decent synths have an oversampling option. I would not increase sample rate just to get a better sound from a synth, because everything else will double the CPU cost.
Equalizers, reverbs, delays, distortions, etc all sound better at 96khz. The true amount of aliasing reduction is 6dB, but if your sounds are bandlimited at around 16-20khz, you get less audible aliasing than -6dB. Aliasing is like a mirror, and with the distance from the mirror increased, the distance is doubled... there's more room for harmonics to exist between the area before the mirror and after it's reflected, above the bandlimit point. You can still have frequencies all the way to 48khz in the signal, but there's really no need.
One thing you must learn to do when working at 96khz is bandlimit. Use lowpass filters to cut away the ultrasonic treble before effects processing AND after.
The thing about plugins and 96khz is not all plugins support it correctly. Lots of plugins have only been designed to work up to 48khz. So take time to find out which ones work correctly, for example an equalizer should always work at the same frequencies at either sample rate. Delays should always delay the same amount of time at any sample rate. When a plugin does not take sample rate into consideration, uping the sample rate can result in effects like time "speeding up" and frequencies increasing, synths as well.
If you are sample heavy, resampling can cause a more unexpected slowdown than only double the CPU. The memory usage is also doubled in size, and this can result in a greater slowdown if the computer's memory subsystem cannot keep up. You can mitigate this somewhat by using faster, lower quality resampling processes.
You should notice the difference between 48 and 96khz for production immediately while using EQs and distortion effects.

There sure seem to be a lot of pitfalls... what speed processor do you recommend? Would upping it to an i7 running at 4 khz do the trick? I tend to use about 10 plug ins per project, but some are low CPU plug ins like Möbius.

Where I'm hoping to get a big benefit is using the new Reaktor 6 blocks. I'm already pretty impressed with Monark and how it deals with audio rate filter FM and feedback. My thoughts are if I could get that sound and possibly a tad better out of an R6 instrument made with blocks, it would be amazing.

My dilemma is this: Should I sell off most of my hardware synths and use the proceeds to buy a blazing CPU and 96 khz audio interface and shift my workflow ITB, or stay where I am and add a SE80 which does things like cross modulation and feedback effortlessly.

The pros of software: flexibility, variety, patch recall, no clutter or spaghetti mess.

Pros of hardware: computationally irrelevant, unique, dedicated interface, plug it in and it works. Clutter and spaghetti make me seem cool (to myself ) like I'm preparing to free Neo from The Matrix.

I'm really torn. On one hand I love hardware. The sound. The totemic nature of it. There's something romantic about it. OTOH, it can drive me crazy. I've definitely accumulated too much of it to be physically ergonomic, but at the same time I love the variety.

What to do... what to do...

[sqigls]

don't sell all of your hardware.
computers supersede. i've done the same thing, spent around 5k on a killer CPU badass rig… say 6 years later i couldn't sell the machine for $100. personally, i don't miss the spaghetti mess, or the power bills, but I miss the hands on and the fact is, I gave that computer away, and I could have bought a flipping OSCar for 4k!!!

2 cents detuned

[sqigls]

…on track, PLEASE DO work at 96. We need to drag this 16bit MP3 bullshit out of the dark ages.
WE HAVE FUNKTION ONE FOR GODS SAKE,
when people hear your live set at 24/96, after listening to "Joe MP3" for the last 2 hours, their ears are gonna think they've been reborn.

[DJ Warmonger]

Should I sell off most of my hardware synths and use the proceeds to buy a blazing CPU and 96 khz audio interface and shift my workflow ITB

Come to the digital side

computers supersede. i've done the same thing, spent around 5k on a killer CPU badass rig… say 6 years later i couldn't sell the machine for $100

It's the other way round. Computers gets faster and more powerful over time, analog synths do not.

[do_androids_dream]

when people hear your live set at 24/96, after listening to "Joe MP3" for the last 2 hours, their ears are gonna think they've been reborn.

When music is played at the sort of volumes hit at live events quality gets thrown out the window anyway. No one's going to notice

[pdxindy]

96khz makes a clear and distinct difference for the soft-synth and effects I use most. There is a more appealing smoothness to the high end. But then I push my synths hard... especially Bazille and ACE. Diva has less benefit at 96khz but I can still hear it on really high notes.

I wish DAW's gave the user the ability to choose per track. I'd rather have my project at 48khz to save disk resources and just set specific synth tracks to render at 96khz and resample down.

[zerocrossing]

don't sell all of your hardware.
computers supersede. i've done the same thing, spent around 5k on a killer CPU badass rig… say 6 years later i couldn't sell the machine for $100. personally, i don't miss the spaghetti mess, or the power bills, but I miss the hands on and the fact is, I gave that computer away, and I could have bought a flipping OSCar for 4k!!!

2 cents detuned

Oh, I have no intention of selling all my hardware analogs, but I'd like to pare it down to two or three.

[cron]

For my money, plugs that sound better at 96 kHz are the ones that obviously, audibly alias at 48 kHz. Who cares what's on the scope? - the ear is king. I recall all (?) Valhalla plugs had a pretty nasty aliasing bug until a fair while after VVV was released. A member spotted it in the scope and Sean was quickly able to identify and fix a problem in his code that existed in all plugs. Nobody would have noticed it otherwise.

The only one I've got in my arsenal that audibly benefits is Cyclop (and that's only when abusing it to do super high end twinkly stuff with the additive oscs). Sadly, I find that just switching up to 96 kHz for the render isn't really feasible most of the time. There's always some plug-in in the chain that sounds radically different at 96 kHz to mess things up. Ironically Cyclop is one of them. The grain engine plays grains at double speed.

I remember something like 10 years ago a (now banned?) member called arke wrote an oversampling plug that would host and oversample individual plug-ins. Does this or something like it still exist? It'd be a fantastic compromise because running the entire project at 96 from the start is just too juice heavy for me.

[acYm]

yeah that's vst oversampler, there's a dl link in the comments of the plugin's kvr page

[camsr]

There sure seem to be a lot of pitfalls... what speed processor do you recommend? Would upping it to an i7 running at 4 khz do the trick?

Yeah that would work better. Important to remember that if sample rate is doubled, the CPU operations are doubled and the memory throughput necessary is also doubled. Do you have a modern system already or does it use DDR2 ram? My core 2 quad with ddr2-800 ram begins to choke at 96khz after about 12 "good" effects. Some newer plugins (like MJUC Jr) didn't even work at all. The clock speed is 3.2ghz. The fastest CPU for desktops today is only double the speed (in CPU ops) of my system, and I bought it almost 7 years ago. The bottleneck is the memory.

[sqigls]

when people hear your live set at 24/96, after listening to "Joe MP3" for the last 2 hours, their ears are gonna think they've been reborn.

When music is played at the sort of volumes hit at live events quality gets thrown out the window anyway. No one's going to notice

I wouldn't say "no-one" is going to notice.
Most people wouldn't consciously, but it certainly DOES make a difference.
Reverb tails, sweet juicy Diva… I certainly notice

[Synthetic Wav]

96khz makes a clear and distinct difference for the soft-synth and effects I use most. There is a more appealing smoothness to the high end. But then I push my synths hard... especially Bazille and ACE. Diva has less benefit at 96khz but I can still hear it on really high notes.

I wish DAW's gave the user the ability to choose per track. I'd rather have my project at 48khz to save disk resources and just set specific synth tracks to render at 96khz and resample down.

bazille hq can dull a lot the sound, so at higher sampling rate it probably don't happen.
dunno if it s a bug, but if it was proper it should not really benefit from using 96khz in many case i guess.

[3ee]

Please, anyone who has doubt that a higher sample rate helps you or not... do some tests!

With higher sample rates:
When and single track/sound is concerned, that will help your -sampling- if you lower the pitch, keeping the spectrum "intact" (it's a matter of mathematical division) for example, 88.2k divided by 2 = 44.1k that means if you have a recording of 88.2k and you lower 12 semitones you have the quality of 44.1k

When multiple tracks (layering) is concerned .... (this get a bit complicated with the PPQ of whatever it's called in your DAW) higher sample rates aligns the freqs more accurate, you really need to know what/how "phase aligning" helps and not only that... effects like chorus and flanging will do their work more accurately.... even if your finale sample rate is 44.1k you still get different results when down-sampling a higher SR mix.

I don't know how to describe it accurately, but I am totally convinced that higher sample rates definitely helps... you just have to have a total understanding on how it works and train your ear to pay close attention on how it affects the sound.

If CPU is concerned, record @ higher sample rates but mix @ 44.1k for example.... but export/render @ higher sample rates like 96k or 192k even if you downsample to 44.1k CD standard for example.... the summing is there...

[Synthetic Wav]

i did some test already in the past and a 1 db eq boost on the master can have more effect than using 44.1 or 96khz sampling rate.
it depends the plugins but if some clearly benefit from using a higher sampling rate in the daw it's cause of compromised/ poor design imo , it shouldn t happen .

one good thing of using higher sampling rate is that you can bypass some poorly implemented resampling that harm the sound even more than the artifacts it's supposed to reduce but best is to avoid these plugs anyway if possible