Is there a way to "smear" audio in the time domain?

[whyterabbyt]

There are a lot of spectral/FFT solutions here, but the OP was asking for time-domain solutions.

Where did they say that?

[mjolnir]

I'd say, emulate the effect you're going for with the best tools you can find to do it.
Maybe do a regular preverb with regular reverse and reverb and then time expand that and save it. There you go, boom.
On to bigger and better things (instead of trying to find the perfect tool which might not exist).

[VariKusBrainZ]

Wouldnt that just be a standard time stretch then

[CinningBao]

There are a lot of spectral/FFT solutions here, but the OP was asking for time-domain solutions.

Where did they say that?

literally in the subject title!
that's why i said that

[BertKoor]

But there's also this remark by the OP:

I don't really know what I'm talking about

I think what he meant by the title is that the smear should be in time (transients getting less sharp) and not in frequency (pitch getting less clear)

[whyterabbyt]

But there's also this remark by the OP:

I don't really know what I'm talking about

exactly... he wasnt sure, and his reference for what he was after was specifically an FFT process, but he just seemed to be assuming that wouldnt work on audio.

Imagine you are looking at a spectrogram. Then increase the FFT size. The frequency resolution gets better and the time resolution gets worse. This results in the display of every audio "event" getting smeared, or elongated, left-to-right (in time).

Is there a way to actually process audio in this way

[osiris]

Is this time based granulization? (possibly not a real word) I used to extreme stretch voices with FL Granulizer and this old VST from GRM Tools, but I can't remember what it was called.

[David Kissam]

I think what he meant by the title is that the smear should be in time (transients getting less sharp) and not in frequency (pitch getting less clear)

Yes, that's right.

I am working on samples for a character piano VI that I am making. Some of the samples have buzzy noises in the 4-10kHz range directly after the transient. These noises are partly what I like about the piano, but some are too ugly and must be fixed. I have been doing this simply by attenuating the noise - RX's Deconstruct module does a good job at this.

But then I thought, instead of just attenuating, what if I could also change the character of the noise to make it less objectionable? As it is, the noises have these little sharp sounds in them - little tiny transients, I guess you could say. If I subtly smeared the entire frequency band where the noises are, it would not hurt the actual harmonics of the piano, as those change relatively slowly in time anyway, but it would soften those sharp sounds which are the most unpleasant.

I maybe should have said all that in my first post, but the replies I've gotten have been quite helpful anyway. Thanks again!

I tried DtBlkFX's Smear effect and that was at least in the ballpark. It does not report its latency to the host, however, which makes it unusable for me because it moves the audio forward in time. Ideally, I need something I can apply to a selection in RX.

[David Kissam]

Oh, and it's possible that one of the tools suggested here already is exactly what I need, but I haven't tried them all yet. What I'm trying to do is a fairly subtle change, not a dramatic effect.

As I understand it, FFT works by analyzing the frequency content of blocks of audio. Larger blocks give you more detailed information about frequency, but you lose detailed information about timing - you end up with an average level of each frequency for that block. A spectrogram with a high FFT setting cannot show rapid changes in amplitude.

And that's what I'm after - to "average out" or simply slow down the changes in amplitude of small frequency bands. More precisely, I want to reduce the rate of change in amplitude of small frequency bands.

[cryophonik]

If I understand the question, it sounds like maybe you want a comb filter that notches out a lot of frequencies with a very narrow bandwidth, then backfill those notches with another filter (with a wider bandwidth) that boosts the notched areas with an average of the frequency bands on either side of the notch. Maybe? Or, maybe I need another coffee

[eLawnMust]

Easier to 'smear' Kaka on a wall...

[David Kissam]

If I understand the question, it sounds like maybe you want a comb filter that notches out a lot of frequencies with a very narrow bandwidth, then backfill those notches with another filter (with a wider bandwidth) that boosts the notched areas with an average of the frequency bands on either side of the notch. Maybe? Or, maybe I need another coffee

Hmm, that sounds like an average across frequencies, but I'm looking for an average across time.

[David Kissam]

Easier to 'smear' Kaka on a wall...

That would be easier, but I don't think I could put up with the smell.

[cryophonik]

If I understand the question, it sounds like maybe you want a comb filter that notches out a lot of frequencies with a very narrow bandwidth, then backfill those notches with another filter (with a wider bandwidth) that boosts the notched areas with an average of the frequency bands on either side of the notch. Maybe? Or, maybe I need another coffee

Hmm, that sounds like an average across frequencies, but I'm looking for an average across time.

OK, got it. Yeah, my approach would manifest itself in playback over time, but you're thinking that frequencies should be smeared by an average (or whatever) of the frequencies across the spectrum that precede and follow it, sorta analogous to a running average. Hmmmmm....

[mjolnir]

Wouldnt that just be a standard time stretch then

who cares what is is, if it sounds alright tho.