Sneaky declick update

[Overture]

Ahoy all,

I've reached my first little milestone on the declicking - a decent removal.

Currently implementing it in a simulation environment but the algorithm should let me do it in real-time in a processor, or if not completely real-time, at least less than 10ms of delay. So far, anyway.
Next will be testing it on larger passages, working out which variables will be adjustable with the strength knob, to what degree, and then I'm going to have a crack at de-crackle.
This is still early days but it's promising.

The original tiny snippet with the click and the removal is attached. Here's a visual as well:

Bonus points if you can guess what album it's from. Not you, Chucky, you know the answer.

Cheers
~D (Mrs)

 

[chucky3042]

Ahoy all,

I've reached my first little milestone on the declicking - a decent removal.

Currently implementing it in a simulation environment but the algorithm should let me do it in real-time in a processor, or if not completely real-time, at least less than 10ms of delay. So far, anyway.
Next will be testing it on larger passages, working out which variables will be adjustable with the strength knob, to what degree, and then I'm going to have a crack at de-crackle.
This is still early days but it's promising.

The original tiny snippet with the click and the removal is attached. Here's a visual as well:

View attachment 110076


Bonus points if you can guess what album it's from. Not you, Chucky, you know the answer.

Cheers
~D (Mrs)

I do know the answer but my lips are sealed. I claim its a 100% removal, not bad for 1.5 years work I say.

 

[Overture]

I do know the answer but my lips are sealed. I claim its a 100% removal, not bad for 1.5 years work I say.

With one or two several-month distractions in the middle.

 

[chucky3042]

With one or two several-month distractions in the middle.

Shiny things

 

[Overture]

Shiny things

Oooh, shiny!

 

[DuncanS]

Ahoy all,

I've reached my first little milestone on the declicking - a decent removal.

Currently implementing it in a simulation environment but the algorithm should let me do it in real-time in a processor, or if not completely real-time, at least less than 10ms of delay. So far, anyway.
Next will be testing it on larger passages, working out which variables will be adjustable with the strength knob, to what degree, and then I'm going to have a crack at de-crackle.
This is still early days but it's promising.

The original tiny snippet with the click and the removal is attached. Here's a visual as well:

View attachment 110076


Bonus points if you can guess what album it's from. Not you, Chucky, you know the answer.

Cheers
~D (Mrs)

Impressive!

 

[J. PUPSTER]

Ahoy all,

I've reached my first little milestone on the declicking - a decent removal.

Currently implementing it in a simulation environment but the algorithm should let me do it in real-time in a processor, or if not completely real-time, at least less than 10ms of delay. So far, anyway.
Next will be testing it on larger passages, working out which variables will be adjustable with the strength knob, to what degree, and then I'm going to have a crack at de-crackle.
This is still early days but it's promising.

The original tiny snippet with the click and the removal is attached. Here's a visual as well:

View attachment 110076


Bonus points if you can guess what album it's from. Not you, Chucky, you know the answer.

Cheers
~D (Mrs)

Is the de-crackle mainly a similar but more sensitive process than de-click?

When I run my SM Quad decodes & Marantz demods thru RX 8; it takes about 4-5 minutes per channel on de-click, but approximately 20 min. for de-crackle.

Would be great to have that in real time; just hope it’s not another 1 1/2- 2 years before finished product arrives 😬

 

[chucky3042]

Is the de-crackle mainly a similar but more sensitive process than de-click?

When I run my SM Quad decodes & Marantz demods thru RX 8; it takes about 4-5 minutes per channel on de-click, but approximately 20 min. for de-crackle.

Would be great to have that in real time; just hope it’s not another 1 1/2- 2 years before finished product arrives 😬

Its a slightly different process complicated by psychoacoustic masking issues meaning the crackle is masked by higher amplitude stuff in the same spectral band but is more apparent in quiet passages. The detection is slightly different also. I also hope its not another 1.5 years but things are really tight on budget issues right now but I hope to crawl out of it.

 

[J. PUPSTER]

Its a slightly different process complicated by psychoacoustic masking issues meaning the crackle is masked by higher amplitude stuff in the same spectral band but is more apparent in quiet passages. The detection is slightly different also. I also hope its not another 1.5 years but things are really tight on budget issues right now but I hope to crawl out of it.

well I'm rooting for you guys, at least you can pencil me in for a SMv4 if it has those features

the de-crackle process almost sounds like mythical snipe hunting

 

[Overture]

Is the de-crackle mainly a similar but more sensitive process than de-click?

When I run my SM Quad decodes & Marantz demods thru RX 8; it takes about 4-5 minutes per channel on de-click, but approximately 20 min. for de-crackle.

Would be great to have that in real time; just hope it’s not another 1 1/2- 2 years before finished product arrives 😬

Only 1.4 years, tops.