SemiEvolved

Have you seen the Wonder Woman 1984 trailer? No? It’s been out for a couple months now! Sheesh. Well, here it is. Watch it. I’ll wait.

Okay, so now you’ve seen it and if you’re like me, the first thing you notice is the epic Blue Monday cover by Sebastian Böhm. I love the bombast, the orchestral hits, the bassline. It’s just a brilliant cover of one of my favorite songs.

Fast forward to a couple days ago. Because we’re in the middle of a pandemic, I’m practicing social distancing by being bored at home and surfing YouTube. The WW1984 trailer pops into my feed at some point and I watch it again. And then, because I’m at my desk, an idea forms. An obvious idea, but a challenging one.

New Order vs Sebastian Böhm – “Blue Monday²” (Download)

This took the better part of a couple days. The Böhm cover is just over 3:30 long, while the original is nearly 7:30. I wanted to keep the structure of the original and lay the Böhm instruments on top to flesh out the arrangement. So I let the original guide the structure until the very end where the New Order track instrumentals itself into a gradual fade out. At that point, I brought in the Böhm version and let it shine.

Well, when I say shine, I need to be clear. The mastering on the Böhm cover is a turd. It’s an over-heated turd, set alight and left on the porch of your ears (okay, pushing this metaphor too far… I digress). It’s waaay too hot, with no dynamic range. And it’s only available as an mp3, which doesn’t help. That means there’s nowhere for the New Order instruments and voice to go. It’s just a big wall of sound.

Here’s where I struggle most when I’m finishing a mix. I lack a good understanding of filtering and mastering. Multiply that by terrible source material, and it’s a recipe for sonic fatigue before you’re halfway through the track.

I did the best I could without losing the will to continue. It’s still too hot and there’s nothing I can do to control it without losing the character of the Böhm cover entirely. My apologies. I’ll probably revisit it at some point when my ears are fresh. Maybe if work closes down I will have some more time to devote to it.

Anyways, I hope you enjoy. It’s not sonically perfect, but I’m pretty proud of the structure of it. Maybe they’ll release an HD version of the Böhm cover so I can remaster this and make it sound how it deserves to sound. For now, enjoy.

I almost didn’t finish this, but Jade made me. So you can blame her.

Enigma vs Nine Inch Nails – “Closer To Innocence” (Download)

There’s a long story behind this. Cheers to you, Yahoo!Groups user gen138. This one’s for you.

VNV Nation vs Sylvester Levay – “Electrowolf” (Download)

Do you watch Black Mirror? No? Then this one is likely to be a bit confusing, and I’m not going to explain why Ashley O exists, looks and sounds like Miley Cyrus, or why she’s singing an obnoxiously poppy rework of Nine Inch Nails’ “Head Like A Hole”. Google it.

Still here? Great. Now, the inspiration for this one is pretty simple. I watched Black Mirror and thought, “Well, somebody must have done that already.” So I looked around. Gave it a few weeks. Looked again. Then I came to the conclusion that I was just gonna have to do this for the internet, since the internet wasn’t doing it for me.

Ashley O vs Nine Inch Nails – “Head On A Roll” (Download)

This was a fun one, since I had the NIN stems to work with. It allowed me a ton of freedom with which sounds to pull in and when… sometimes that’s too much freedom, and I spent a lot of time experimenting with how much was too much. In the end, I let the Ashley O song take center stage and let NIN add some pop, primarily with snares, hats, and eventually some guitar and bass towards the end to kick things into high gear.

The Ashley O version was three semitones higher in pitch than NIN, which I mostly ignored. I changed the pitch of the aforementioned guitars and bassline when they kick in behind the chorus, just to keep from muddying the lyrics. I think it worked fairly seamlessly. I didn’t pitch Trent’s vocals because it gave way too much of a chipmunk effect. I did use a centering process to reduce the echo and reverb on Trent’s vocals to bring them into more atmospheric similarity with Ashley O’s. A minor adjustment, but I feel like it helped make a more cohesive song.

I’m especially proud of repitching a single synth hit that allowed me to extend the intro of the non-video edit of the song. I realized that the hit was made of two notes, and in the version I was trying to replicate without vocals, the notes changed pitch in opposite directions. So I couldn’t simply repitch the entire sample. I instead used the spectral frequency viewer to locate the individual notes, isolated them to new tracks (low, mid, high) then repitched the individual instruments in the sequencer. It worked perfectly, which was a bit of a surprise. So that’s another tool in the shed.

The instrumental for On A Roll took some doing. I was able to recreate some of it (see above) but there are entire sections that are completely obscured by vocals. So I did a little hunting, but couldn’t find an instrumental. I found a few filtered acapellas, ie: not studio acapellas but ones derived from the recording by filtering specific frequencies. It’s an inelegant way to get an acapella, because one often winds up extracting too much music or not enough vocals out of the mix. I found one that reduced the least amount of music, then applied wave inversion to remove the vocals from the full recording. The result was an instrumental that still had the faint echo of the vocals, but most of the music remained intact. I had to re-EQ that track to bring some of the midrange back up, but I feel like it was largely successful.

I made a video for this one, because why not? I used NIN’s live version since the music video for Head Like A Hole is… jarringly different. Also, it is only available in 480p, which would look terrible intercut with the 1080p of the Ashley O video. I also shortened the video version of the song, just so I wouldn’t have to invent another 30 seconds of footage. That doesn’t sound like a lot, but I had a hard time doing something similar with the InSoc/Sister’s of Mercy video, and I used three sources for that. Why complicate matters?

I had time off work, and this took a good chunk of my time for two days. The structure came together quickly, but the details, the levels, EQ, reverb and other processing, took a loooong time. But I feel really good about the outcome. I mean… it’s absurd and weird and not actually good, but I’m happy with it. I set out what I intended to achieve, which is all that matters.

Both of these songs are fairly old, of course. Monument came out in 2014 and Bloody Mary came out… longer ago than I care to consider. So this is yet another mashup that is woefully out of date, but I don’t particularly care because I think it sounds pretty cool. This one came about when I was listening to my Freezepop station on Pandora. I heard the Royksopp track and was immediately entranced by what I was hearing. Then the chorus hit and I thought, “Wait a fucking minute.” I took a note in my phone and the first chance I had I got to work on the mix.

Royksopp ft. Robyn vs Lady Gaga – “Bloody Monument” (Download)

Media used:

• Royksopp featuring Robyn – “Monument (The Inevitable End Version)”
• Lady Gaga – “Bloody Mary” (Studio stems)

Not a huge amount to say here. It came together fairly easily. I started a “rough” mix that I ended up working with until I was happy with the composition as a whole. So my “rough” canvas became the finished product. Almost.

I’ve started a new mastering process that I hope will be successful. I’ve noticed in the past, especially on tracks like “Dominion – Even More“, that the more I throw into the mix, the more fatiguing and crowded the sound feels. This isn’t anything new to music, and I was aware of the issue, but not really sure what to do with it given the source files I was working with. However, I’ve started experimenting with some things I can do to help make room for sounds to live together.

On this track, once I had finished the mix, I exported it as several files: Music, Vocals, and Noises, then brought those tracks into their own mix session so I could manipulate them more holistically. This allowed me to really play with levels and implement filtering of the music at the parts where things got crowded. For example, at around 4:28 in the track I have the vocoded “gaga” sound overlapping a fairly busy portion of Monument. So I moved the portion of Monument into a new track and implemented a noise reduction routine with the “noise” defined by the “gaga” sound. Basically, I subtly reduced the frequencies in Monument that were clashing with the frequencies in the “gaga” sample.

That’s just one trick I used, among others. Center channel reduction was another, which I used to make room in Monument for the majority of the Lady Gaga vocals by filtering a small frequency range out of the apparent center of the soundstage, where the Gaga vocals reside. This was especially key at the end where there is a LOT GOING ON around the vocals. It’s still a bit crowded, but I didn’t want to make changes so drastic that I seemed to be forcing the issue. I feel like it all works well with how I’ve presented it. Could it be better? Sure! Will I revisit this again in the future? Probably!

I sorta wanted to make this more of a mashup by bringing the entire Gaga song into the mix instead of working with just the stems. And maybe I will? I dunno… This felt right.

For now, however, I’m pretty happy with how this one turned out. And I gotta say… that Royksopp track is beautiful. Either the TIE version or the original. I feel like I could listen to it for days. And I just did, so that’s not an idle statement.

I wanted to title this post “There’s nothing worth doing that isn’t worth doing again after someone else has done a crappy job of it”, but I realized that was shockingly uncharitable. Many people have mashed these two songs together. It seems like a right of passage for aspiring mashup DJs. YouTube is rife with examples. So I’m not exactly breaking ground here. But this was an exercise worth posting for two reasons:

1. It really is a great combination and I’m pleased with the results.
2. I used wave inversion to good effect and I feel like that’s worth documenting.

Culture Beat vs Snap! “Mr. Vain is a Dancer” (Download)

As for #1, that’s pretty self-explanatory. It’s a good mix, the songs are basically the same in structure, and I feel like I brought them together in a harmony that’s lacking from most of what I’ve been able to find on YouTube and elsewhere.

As for #2? That’s the fun bit. What’s wave inversion, how did I use it, and so what?
Wave inversion is simple: If you take a wave file, invert it, then mix it with itself, it cancels itself out and the result is silence. This is how many people can create high quality acapellas. They take an instrumental version of a song, invert the wave, then mix it with the vocal version of the song. The result is that the difference between the instrumental and vocal versions remains: just the vocals. Now, this technique only works if both files are from the same source quality, preferably CD/FLAC/AAC lossless quality. There are many variables that factor into whether or not this works, but that’s the basic gist.

I’ve been aware of this technique for some time, and have used it to some minor success with certain projects. What’s so different about this project then? Simple. I used a subset of a sample to delete a specific sound, rather than using a full-range sample to delete multiple instruments. By using the frequency lasso tool in Audition, I was able to select the kick drum sound from Rhythm Is A Dancer. I then copied that selection to a new file, inverted it, then applied it to every instance of a kick drum that I wanted to delete.

It was tedious. First I would copy the inverted sample near a kick sound I wanted to remove. Then I’d try to align the samples based on the waveform. Listen to the single kick. Nudge the sample. Listen again. Move on  or keep nudging. Some kicks were easy. Some were recalcitrant. It just took a ton of time to get everything aligned, though honestly not as much as I was expecting when I started my experiment. Eventually, I had all the samples I needed with the kick essentially deleted.

I wanted the kick from Mr Vain to be the only kick in the resulting mix. Layering kicks over kicks just creates a spike in the waveform. Trying to filter out the bass frequencies of the kick just affects other areas of the track that I wanted to leave intact. Removing the kick, and only the kick, was the best course and it worked beautifully.

Anyways, it was a fun learning experience, and I feel like it’s a useful technique I will start applying to other projects as they arise. I may even go back to older projects and see if it’s something I can use to clean them up. For now, I hope you enjoy my take on this popular mashup.

Let’s get this out of the way up front:

Lady Gaga vs Depeche Mode “Silence the Paparazzi” (Download)

Lady Gaga vs Human League “Don’t You Want Romance” (Download)

I’ve had vocal stems for several Lady Gaga songs gathering virtual dust on my PC for years. I recently suffered (ie: inflicted) a serious RAID crash, losing a decent number of files and projects with it. I spent a long time away from the PC as a result. Facing the loss was depressing (on top of everything else that was happening in The Year That Sucked The Most), so I avoided restoring from an old back-up and starting over. It took a chance conversation on Twitter to get me to recover what was left and move on:

This is the fault of @littlelull @thefoodgeek @alvyon. It's not me, it's them. (Warning: sound. Terrible, terrible sound.) pic.twitter.com/lthHhAim8H

— Porcupine Rind (@SemiEvolved) November 16, 2016

This was hugely important for me. Not because it sounded good or worked (because it really, REALLY didn’t work), but it made me get back onto the PC and get on with making stuff. I decided to go back to basics and see what I could work with that I had left over from the crash. That’s when I started playing with the Gaga stems.

Long story short, it’s been a productive weekend. Yeah, the Gaga tracks are old and busted and overused. Hell, the Bad Romance/Don’t You Want Me mashup has been done before. Many times. But it felt great to be poking at details, tweaking levels, and nudging samples around again. I’m looking forward to whatever my idiot brain comes up with next.

I may update this with the usual write-up later, but for now I need to step away from All Gaga, All The Time weekend and focus on other things for a bit. That Gaga/DM mashup is pretty sweet tho, if I do say so. Hope you enjoy, and thanks for listening.

I’m super proud of this one. Like, really fucking proud. I love it a ton. Yes, I’m normally this humble. Lemme ‘splain.

Information Society has been one of my favorite bands since the late ’80s. I was hooked the first time I heard “What’s On Your Mind”. After releasing their third album in 1992, they called it quits. The lead singer bought the name and released an album that was stylistically different, though very good, but I despaired of ever seeing them live or hearing new music from the original lineup.

Then they announced a reunion (which was inevitable, given the number of ’80s groups reunion tours I’ve seen over the past decade). Two of the members released a new album in 2007, and I finally got to see them perform live. Then all three original members collaborated for another album in 2014 titled “_Hello World“. It was the album I’d been waiting 22 years for, and I was not disappointed.

Last month, they released Orders of Magnitude, an album of covers. There’s some crazy stuff on there, like “Capital I” from Sesame Street and “Heffalumps and Woozles” from Winnie the Pooh. Included on the tracklist, however, is a truly epic cover of Sister of Mercy’s “Dominion/Mother Russia”. As I listened to it again and again (and again), ideas began to coalesce.

DOWNLOAD

Media used:

• Information Society – Dominion
• Sisters Of Mercy – More (Extended Version)
• Sisters Of Mercy – More

Production notes:

• Total production time: Again, no really good count. I guess I should forego even trying to record this. Probably 20ish hours, more or less.
• I was in the middle of a different mashup when the idea for mixing InSoc’s cover of “Dominion” with SoM’s original popped into my head, but the tempos are so different I decided not to subject anything to that level of compression. While playing with that I realized that “More” was a much better candidate for mashing up with the InSoc cover. So I started working on that. The other mashup may come out at a later date, but it’s not nearly as awesome as this one.
• The chord progression is slightly different between the songs, so where I could, and if it made sense, I changed the chords on the InSoc track. It’s subtle, but I think makes the combination smoother overall.
• Fortunately there was no mucking about with pitch. Both songs were the same key. Yay!
• I knew that the mix would be based primarily around “Dominion”, so the challenge was how to bring “More” into the mix without overwhelming either track. Though I definitely threw caution to the wind on the final chorus of “Dominion”, putting pretty much everything I had into it. I had some trouble with the ending, as I usually do, but I think it came out decently well.
• Not a lot else. There was a lot of fiddling, some EQing, and a TON of adjusting the timing of “More” to match the metronomic beat of “Dominion”, but the result was totes worth it.
• Did I mention I love this mix? Because I do.