Rescuing Listenability Challenged Recordings.

[fas42]

Time to move this thought stream to its own space; it's all about an exercise, using computing power, in reversing the damage done to so many recordings over the last decade or so to a tolerable degree, to be available at least to the end users, us poor audiophiles. So if people would like to continue here ...

Frank

 

[JackD201]

Mastering Engineers and Archivists all over the world are doing this on a daily basis.

Are you asking if it can be done with the current "audiophile approved" tool set including DSP? Yes, but not in any way comparable to a fully equipped workstation manned by someone who actually knows how to use it.

 

[Phelonious Ponk]

I would like to see it start with how you're going to identify which zeros and ones have been altered by the digital compression process and how much they've been moved. Because if you can't do this, you're just creating another signal-degrading effect, nothing more.

Tim

 

[JackD201]

Have compressor plug-ins supplanted outboard analog compressors in most commercial studios for it to actually be an issue with the current crop of recordings? Somehow I am doubtful. Normalization on the other hand........

In a DAW environment, there's no need to find the zeros and ones. Since they are non-destructive, all you need is the original session file.

 

[Bruce B]

Have compressor plug-ins supplanted outboard analog compressors in most commercial studios for it to actually be an issue with the current crop of recordings? .

No

In a DAW environment, there's no need to find the zeros and ones. Since they are non-destructive, all you need is the original session file.

Yeah... that's the problem. No one HAS the original 2-track mixdown of most of the great music.

 

[fas42]

Mastering Engineers and Archivists all over the world are doing this on a daily basis.

Are you asking if it can be done with the current "audiophile approved" tool set including DSP? Yes, but not in any way comparable to a fully equipped workstation manned by someone who actually knows how to use it.

If it can be done with the right tool set, and the right mind set, then how come alternative, decent versions of these albums and music aren't out there for people to purchase. There are literally hundreds of bits of software out there that can mangl..., er, "enhance" the recording, but very, very little to undo to some degree damage done.

As Bruce says, this is no walk in the park, it's preeeety difficult. Which is not the same thing as impossible. The trouble is, the exercise is really a form of reverse engineering, working out from subtle clues within the final result what was done to it to get the end product. Once you have a decent handle on this then you can reverse the manipulation to a decent level of precision.

Just to emphasise that this sort of thing is my cup of tea, most of my creative computing years was enhancing and playing with 4th generation software development tools: programs that make programs. Most of the current software development tools are still very primitive, I thought the dark ages would have left us 10 years ago. I once started a major effort reverse engineering an obscure French 4GL, as they're called, which had zero documentation. And on the music front, my wife got, still has, a consumer Yamaha keyboard, a Portatone, with a propriety form of MIDI, tracks were stored on diskettes. I reverse engineered the storage format on the disks, and the structure of that music coding; my intent was to translate MIDI tracks to this propriety format, for my wife's benefit.

Frank

 

[JackD201]

Even with all the tools at their disposal, mastering engineers like Bruce can only do so much. Supposing the problem is a hot high hat. Bruce could tone it down some but it would also tone down everything else in that adjustment range. In the end he has to make a judgement call balancing between not usurping the original intent and what the customer would like to hear. That's if Bruce is lucky. Some customers I bet want the loud stuff and actually demand it.

Now we're talking on a track by track basis. I suspect Frank's intention is to somehow set this thread in the "if the power supply is behaving every recording will sound good" direction. My personal answer is there are no silver bullets. There will always be distortion in the recordings themselves and a lot of it is intentional. Leave the track by track stuff to the pros and just worry about our own systems not mangling the software we get any further. If accurate to a bad signal turns you off to what is otherwise a song worth listening to, EQ away.

 

[MylesBAstor]

Even with all the tools at their disposal, mastering engineers like Bruce can only do so much. Supposing the problem is a hot high hat. Bruce could tone it down some but it would also tone down everything else in that adjustment range. In the end he has to make a judgement call balancing between not usurping the original intent and what the customer would like to hear. That's if Bruce is lucky. Some customers I bet want the loud stuff and actually demand it.

Now we're talking on a track by track basis. I suspect Frank's intention is to somehow set this thread in the "if the power supply is behaving every recording will sound good" direction. My personal answer is there are no silver bullets. There will always be distortion in the recordings themselves and a lot of it is intentional. Leave the track by track stuff to the pros and just worry about our own systems not mangling the software we get any further. If accurate to a bad signal turns you off to what is otherwise a song worth listening to, EQ away.

Getting back to that pesky microphone

 

[fas42]

Even with all the tools at their disposal, mastering engineers like Bruce can only do so much. Supposing the problem is a hot high hat. Bruce could tone it down some but it would also tone down everything else in that adjustment range. In the end he has to make a judgement call balancing between not usurping the original intent and what the customer would like to hear. That's if Bruce is lucky. Some customers I bet want the loud stuff and actually demand it.

Now we're talking on a track by track basis. I suspect Frank's intention is to somehow set this thread in the "if the power supply is behaving every recording will sound good" direction. My personal answer is there are no silver bullets. There will always be distortion in the recordings themselves and a lot of it is intentional. Leave the track by track stuff to the pros and just worry about our own systems not mangling the software we get any further. If accurate to a bad signal turns you off to what is otherwise a song worth listening to, EQ away.

The trouble is, Jack, that the current, easily available tools aren't good enough. So it's up to other people, ahem, to improve the situation. And we are not talking EQ'ing, your high hat situation, that's falling off a log stuff. Or eliminating real distortion, that's a whole different, much, much harder kettle of fish.

No, not track by track. Split second by split second adjustments have to be made, there are some really gee whiz compressors out there that do very clever stuff, like dynamically altering their processing parameters on the fly, depending on what's coming through at every moment. It's getting easier and easier to really subtlely play with everything, which may be for the good, or may be for the bad. And, no, it's not about having a correctly working system: I want the recording to not sound like a group of 4 Hulk's mindlessly pounding away, which is how it comes across on the HT ...

EQ'ing will never do it, this is a very sloppy band aid, also applied by a Hulk. The only answer is to some degree unscramble the egg, which can be done using very careful and fiddly techniques. Now, who does that remind you of ... :-|)

Frank

 

[fas42]

Okay, folks, here's a first attempt at getting a much cleaner copy of "damaged" audio: something that's mentioned often as a "bad" example -- Californication, by Red Hot Chilli Peppers. Now, this is a real roughie, and I'm talking about my my first take on a cleanup effort, not the original track: I have done nothing to optimise or filter or EQ, or take extra care, so probably all sorts of faults that Bruce, say, will pick up. But it's a quickie that shows, gives a flavour of what can be done ...

The original is an MP3 file, so already we've lost quality, but it's a start. I got a snippet of about the worst area on the track, cleaned it up and added that after a bit of silence to the original. Then repeated the original (O) and fixed (F) another two times. So the sequence on the track is O - F - O - F - O - F, this makes it easier to get into a rhythm of picking what's been done ...

I've probably done the wrong thing by making the upload and download too large, through leaving it in .WAV form and doing repeats, you'll let me know! Anyway, if people like to have a listen and do some analysis, and let me me know the lot, bouquets and brickbats -- any feedback, I would appreciate it !!

As the first place that came to mind, I've put the file here: www.megaupload.com/?d=SSWP98S3

Thanks for listening ...

Frank