I have some old recordings that were encoded using very early versions of mp3 (and mp2!) encoders that have obvious flaws, the most notable of which is smeared transients (sounds somewhat like 50-100ms of preverb). In some cases I probably still have master recordings but both the 1/4" and DAT are broken at the moment…
I’m confident that the “deconvolution-adjacent” technology embodied in SpectraLayers could be trained to undo this effect, but I don’t know if the current version can. Has anyone dealt with this particular use case yet?
Not expecting or demanding miracles.
I have not, but you should definately take a look at the two plugins Unchirp and Punch from Zynaptiq if you haven’t.
They address artefacts from badly coded mp3:s and overly compressed audio respectively.
1 Like
Thanks! I’ll definitely give it a try.
Folks selling stuff like Unchirp should IMO adopt the REAPER pricing scheme, don’t want your money for a month, then pick a user category/price, and nagware if you never do. I’m extremely unlikely to pay $300 for it regardless of whether it’s magic, but if it was $75 they would sell it 60 seconds after I successfully demoed it.
Yes I realize this bears no resemblance to any Steinberg pricing model, ok, whatever. 
I don’t have any one good reason to spend money on SL Pro but I have enough semi-reasons that I’ll probably buy it soon. I’m curious if de-reverb has any useful effect in this situation.
(To be clear, I think SL Pro is worth the $ for sure, Unchirp, on the other hand, not an obvious value proposition for me.)
SL is worth the money imho.
… but the VST-interface, besides other things, needs a lot of improvements; technically from a pure mathematical/algorithmic viewpoint and ergonomically. SL today feels, at least to me, like a good beta version of something great to become; like a work horse of a very creative and ingenious (no question about that) developer.
I consider SL as a piece of majestic art to work with sound and music. The concept of spectral layers is second to none.
Some decades ago, I had the idea of working in the spectrum like in Photoshop and SL does just that (or at least points into this direction).
But, and this is a very big “but”, spectral editing seems still in its infancy as it is today.
It might take another 5, 10 oder 20 years, until it is as failsafe and reliable (from a mathematical point of view), as modern picture editing is today.
1 Like
Yeah, I know the Zynaptiq tools are pricey. They do have sales now and then, so if one can wait there will be a possibility to get them at considerably lower price. (I may seem like a Zynaptiq rep, but I ain’t).
They also have a de-reverb plugin - Unveil - and while you don’t exactly get unreal results it is indeed very good at what it does.
I use it as a VST3 plugin in combination with the built in SL de-reverb with mostly good results.
De-reverb in SL is nice and does what it should; with moderate settings, and a little more, I hear no significant artefacts.
Unveil gives more control and can add a bit contour or clarity at certain settings.
Those plugins work very well together with other in a chain in SpectraLayers’ VST3 function to restore old recordings. Except for Punch that is, in the current version it crashes SL.
The VST3 function in SL is offline with a realtime preview possibility and is really nice to have at hand when restoring audio.
1 Like
This looks promising for low bitrate audio, I think it’s been implemented in the latest UVR5 beta…
2 Likes
Wasn’t aware of UVR … I’ll give that a look too.