Kilohearts Plug-ins (My Misunderstanding)

[whyterabbyt]

It's a "Yes, I put you on the block list for arguing just for the sake of arguing". Goodbye.

Oh woe is me.

Thanks for the insight into the paucity of your argument, though.

[martinjuenke]

The whole thread zoundz wacky and limitlessssssssssssssssssss
I'll sample this noise, it will be my new killer hit!

[stoopicus]

There has to be a way to collect these bits and make them into a new generator in Phase Plant

[foosnark]

I'm shocked how few users here understand how summing works. It doesn't take many instances of those plugins to end up with a noise floor that ruins even a 16 bit integer signal.

This contradicts the first post in the thread. 64 of these plugins have a noise floor of -101dBFS. That's well below any "professional" threshold for an acceptable noise floor. As is the extreme example of 1024 plugin instances, at -77dBFS.

From the first post in the thread:

I understand some may say that these are not audible and therefore there's nothing to fix.

That's indeed what most of us say.

It's 2024, people (including "professionals") have been using plugins for quite some time now, and digital audio hardware longer still. Much of it with much worse specs than we're talking about here. Where are all the examples of ruined music?

[Igro]

And then, everybody cares which audio interface to get. I mean, if it doesn't matter, why bother then. Get it from AliExpress starting from 20 Euro. You can put in your studio and show it to your clients.

[Examigan]

I guess I have to delete all my ruined tracks.

[mothra]

OP has a point as Kilohearts is designed to stack multiple plugins in series with extreme settings, this is different that using one eq however I think fabfilter should also upgrade to 64float. I am not saying you can always hear the difference but its stupid to say its just marketing, if anything the plugin companies saying it does not matter is marketing. Hopefully OP updates this thread if Killohearts reply's back.

I don't know, I think I'm more inclined to go with the word of the coders who do this for a living and actually spent their time testing it out than an armchair non-dev from the internet who 'knows'..

[jrwaltb]

I mean Ableton and Cubase both rewrote their audio engines and they said the same thing as these vendors are saying now before it was ready and while they were developing it. No ones is saying Phaseplant is bad or sounds bad its just advancing to 64-bit would be better for heavy snapin use.

[WackyZoundz]

64 of these plugins have a noise floor of -101dBFS.

Yes, but only if you never use makeup gain after compression. Or if you never use soft clippers for gain. Or if you stay away from any distortion (as a special effect). Or if you avoid any type of parallel processing (at least +6 dB to the noise floor because it's a coherent signal) and so on. And don't forget fields where 24 bit integer is required like commercial sample libraries, film, video games or high-dynamic range music formats (Audio DVDs, FLACs for audiophiles etc). Just 5 Kilohearts plugins in a row on one channel (without makeup gain after compression of course) and you can forget all of that. The minimum for digital audio should be no less than 32 bit integer so it can cover all kinds of situations.

Just try it yourself. You will see how fast the noise goes past the 16 bit integer threshold with an average Pop project which needs plenty of compression and other things. Heck, I had vocal chains which featured up to 16 plugins alone and that's just one channel.

Sure, noise can be removed afterwards but that introduces new artifacts which you have to mask with dithering again with rather poor results because there is still no dither type that can handle the mess FIR filters leave behind. So it's better to work clean so you don't have to waste time and resources on repairing a broken signal.

[WackyZoundz]

And then, everybody cares which audio interface to get. I mean, if it doesn't matter, why bother then. Get it from AliExpress starting from 20 Euro. You can put in your studio and show it to your clients.

Which is reasonable for bedroom producers. A good audio interface is a waste of money if you don't know how to use it nor how to deal with software nor if you don't have the required acoustic treatment to record or mix properly. Everything has to be redone in a studio anyway if you're aiming at good results no matter if it's just a hobby or commercial.

[WackyZoundz]

Again: A chain of 16 plugins isn't much. Add 18 dB of gain with compression and highshelf filters and the noise is above -96 dBFS already - for a single channel. That's something you'd expect from freeware programmed in 1998.

[limitlesssss]

OP has a point as Kilohearts is designed to stack multiple plugins in series with extreme settings, this is different that using one eq however I think fabfilter should also upgrade to 64float. I am not saying you can always hear the difference but its stupid to say its just marketing, if anything the plugin companies saying it does not matter is marketing. Hopefully OP updates this thread if Killohearts reply's back.

I don't know, I think I'm more inclined to go with the word of the coders who do this for a living and actually spent their time testing it out than an armchair non-dev from the internet who 'knows'..

I'm still waiting for the developers to reply to my email.

[FrogsInPants]

If I set up a chain of 32 kHz Gain plugins, half +3dB and half -3dB as suggested, I see a bit of signal peeking through at -120dB, and noise down around -170dB. That's with an input signal at -12dB. If I reduce the volume of the input by 24dB, the entire output also drops by 24dB. So this isn't introducing a constant noise floor; it's proportional to the input signal. Sure, you can easily apply a huge gain or distortion to bring the output of the null test up to audible levels, but can you make it audible in the presence of the actual signal? With the noise proportional to the input signal, something like make-up gain on a compressor certainly isn't going to do it.

I would be delighted to learn how this can actually become significant. In contrast to something that introduces a noise floor independent of the signal level, I'm just not seeing it outside of niche conditions like a null test or extreme numbers of chained plugins. But I'm no expert.

[limitlesssss]

All I'm asking is, why should there be a noise floor to begin with in a purely digital signal that is being operated in a 64-Bit FP DAW environment? Is it dithering noise? I doubt it. In a floating point environment dithering is odd. What is it? Why can I perfectly null a duplicated track when I flip the polarity of one of the tracks, but as soon as I insert even *one* Kilohearts Gain plug-in (untouched) on one of the tracks, all of a sudden the noise floor of the null raises to -150dBFS? Why? What is it? I want to know. What is going on behind the scene of these plug-ins? How about when I insert two Kilohearts Gain plug-ins on one of the tracks, one of which is cutting exactly 3dB and the other is boosting exactly 3dB? Shouldn't the result of the null be absolute silence instead of -129dBFS? Why isn't it absolute silence? What is going on behind the scene?

Is that nitpicking? Okay, I understand. Those who don't care because they're busy releasing hit records after hit records are welcome to ignore this thread.

[csscp]

All I'm asking is, why should there be a noise floor to begin with in a purely digital signal that is being operated in a 64-Bit FP DAW environment? Is it dithering noise? I doubt it. In a floating point environment dithering is odd. What is it? Why can I perfectly null a duplicated track when I flip the polarity of one of the tracks, but as soon as I insert even *one* Kilohearts Gain plug-in (untouched) on one of the tracks, all of a sudden the noise floor of the null raises to -150dBFS? Why? What is it? I want to know. What is going on behind the scene of these plug-ins? How about when I insert two Kilohearts Gain plug-ins on one of the tracks, one of which is cutting exactly 3dB and the other is boosting exactly 3dB? Shouldn't the result of the null be absolute silence instead of -129dBFS? Why isn't it absolute silence? What is going on behind the scene?

Is that nitpicking? Okay, I understand. Those who don't care because they're busy releasing hit records after hit records are welcome to ignore this thread.

These are quite valid questions and as an intellectual curiosity they are valuable. However, in my humble opinion, answers to some of these questions would require a mental set, a wealth of hands-on experience with digital signal processing, and nitpicks of common digital signal processing frameworks. That being said, I'd humbly recommend not obsessing over these questions and not let them cloud your music making process.