Well...In this case, it doesn't.

I would say, using the picture analogy...I'm not sure if you're familiar with photoshop and image editing, but lets say you have an image of a pretty model and you have to edit it on Photoshop and then save to a highly compressed 800x600 pixels format for web purposes. In which case you'll get better results?

a) You get this picture into Photoshop with the highest resolution you can, so you can meticulously edit all its flaws, remove its imperfections, improve the tonal balance, improve the lighting, etc...And only then, after you're done, exporting this image to the final 800x600 format...

OR...

b) You downgrade the image to 800x600 before importing it into Photshop, then perform all these highly meticulous tasks already on the final crappy resolution and them simply save it and call it a day.

I'd say case A will give you way better results, without a doubt.

 

Cant follow that analogy. Your photoshop analogy compared to audio makes no sense for me. On a picture you zoom to see redicolously fine pixels to edit them. You cant do something like "zoom" with audio so you "hear" it better or hear above 20khz.
I belive Craig Anderton. You cant hear a difference if something is recorded in 96k or 44.1k. It only makes a differnce in 1 case. You have to be completly in the box and you have to use a VST that dosent oversample completly or does not oversample at all. Than you could hear a difference.

I dont wana talk bad about a vst but a little example. I stoped using Guitar Rig as i found out it only overamples the amps but not the effects in the plugin. I dont know if the newest version still does that. But with that version of Guitar rig you can easily hear a difference if you work in 96khz or 44.1khz.

I am a strong beliver that if you hear a differncece in your recordings 96 - 44.1 you should check your plugins and throw some of them away and never touch them or use hardware.

 

Somewhat. Enough to make that gif.

Of course I'd pick (a) given options like 1600x1200 vs. 800x600, that's not the case. In case of 48k vs. 44.1k, my "highest resolution possible," 48k, is only fractionally higher (<10%) than my render resolution, and the math is hard (not 2:1, as with 88.2k/44.1 or 1600/800, it's 48000/44100 =1.0884353...

My gif shows doing just what you suggest: I downsample the original (frame 1) to 92pix (frame 2, think 44k sample rate), and, for frame 3, I first downsample the original to 100pix (think 48k sample rate), and then resample [the 100pix result] again to 92pix (frame 3). See the parallels?

Granted, no editing was done along the way -- straight-up sample rate (resolution) conversions. Editing may or may not make conversion mess less noticeable.

 

You realize there are two possibilities here:
1. Bad analogy
2. You fail to see/grasp it.

I try not to mix faith and tech in the same thread. Seems like a recipe for a long and fruitless conversation.

 

Ok i change that ...
if some one like Craig Anderton who did classical recordings with some of the best classic musicians tells he or the musicians (really trained and great ears) could not hear a differnce between a 44.1 khz and a 96 khz. Noone can. Fact.

And I change i am a "strong beliver", to "I know"

Everyone that isnt mixing completly in the box or uses correctly oversampling plugins only wastes cpu power and hd space if he mixes in 96khz.

And about the analogy- My english isnt very well ... still the analogy makes no sense for me. Apples and orange. Picture editing and Audio mixing. Hearing and seeing ... "zoom" and "???"

 

If I suggested that I can hear the difference between 44.1 & 96k sample rates, or that anyone could, point to where I've made this nasty blunder. I'd like to redact (or, at least, correct) it.

What I did was cast a bit of doubt re. the benefits of recording/mixing @48k sample rate, if the final render is @44.1k.

You might also "know" that 2+2=7. As for me, I'm a bit skeptical.

oic. Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn?

Please read the thread, or just a few of @RMorgan 's comments, to learn what's being discussed.

 

I see no reason to record higher than 48khz.

 

Wow you seem to be a lovley person. How about you follow your own advise?
I am out a here, have a nice day, Not in the mood for a smartass

 

Allright...Let's clear up that analogy, because it was a false one from the beginning.

When you start your mix in 44.1kHz and use a lot of coloration plugins, you know, those who generate harmonics, you're giving the less resolution to work with in the first place, thus they'll produce less precise results compared to analogue, which has infinite resolution.

Well, but does it matter if I'll mixdown to 44.1kHz anyway?

The answer is, it depends.

If you use only clean plugins which, getting back to the original op's subject, are not supposed to produce aliasing, then no. Go on and mix everything in 44.1kHz.

However, if you use a lot of analog emulations and plugins which generate distortion, such distortion will produce undesirable, atonal aliasing, which will bounce back throughout the whole audible spectrum. This effect is cumulative, so the more analog emulation you pile up on your tracks, the more aliasing you'll get.

Again, but does it matter if I will eventually mixdown to 44.1kHz, even if I mix everything up on higher sample rates?

In short. Yes, it does.

When you mix in higher resolutions and, like most people, use analog emulation plugins and/or other distortion plugins, they'll produce very little or no aliasing at all. You'll end up with a much cleaner mix, and most importantly, it will remain clean even when you downsample it to 44.1hKz...With one condition: You must first render it in higher resolution, then reduce it directly from this already "printed" track.

But hey, why does it matter? What's the difference?

When you downsample from a higher resolution, already summed up audio file, the "print" is already there. There are no plugins producing aliasing and piling up horrible artifacts as cumulative effect.

If you simply get into a 88.2kHz session and go like "save as 44.1kHz", then it will be the same as mixing it in 44.1kHz from the start, because it will downsample all individual tracks and their respective plugins during the mixdown/summing process.

That's why people usually send higher resolution files for mastering, and only then, during the mastering process, downsample to comercial standards...

Besides, you never know what the standard resolution will be like, 15 years from now...If all your cool songs are recorded and produced in 44.1kHz, you'll never be able to upsample them to a higher resolution. You can't put back information that was never there in the first place (unless they come up with some weird alien technology).

That's it.

 

Last edited: Oct 11, 2017

How so? What makes it a "false analogy"?

Analog gear does not have infinite resolution, not sure what you mean. If "something analogous to sample rate," it's less than 44k for most high-end analog gear (most don't bother plotting frequency response above 22kHz). If you mean bit depth, then we should bring up linearity/repeatability: number of distinct [air] pressure/voltage pairs the gear is capable of, unique mappings.

No one cares if an amp could produce an infinite number of random voltages. Just like no one cares about a DAC which can convert the same number to an infinite number of random voltages. And linearity? Noise? Anything but "infinite resolution."

As for the rest? Only if you do it wrong & ignore the first rule of digital: Filter everything above Nyquist limit. If your "analog emulator" plugs fail to filter @20k when running @44.1k, they're not "analog emulators" -- they're shit. Into the trash they go.

A question I never asked, but if you must.

Again, if the plugs you use ignore the first rule of digital &, for reasons known only to their creators, fail to filter teh part of the audio spectrum audible only to dogs, they will generate aliasing. There will be a couple of notes (guessing 1 to 3, without doing the math -- defined far fewer than an octave) where 44.1k will create aliasing & 48k won't. Is that your concern?

No simply *additive,* if "distortion grows faster than signal" is what you mean. And certainly not if you filter.

Again,

Point to me asking this ridiculous question, rub my nose in it. @TW (guy above you) thinks I can hear the difference between 44.1k & 96k sample rates, you answer questions I never asked... Let's never speak of it again.

Back on topic: Explain how sampling a pic @1.0884353...x pixels per line, and then downsampling the result to x pixels per line is fundamentally different from sampling a sound @1.0884353...x samples per second, and then downsampling the result to x samples per second.

Or care. I know [paying] canine demand for my bangerz won't increase much in 15 years, and h00man hearing won't change -- will top out ~20k. For children. Less for adults

 

Chandler Limited makes a software version of lil devil?...

also... soundtoys devil-loc is not setup as a compressor though... despite the fact that yes, the effect can reduce dynamic range.

Also, what are you implying with the Acoustic Audio link?..and how does it relate to this?

Or is it all the emu;s for nebula?... cuz I can't do dick with the free nebula...

 

Last edited: Oct 13, 2017