Why Don't You Remix New VGM?

[Nase]

My apologies: my music vocab is almost non-existent. Detailed high resolution images can be downsampled, causing pixelation. Downsample to a small enough resolution and (as a fun side effect!) you're left with a 'core' pattern, showing only the most common pixel colours used (or average colours).

Is there an equivalent procedure to open a track from a modern game (an orchestral or 'fleshed out'/complex track)... downsample it somehow... then come out with a more 'pixelated' pattern? Wouldn't that be a good way to more easily identify the melody in otherwise complex music? From working off such a core/pixelated version of a modern track, it'd be easier to remix maybe?

Fun analogy. Though what you're left with by downing the resolution is just a very basic piece of color mood. To preserve the essence, form and information in a lo pixel version, you'd have to get your hands dirty and do some careful pixel art.

It's similar in music. Current techology would allow for a melodyne-like analyzation of a complex piece's tonal and timbral values, followed by a crass algorithmic simplification. What it spits out might carry over some basic qualities of the original mood, but most of the information that pertains to the compositional essence will likely be lost.

Interestingly, just like a human being (in fact much moreso) an algorithm will have an easier time adding embellishments to something simple than downgrading complexity tastefully.

Branching out offers the freedom of diversity, while focusing in on core qualities forces you to be very deliberate. Computers aren't too good at that..

Just my 2 c, i can't even program but it seems relatively common sense.

[Brandon Strader]

You could use something like midi hit point mapping in cubase to make a midi from a big orchestration (I think, I haven't tried it) but it'd probably be a big mess of all the melodic notes playing at the same time through 1 instrument, which would be... insane.

It's reasonable to just hum along to a big song. If you can hum along to it, and recognize what you hum as being THAT SONG, then you are effectively deconstructing it to its source... you know, like a cheeseburger.

[Nase]

Yup, you need human taste to get to the meat. A computer might just chew on the cheese and pickles and spit out something rather hideous

[swansdown]

.

Edited September 8, 2023 by swansdown

[Nabeel Ansari]

There's a real term for this, it's called Textural Reduction.

There's no Wiki for some reason, but this may yield similar info.

[SkyRiderX]

I really feel its a lack of resources to do it mixed with song complexity. You can find Midi's and sheet music alot easier for older games than you can for new ones which means you have to remix the melody by ear and that can seem intimidating to remixers I think, especially new ones. On top of that songs nowadays already have a crap ton of layers to them and a fusion of multiple genres which makes it harder to think of new or interesting ways to remix songs.

This is just my thought on this...

[SystemsReady]

I was thinking about all this this morning, and I really think a lot of it does come down to complexity and the fact that it's really hard to fix something that isn't broken.

I read a lot of posts talking about how modern game music isn't as catchy, etc, but I can literally sing to you most of the themes from Okami, Odin Sphere, and most of Skyrim, and those are all orchestral soundtracks. I had Odin Sphere's miniboss theme stuck in my head while showering today.

But I can't write remixes to them. Granted, there's always the sure-fire method of just going "well what if [x] song was in [y] genre" (I've come across an electronic remix of Skyrim's "Dragonborn", which was...amusing), but that's not really an improvement, or that much of an actual change. I can't look at a sublime song like

and name a single thing I'd like to change about it, let alone enough to warrant a remix. And I think that's the biggest problem. It's way, way harder to remix something that is absolutely, completely fantastic, from songwriting to arrangement to instrumentation. It's like trying to improve a Van Gogh painting.

[Hyperion5182]

The above post is the single biggest reason.

Music in video games has taken a jump to ABOVE the music found in movies. And a lot of it has become down right iconic.

To have the balls to remix a soundtrack like The Normandy Reborn or Bayonetta's Now Epic Fly Me To the Moon is one thing. To get it done right is entirely something else.

You have a real limit especially with the growing quality of soundtracks in video games. The good games are getting music so strong you're just dumb struck.

In older games there weren't many songs outside of the SNES era that really put chills up your spine when you listened. And even then they're one or two per game leaving a remixer a lot of room to work.

These newer ones? They're being made with AAA class star power. Defiance's entire game soundtrack was created by none other than Bear McCreary himself. (Aside from Radioactive) And we've had OCR guys ON games already.

Compared to previous games where you really have room to work. The real new GOOD songs leave it all on the floor. It took real courage to take on the opening theme to Skyrim especially after i read its full translation. Sometimes you just hit a song so right no one wants to touch it.

At which point you have to raise the question would original compositions be preferable in dealing with future games?

[Eino Keskitalo]

I find the "it can't be improved" angle weird.. well, different from mine I guess. To me the point is almost never to "improve" on the original. The best NES or C64 soundtracks for example are perfect as they are already. You get inspired & do your own version, that's all.

--Eino

[Brandon Strader]

To have the balls to remix a soundtrack like Bayonetta's Now Epic Fly Me To the Moon is one thing. To get it done right is entirely something else.

I would argue that Frank Sinatra got it right when he covered Bayonetta.

I'm with you on the intimidation from how large soundtracks are these days. But I also feel like needing to match the quality or scope of the original is not necessary for a good remix -- and I think that might be consistent with OCR's beliefs as well, if I could be so presumptuous.

[Flexstyle]

I also feel like needing to match the quality or scope of the original is not necessary for a good remix -- and I think that might be consistent with OCR's beliefs as well, if I could be so presumptuous.

I seem to recall this being covered in an episode of OCR Talkback, and I think you're right. Also, as someone who's ReMixed a small chunk of an epic song from a modern OST, I think I can comfortably say that OCR has no problem with submissions only using a small portion of a track, as long as the song flows together properly.

[Nase]

The best NES or C64 soundtracks for example are perfect as they are already.

Agreed. They are what they are and it is good they are.

bigger =! better

you don't need to make something bigger OR better, you just wanna make something good.

making something small bigger + good is easier than making something big smaller + good.

making a good variation of something small is also easier than of something big.

Edited February 24, 2014 by Nase

[timaeus222]

Sometimes the VGM is just too layered to dissect. e.g. the sound design melds so well that it's hard to distinguish what's playing what. It's not an issue of no available MIDI for me---I stopped using MIDIs over a year ago---it's more about the notes that can be distinctly pulled out of the source tune.

For example, with the Vampire Variations II album, I picked a hell of a source called Dancing in Phantasmal Hell (get it?). It was... heh... difficult to interpret, to say the least. I'm surprised it's even cohesive enough nearing its loop point, but anyways...

The original track was fully orchestral, and I had to find a "downgraded" version which I presumed was somewhere out there, because I just couldn't distinguish all the notes as it was.

Original

Synthy Version

Clearly, you can hear the notes in the synthy version more easily, and that's the one I chose to reference while remixing. The orchestral version was more convoluted to me, especially after 0:14.

[SystemsReady]

I should probably clarify...my syntax is a little misleading, and written at 7 in the morning.

After listening to songs on here and reading judges' critiques, you need to change enough of an existing song for it to stand out on its own. Additions to the melody, new instrumentation, introducing new countermelodies etc etc.

This is easy to do for an old VGM song, because most songs are a single looping melodic line, a harmony, and maybe a countermelody. The quality of it is generally pretty low too, so you can add a lot simply by going "let's make [x part] [y instrument]". You can improvise more, and have more room to do things with it.

But when you look at new stuff, something like, say,

...what is there for you to change or add to make a decent remix out of? The melody is much longer than anything that shows up in a Sega Genesis title, there are a ton of instrument parts, there's so much more texture, the instrumentation is fantastic. It's so much more complex. It's so finished and everything works. Again, like trying to make your own version of a Van Gogh painting or something. I can't listen to this song and imagine how I would change it in any way, shape, or form, even as I can pick out each and every part of it. There's absolutely nothing I can add to it if I ever proposed writing a remix for it.

Earlier posts in the topic tried to make it an issue of catchiness, and I disagree. A ton of modern music is just as catchy, the issue is that it's really hard if not impossible to adequately add or re-interpret music that already has the parts, bells, and whistles to be a fully-fledged, imagined piece.

[timaeus222]

I should probably clarify...my syntax is a little misleading, and written at 7 in the morning.

After listening to songs on here and reading judges' critiques, you need to change enough of an existing song for it to stand out on its own. Additions to the melody, new instrumentation, introducing new countermelodies etc etc.

This is easy to do for an old VGM song, because most songs are a single looping melodic line, a harmony, and maybe a countermelody. The quality of it is generally pretty low too, so you can add a lot simply by going "let's make [x part] [y instrument]". You can improvise more, and have more room to do things with it.

But when you look at new stuff, something like, say,

...what is there for you to change or add to make a decent remix out of? The melody is much longer than anything that shows up in a Sega Genesis title, there are a ton of instrument parts, there's so much more texture, the instrumentation is fantastic. It's so much more complex. It's so finished and everything works. Again, like trying to make your own version of a Van Gogh painting or something. I can't listen to this song and imagine how I would change it in any way, shape, or form, even as I can pick out each and every part of it. There's absolutely nothing I can add to it if I ever proposed writing a remix for it.

Earlier posts in the topic tried to make it an issue of catchiness, and I disagree. A ton of modern music is just as catchy, the issue is that it's really hard if not impossible to adequately add or re-interpret music that already has the parts, bells, and whistles to be a fully-fledged, imagined piece.

I actually agree with the "catchiness" in that it seems to be a popular opinion, but I would call that mostly an effect of effective melodic contour. Instrumentation of course also plays a role, but I think that role is smaller than that of the melodic contour because if it wasn't, no one who hated chiptunes would like soundtracks for Genesis, NES, GBC, etc. Older soundtracks had less resources than modern soundtracks, so there was more emphasis on arrangement back then than now. The lower the quality of the instruments, the better the arrangement has to be to make up for that, so composers for those older games composed more carefully such that the melodies were memorable, able to be reinterpreted in many ways, and simplistic yet completely fitting.

Therefore, if a person says they "can't remix a source tune" (e.g. "don't think they can") because "it's too good", assuming the VGM in question is well-mixed, they're either 1) not confident they can deconstruct it, 2) not actually able to deconstruct it well enough to satisfy their needs, 3) it's actually too complex for anyone to deconstruct, 4) the arrangement is so specific to the instrumentation that it just wouldn't sound right in other instrumentation without a bunch of reworking of parts, or 5) they like it too much to do anything with it. I think possibility 4 is the most common objectively and possibility 5 is most common on the surface if surveying people, but those are just educated guesses. However, that means there's at least one person out there who can, so it wouldn't be out of the question to try remixing whatever source tune you're thinking of, as long as you feel "right" doing it. In other words, the textural complexity isn't unlimited, and all VGM is able to be remixed by someone who has gotten comfortable enough with remixing in the particular style that comes to their mind.

For example, I would consider this a "fully-fledged, imagined piece", and I would never have thought of remixing it (as in it never came to mind until recently) until I somehow got the inspiration to start at one point, but right now I'm pretty far into writing a remix incorporating that, this, and

(and I'm definitely going to finish it! ). I consider it difficult to deconstruct the chords in the first track (possibility 2) and a select few parts of the bass line in the second track (possibility 2) because you just have to be able to hear harmonies with relative certainty to examine those two in detail. Fact is, I really loved all three tracks opinion-wise as well as production-wise, so I guess I don't fit possibility 5. Edited February 26, 2014 by timaeus222