Remastering Plugin Tools

Greetings … I have found out about some pages on the web that Remaster music however some of them offer the type of tone like Clarity, Bass, Treble etc. but I found a page that Automatically did everything, it was Loud and clear without Distortions, the bad news I just got a little frame from the song and I have to pay for the full song, so I would like to know if there is a Process to do this on Audacity to remaster on that High Quality all Loud , clear without distortions or Sibilances all Loud and Original Transparent at the Remastering Process, like if those shrink waveforms from old songs were New. Please Thanks

I think you’re looking for the mythical “professional audio filter”, which, if it existed, would put all audio-engineers out of work.

There are levelator type filters, e.g., which will make the audio more uniformly loud.
It is also be possible to match the overall equalization to (approximately) match some reference.
DeEsser plugins are available to reduce sibilance, e.g., but they are designed to work on the vocal track in isolation, (without the musical instruments).

It would be possible to chain some of those processes together in Audacity, to automate remastering.

I think you’re looking for the mythical “professional audio filter”, which, if it existed, would put all audio-engineers out of work.

What was the posting date on that web page? April First by any chance? The last one I posted could turn any voice track in any location or condition into a studio production. The best example was filtering out a Harley on Venice Blvd in Mar Vista.

I gotta see if I can find those. They’re on the Dark Web. That means I don’t remember where I put them.


I found a page

Can you find it again? Can you post it?


The Pages are , and, the Page One I got the sample was from MaximalSound. I like the resolution from the Sound was so High and Clear the other Pages offers you how could be master the music if adding Bass or Trebble, bright, Stereo many effects like this etc…

I found one …
From the landr examples shown, it’s compressing, equalizing, adding stereo width, & maybe de-clicking.
(It’s not de-essing though).

Do You know if this sites adds Bass Boost to the Equalization or how they do this adds Trebble too …

Landr has different templates, the R&B one is compressing, limiting, boosting treble* and widening stereo
Before-After Landr's R&B processing on ''Freakazoid''.gif
It’s not de-clicking though.

[ * some of the treble-boost is a side-effect of using a limiter which has more effect on the bass ].

If you are looking to remaster a commercial release, I have found that it is best to try to bring it back down to pristine. Luckily, nowadays much older music (e.g. from HDtracks and other good sources) sell recordings that are ‘close to’ DolbyA encoded. I think that they just might be running off master tapes without decoding? This means that a reasonable sidechain type expander (like my DolbyA software) can remove alot of the compression artifacts. I have a realtime DolbyA decoder that does things like taking the overly sibilant, bouncy high frequencies (from compression) and converts into the music that I remember when I was young. Doing a DolbyA decode isn’t as simple as DolbyB/C, and the decoder is scarce as hens’ teeth (euphemism.) I just wrote/tested/verified a software decoder that works in single L+R mode, L+R&M+S mode, and also L+R,M+S,inbetween mode. Apparently alot ot the recordings had been created like that so that there wouldn’t be worse hiss when using a quad matrix.
So, what I would suggest is: repair in a neutral way, re-equalize, and the re-master by using a minimally intrusive finalizer of some kind. I have been developing such technology (e.g. my finalizer has been so unintrusive that I had to make mods to make the compression more obvious.) I am not suggesting my work product, but do know that it is problematic unless expanding it to remove the commercial compression/limiting. (Limiting is very problematical, but compression can often be partially undone.
I have been working on remastering some music from an old, very big group – and their stuff came VERY DolbyA encoded…
Again, much older music out there is either DolbyA encoded or uses lots of fast sidechain compression. A good example of such is ‘Shake it off’. My analysis shows that it has been compressed (one way or other – not necessarily real DolbyA) by about 6 passes of 3 axis DolbyA compression (yes, it works as a very intense compressor, not just noise reduction.) I ran ‘Shake it off’ through 6 passes of my 3phase DolbyA software, and finally could detect a relatively clean background.

Good luck/Best Wishes

What’s “lucky” about that? As you said yourself, software Dolby A decoders are rare, so it would certainly not be “Greatest-Sounding Music Downloads”, as HDtracks advertise, it would actually sound pretty bad.

(Dolby Laboratories only ever made hardware decoders for Dolby A, though for a around $100 an emulator is available from U-He).

DolbyA appears to sound like a compressed, HF intense version of ideal. Frankly, it doesn’t seem all that much worse than DolbyC – just compressing LF, MF and HF in two different attack/decay rate pairs. The attack/release times aren’t bad enough to terribly intermodulate with the audio (it does a bit, but gets undone by decoding.) The release times (different for HF and LF) are somewhere around a constant 50msec – faster than desired for ‘normal compression’ , but not terrible. The attack time is nonlinear, so cant really be specified as as simple value – it gets faster as the signal level difference increases. So, slow moving signal levels get attacked more slowly than a simple RC type attack timeconstant. The max compression depth for a signal between -40dB and 0dB is approx 10dB – so it isn’t like a hard limiter. The compression ratio is odd – it slides around because basically DolbyA is a compressor/expander designed based upon existing components and not apparently designed from spec. If you look at it, it looks like a modified typical JFET compressor. Dolby even uses the now well-known 50% feedback between the JFETs gate and the drain/source (providing significant distortion reduction.) It is a pretty damned good design for 1965 timeframe!!!

So, yes DolbyA doesn’t sound pretty – but the overly hard (and boingy HF) in alot of old releases is what I’d expect from the schematics and the specs. Since I did a massive iterative attempt to decode some stuff (literally 100’s or 1000’s of iterations) because the specs are incomplete and the schematic mostly gives a general concept (and since I have done alot of similar compressors, I know approximately where to start.)

I am NOT 100% claiming DolbyA, but I do know that a DolbyA style sidechain expander undoes ALOT of the nastiness that I have heard in alot of old releases.

John Dyson

Yeah, Im Looking for a Natural Remastering like I said before I must care about Quality more than Loudness or adding effects, But the Natural Loud

The killer problem with that ‘remastered’ version is that it is so very compressed/limited, there is not much dynamic range EVEN for an expander (a kind of partial correction for a compressor) to be able to work with. Music that is so very compressed/limited like that just does NOT make sense. To give you an idea as to how much intent there was to totally destroy music – I actually had to make my finalizer (software that I wrote) be very aggressive to even start looking like that. The goal of my finalizer is to bring the dynamic range into reasonable bounds, not to totally drop the dynamic range to nothing.
In my experiments, there are at least two kinds of dynamic range – long term, and short term. To me, it appears that they bottled the music all the way up for long term and short term. My finalizer looks to tightening up the short term and controls the long term such that it still has audible ‘dynamic range’ and even works VERY WELL and if adjusted, not even very audibly – but electronically the audio has perhaps a 20dB range of levels instead of 30-35dB (that is, for orchestral.) For popular, you can tie the long term up pretty well, but the short term is critical for musicality (IMO.) Of course, in the long term you don’t want the background to come up too high, but in the short term YOU DON"T WANT TO FILL IN EVERY SYLLABLE – and if filling in, leave a few dB for the hint of dynamic range!!!
That mess that you showed really upsets me, because it is easy (with some math/dsp tricks) to totally smash the music, but it isn’t music any more. It requires skill and carefully designed technology to create a good trade-off of good/pleasant dynamic range within the capability of the listening environment and/or media dynamic range.
People designing audio processors (or post processing music with them) ARE NOT the musicians/artist that people want to listen to…
THINK ABOUT THIS: when is the last time that someone thought – I like the sound of the XXXYYYS processing, and I want to purchase music processed by that person?
People want to listen to music, NOT PROCESSING!!!

John Dyson

Phil Spector
George Martin
Joe Meek
Brian Eno
Brian Wilson
Dr. Dre
Lee Scratch Perry
Pharrell Williams and Chad Hugo
Jim Steinman
Rick Rubin
to name a few.

And Joe Boyd :sunglasses:

Look – I wasn’t talking about those people on the early part of the real creative process – I am talking about the person who takes the 20-25dB long term dynamic range tape (still very compressed – like ABBA), and then changes that into a 6dB or less dynamic range. People that you are talking about are on the FRONT END of the process, and probably leave some dynamic range. It is those people who take those tapes (source material now) and destroy that. I have the luck of hearing some of ABBA 2ch master tapes – and I have heard a few publically released materials that come close to it. The originals are still pretty well compressed, but… But the normal case is more like my recently purchased ABBA CD that is totally destroyed. They even did a matrix shift to subvert an intentional harsh chorus effect. (the M+S version isn’t as compressed sounding as the normal L+R – so they shifted to the M+S!!!) If they wanted to ‘fix’ that, then a proper expansion would have helped a lot instead. In fact, on that CD I heard numerous matrixing games that brought certain components up to the foreground. Still, after those games, the 18-25dB instantaneous dynamic range had been smashed into a 12-13dB peak to RMS ratio. That CD is so ugly as to be unlistenable. On the other hand, I heard (AFAIR) a Polar CD that didn’t even have the common severe intermod in Super Trouper (at the beginning of the song, there is an extreme, extreme roughness – that sounds to me like someone did alot of fast compression but didn’t keep the LF products from being produced… I can make that kind of terrible distortion also, but avoid it like the plague.) I think that happens when the two female voices are mixing together with a very fast compressor/limiter causing modulation (intermodulation) products.

I have heard several sources before they were damaged by the final production of the distributed version – and the difference is amazing, and actually the non-broken (non-modified) versions were very good, but the resulting distributed versions sucked badly to the point that I won’t even listen because it is so ugly and loud.

I hope that you aren’t saying that you are even aware of the person who destroyed the sound quality before producing the released version? I know - ABBA is really a special case where the listeners are so desperate for any reasonable quality that they do HAPPEN to know the names of those people last touching the recordings. I don’t know much about the CD that I have, other than it is so ugly as to be TOTALLY unlistenable. I wish I know who destroyed that – or maybe the reason why he/she was motivated to do so.

John Dyson

I have a REALLY GOOD DolbyA decoder that really works, and been professionally compared to be generally superior and more desirable than anything else including a real DolbyA unit. It isn’t perfect, but nothing is. It is available for Windows with recent CPUS and I have a slightly older Linux version (which is only because of demand, all development is done on Linux – windows is too slow and clunky – even today.) It incredibly closely emulates a real DolbyA DECODE OPERATION ONLY – to a point. It goes beyond anything else for intermodulation avoidance and removal (not simple attack/decay – quite sophisticated), so that complex material comes through clean with a more similar tonal balance to a real DolbyA. (All tested/verified.)
It is available on my current repository (updated fairly often as soon as any negative feedback or improvements that I find):
filename:, e.g.

There are some decoding examples online in that repository. It really works, and works well. Too bad it isn’t a plugin (yet.) It will suck up an entire CPU (Haswell 1 core) at realtime at 124ksamples/per sec, and is slightly faster than realtime at 96k, runs as slow as 44.1k, but likes 48k or better. Slower version works on recent ATOMS. Both are in the distribution. There is a ‘DecoderA.pdf’ file that describes command line use. The reason for such large CPU usage is that it is NOT a trivial piece of software, and does all kinds of intermod avoidance so that complex choruses and things like that come through more clearly. It can run more quickly (command line options), but only makes it have the intermod of most other kinds of decoder.


I’m getting:

$ ./da-avx -thresh=-15.50 input.wav output.wav
LEFT-RIGHT dBthresh: (-15.50,-15.50), dBingain: (  0.00,  0.00), dBoutgain: (  0.00,  0.00)
Cannot do audio input from tty
Failure to start -- input file is incompatible

The problem that you are having is exactly the kind of problem that happens because I am too stupid to write a GUI based tool (or plugin) instead of command line… Here is the mistake, and i also have a hint (at least the program started for you – that is good!!!)

The command that you want to issue is this:

./da-avx --thresh=-15.50 <input.wav >output.wav

Here are some hints: note the two dashes in front of ‘thresh’. Note the less than ‘<’ sign in front of the input file and the greater than ‘>’ sign in front of the output file. Each of those are critically important (and admittedly arcane) punctuation.

One more ‘extra’ hint: use the --info (that is two dashes again) switch with a space right after the command name for a bit of a ‘bordom minimizer’ – it gives a second by second running log of the input/output levels and the gains for each channel – the gains are (min, avg, max), because, of course, the gain can move a long way during a second.

One more ‘tuning’ hint is that adding on the --ai=high would slightly improve the quality if there is a complex chorus or something like that in the music. I’ll probably make the equivalent of --ai=high the default some time soon. I haven’t found any disadvantages other than the program runs perhaps 10-20% slower. I have recently made a minor speed improvement that cancelled the slowdown. I am not really ‘suggesting’ the --ai=high simply because it complicates the command and isn’t absolutely needed.

So, my suggested command with the ‘bordom minimizer’ enabled would be:

./da-avx --info --thresh=-15.50 <input.wav >output.wav

Let me know if you have further problems. I have been strongly suggested to produce a plug-in version – but it WILL suck up an entire core of a CPU when running realtime at 96k.


You need to update your documentation as that shows only one hyphen.

That should be made explicit in your documentation because it is very common for chevrons to be used to indicate that the enclose string is an example / place-holder text.
A more conventional syntax would be to use “-i” and “-o” switches.

A nice enhancement to the command-line tool would be to implement a “–help” / “-h” switch.