Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Manual dynamically enriching remixing (Read 3545 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Manual dynamically enriching remixing

Hello folks,

Recently I've found out that for some reason I like the song "Poker Face" by "Lady GaGa" pretty much. Unfortunately it's mastered badly (which really was expectable...). So I've tried my best to increase it's dynamic range by overlaying the drums with dynamic sounds. It's a lot of work, but I find it much more listenable now.

What do you think about it? I personally think that's the only possibility to really improve the sound of a clippressed recording like this one.

What I really want is a discussion whether it's worth doing something like this, and whether you think the result sounds better than the original.

Maybe somehow the process could be automated (IIRC there are already algorithms for finding rhythms and so...). I wouldn't hesitate to write a program for that if it'd not be too complicated for me

Cheers!

Moderation: Removed potential TOS #9 violation.
flac 1.2.1 -8 (archive) | aoTuVb5.7 -q 4 (pc, s1mp3)

Manual dynamically enriching remixing

Reply #1 –
While I'm not sure if the specific way you've remixed this track is something I find to be an improvement overall, the basic concept itself is a pretty sound one. You'll find that many mixers and editors use this technique for film trailers, overlaying various transient booms or snaps to an existing track to add a certain level of tension or to otherwise improve upon the track's dynamics (when required). In fact, this practice has become quite commonplace in the past five years or so.

As for designing an algorithm to do this, I'm not certain. The 'trick', if there exists one, would be to overlay transient sounds that are similar in nature to those in the original mix or sounds that are otherwise lacking in any kind of definable character. Because drum tracks are often processed heavily, especially in pop, industrial and other genres, designing an algorithm that can not only detect appropriate transients but choose 'appropriate' sounds to overlay would be especially difficult. Unless you have a generic set of very short, probably synthesized transient sounds, and the algorithm is accurate enough to lay them in precisely (a difficult process considering how heavily-compressed music can be -- how well can any algorithm find transients in such a mess?), I can't foresee this working acceptably.

That being said, it would be a noble endeavor to try

Manual dynamically enriching remixing

Reply #2 –
I've thought the same before. For percussion, it seems to me as though multiplying a single sample directly on the beat by some factor above 1 could produce listenable results. That way you get an impulse at precisely the right time. Maybe not though, I've never tested this.
 

Manual dynamically enriching remixing

Reply #3 –
Quote from: Ron Jones on
While I'm not sure if the specific way you've remixed this track is something I find to be an improvement overall,

Okay, maybe I'm just not good enough at choosing the overlay sounds. 
Quote from: Ron Jones on
the basic concept itself is a pretty sound one. You'll find that many mixers and editors use this technique for film trailers, overlaying various transient booms or snaps to an existing track to add a certain level of tension or to otherwise improve upon the track's dynamics (when required). In fact, this practice has become quite commonplace in the past five years or so.

So it's nothing new... Okay.

Quote from: Ron Jones on
As for designing an algorithm to do this, I'm not certain. The 'trick', if there exists one, would be to overlay transient sounds that are similar in nature to those in the original mix or sounds that are otherwise lacking in any kind of definable character. Because drum tracks are often processed heavily, especially in pop, industrial and other genres, designing an algorithm that can not only detect appropriate transients but choose 'appropriate' sounds to overlay would be especially difficult. Unless you have a generic set of very short, probably synthesized transient sounds, and the algorithm is accurate enough to lay them in precisely (a difficult process considering how heavily-compressed music can be -- how well can any algorithm find transients in such a mess?), I can't foresee this working acceptably.

That discourages me. I understand the reasons given. It's already hard for me as a human to find the rhythm when I look at this heavily compressed waveform. On a dynamic waveform, you can see sharp peaks; these could easily be detected automatically, but in this mess... too bad. Filtering out WHAT has to be boosted (i.e. automatically finding overlay sounds) is even more difficult, right.

Quote from: Ron Jones on
That being said, it would be a noble endeavor to try

Yes, but somehow I don't really know where to start at all...

Quote from: Canar on
I've thought the same before. For percussion, it seems to me as though multiplying a single sample directly on the beat by some factor above 1 could produce listenable results. That way you get an impulse at precisely the right time. Maybe not though, I've never tested this.

I've tried that, too. I've found that doesn't work well, because while the percussion is going on, there are other sounds playing at the same time, which are then increased in loudness, too. This sounds messy (just think of a singer whose voice rhythmically gets louder and quiter... ugly...). That would likely be much easier to automate, but given that the results aren't good in my opinion, I couldn't be bothered to try it.

Thanks to the moderation for acting correctly. My fault, sorry. I'll remember!
flac 1.2.1 -8 (archive) | aoTuVb5.7 -q 4 (pc, s1mp3)