I still think that's such a creative way to assign notes. And you never know what you're going to get!
Well, the idea is that you can know what you're going to get by analyzing the functions, so if the result is unsatisfying, it's probably that I'm using the wrong functions.
@Elaro — While I can't say that it's particularly listenable, I appreciate the intent. I'm not sure what your musical ambitions are, but I think it'd be interesting to hear what you could create if you were to edit and elaborate upon these fractal skeletons. From a compositional perspective, a lot of music is already likely to incorporate structural behaviors akin to a fractal (like a fugue), so you might enjoy searching for various musics that could inspire a more refined expression of the maths that are inspiring you.
Thank you. Like, I understand my music (I know what to expect from it), but I needed the opinion of someone who didn't. You say it's bad? I believe you. Although I am curious: what exactly makes it unlistenable? Too repetitive? Bad tempo?
I do think this "composition by math" approach has merit, I just need to some form of Machine Learning to actually learn the math of good music.
I don't know if anyone else in here uses Reaper, but there's a free extension for managing extensions. It's like a Linux package manager, but for Reaper. There's a bunch of useful utilities.
@Elaro — To some extent what I perceive as "unlistenable" about it is that, by being purely dictated by the outcomes of mathematical formulae, there is no emotional, human core. The rhythms are disjointed, all lurching and stumbling — but they aren't spontaneous decisions of a pianist, they are the precise determinations of your maths. As such, there may be no revision that could please me, personally.
That said, I can imagine tha what you've presented could be more enjoyable if some adjustments were made such that it yielded a more distinct melodic phrase, as it's currently more of a "complex tonal sequence", as opposed to a chord progression that yields a truly harmonic sense of movement and purpose.
@Elaro — To some extent what I perceive as "unlistenable" about it is that, by being purely dictated by the outcomes of mathematical formulae, there is no emotional, human core. The rhythms are disjointed, all lurching and stumbling — but they aren't spontaneous decisions of a pianist, they are the precise determinations of your maths. As such, there may be no revision that could please me, personally.
That said, I can imagine tha what you've presented could be more enjoyable if some adjustments were made such that it yielded a more distinct melodic phrase, as it's currently more of a "complex tonal sequence", as opposed to a chord progression that yields a truly harmonic sense of movement and purpose.
Ah, you misunderstand. This music is not "dictated" by formulas, it is the formulas, so to speak. Every note, a note being a triplet of numbers Pitch, Length, Velocity, there is a sequence of numbers P, L and V such that Pitch of 1st note is P(1st), Length of 1st note is L(1st), Velocity of 1st note is V(1st), Pitch of 2nd note is P(2nd), and so on. P(n), L(n), V(n)? I choose that. I can literally express any piece of music with this notation, I just have to find P(n), L(n) and V(n) by hand right now.
So it's not the fact that this was created using formulas that you don't like, because every piece of music can be expressed with formulas, even the ones you like. What you don't like is the formulas I chose, and that's okay.
Don't worry, I'm not offended! I'm just slightly annoyed that I was misunderstood.
Oh cool a music thread. Im studying composition and had the opportunity to put together a few live performances for my compositions. Here is a recording of one and another of me playing a version off a computer while the camera is awkwardly trained on me.
@Elaro — To some extent what I perceive as "unlistenable" about it is that, by being purely dictated by the outcomes of mathematical formulae, there is no emotional, human core. The rhythms are disjointed, all lurching and stumbling — but they aren't spontaneous decisions of a pianist, they are the precise determinations of your maths. As such, there may be no revision that could please me, personally.
That said, I can imagine tha what you've presented could be more enjoyable if some adjustments were made such that it yielded a more distinct melodic phrase, as it's currently more of a "complex tonal sequence", as opposed to a chord progression that yields a truly harmonic sense of movement and purpose.
Ah, you misunderstand. This music is not "dictated" by formulas, it is the formulas, so to speak. Every note, a note being a triplet of numbers Pitch, Length, Velocity, there is a sequence of numbers P, L and V such that Pitch of 1st note is P(1st), Length of 1st note is L(1st), Velocity of 1st note is V(1st), Pitch of 2nd note is P(2nd), and so on. P(n), L(n), V(n)? I choose that. I can literally express any piece of music with this notation, I just have to find P(n), L(n) and V(n) by hand right now.
So it's not the fact that this was created using formulas that you don't like, because every piece of music can be expressed with formulas, even the ones you like. What you don't like is the formulas I chose, and that's okay.
Don't worry, I'm not offended! I'm just slightly annoyed that I was misunderstood.
There's actually some pretty cool research by Dr. Darrell Conklin, at.. uh.. University of Basque County I want to say? He does a lot of work on models for music prediction(and a fair bit on music generation, but most of the stuff I've read from him has been predictive). And specifically, about finding ways to reduce the amount of data needed to correctly model different styles of music. There's one great paper called Multiple viewpoint systems for music prediction that's a little dated, but still good as a general idea of not just using Pitch, Duration, and Velocity, but also using contour, harmonic structures, etc. I sort of ganked the idea for a lot of my work.
Just dropping by with a recent piece of mine, written for a Ace Attorney-inspired point-and-click. The composition is meant as a suspense theme, accompanying the initial discovery of crime scenes. Check it out
Here's another one I loaded up for the game. It's associated with the cross-examination segments. I'm looking for feedback on whether it properly conveys the mood - thanks for the listen!
@evocurio that's sounds really polished to me. I'd say you capture the mood.
So I'm very new to producing digital music, but I'm getting some ideas down in a DAW. I tend to like music (and compose music) that often has bass (as in bass guitar) carrying the melodic interest at times. I'm noticing that on headphones this sounds great, but playing it back through other speakers means you can barely hear what's happening. Does anyone have any tips for a newbie on this particular mixing problem?
I don't know where he got the scorpions, or how he got them into my mattress.
So I'm very new to producing digital music, but I'm getting some ideas down in a DAW. I tend to like music (and compose music) that often has bass (as in bass guitar) carrying the melodic interest at times. I'm noticing that on headphones this sounds great, but playing it back through other speakers means you can barely hear what's happening. Does anyone have any tips for a newbie on this particular mixing problem?
The most basic rule to abide by is: don't mix bass-centric music in headphones (or any music, really) — you will always be deceived, no matter how much your headphones purport to be "flat-response" or whatever.
Beyond that, my basic process for reinforcing melodic bass is to work with stereo-field and EQ plug-ins that help ensure that the bass is centered and present, positioned in the front of the mix (thus minimal reverb).
So I'm very new to producing digital music, but I'm getting some ideas down in a DAW. I tend to like music (and compose music) that often has bass (as in bass guitar) carrying the melodic interest at times. I'm noticing that on headphones this sounds great, but playing it back through other speakers means you can barely hear what's happening. Does anyone have any tips for a newbie on this particular mixing problem?
A few different tips, actually!
One thing you'll have to think about is what the likely environment is that whatever you're working on might be played. If you know, for instance, that this is probably just going to be through computer speakers, then mix to computer speakers as your primary source, and then check against other, nicer setups. If this is going to be performed in a big ol' space with giant 8-channel speakers and a subwoofer the size of a car, then mix with a great speaker setup.
A fun thing to do is to try doing a mix on your crappiest setup, and then checking against the good setup, rather than working in a perfect setting and then going "Oh that sounds shitty anywhere else". Only do this when you're getting to the mastering part though, don't bother with it during composition or you might end up making choices that are shitty, but are covered up by the shittiness of the medium.
At a certain point, you'll run into problems. If you have $20 speakers and no sub, you're not going to hear a big ol' beefy bass, no matter what you do to it. If you think that the general listeners will be on that setup then you're going to have to take that into account at the compositional level, rather than the mixing level.
A thing that you might try is making sure that the bass is super clean. Don't muddy it up too much with reverb or weird balancing, don't abuse it with filters, just keep a clean bass line through the rest of the mix, and it should help it pop out more.
Also, on the best setup you still might run into a compositional issue with putting too much bass doing melodic stuff. Try popping it up an octave, even on the same instrument, and see if that clears it up. The lowest thing will still sound like the bass note, so as long as you don't have really muddy voicings, you should be able to pop the bass up a little and still make out that it's the bass. Also, try just transposing the entire piece and see if that helps. Maybe go up a third, or up a 5th. If your bass stuff is all at the bottom of the bass clef staff, it'll mud up really easily in a way that even just being above the mid-staff F won't.
Something you can do if you dont have monitors (You should prioritize a decent set of monitors, though. I highly reccomend something like KRK Rockits. You can get a new pair of Rockit 5s for ~$300, and used ones for considerably less if you prowl craigslist) is while you're mixing throw down 2 or 3 test mixes, pushing things a little differently in each. Listen to them on as many different setups as you can. Listen on headphones, on the car, on a nice stereo, through your tv. Take extensive notes on how well represented each one is. The ultimate goal is to zero in on a mix that sounds accurate over a variety of speakers. Normally I do this during mastering, but without fairly flat response monitors its helpful to do it a couple times during the initial mixing as well.
I really like that track, those are live strings, yeah? Did you just compose it or are you performing as well? I know so little about orchestral composition, it seems like black sorcery to me, I am impressed
The whole record is on iTunes, Amazon Music, and Google Play Music. If you'd prefer a physical cd, they're available at http://jawbar.bigcartel.com
I really like that track, those are live strings, yeah? Did you just compose it or are you performing as well? I know so little about orchestral composition, it seems like black sorcery to me, I am impressed
The whole record is on iTunes, Amazon Music, and Google Play Music. If you'd prefer a physical cd, they're available at http://jawbar.bigcartel.com
Awesome thank you I should have looked on itunes first I guess.
I just composed it. I can technically play the viola but the only instrument Im really passable at is the guitar. Its a synth version. Im hoping to get some players together and do a live recording next year.
So @nightmarenny, there's a fun cheat technique I use often, which is basically a cheat towards making music seem ultra-technical. I notice that in this piece there's a lot of long melodic lines, and that's sort of fine with strings, but you can cheat lines and seem like a badass pretty easily. A trick I like to use is basically to breathe through phrases, and end any instruments participation in the phrase when you run out of breath. Generally let's say about 6 quarter notes at 120 is about the max that I like to have an instrument. I took your first melody introduction and tried re-voicing it for fun.*(in linked PDF) Also it sounds like:Soundcloud link!
Try splitting your melody into sub-melodies, and it's basically a way to seem like you're some technical genius without much thought
Hello again music makers. I finally managed to actually finish a track! I would be very grateful for feedback from anyone and everyone. I have a long list of things I need to work on in digital music production but maybe you all can tell me what needs to go straight to the top of the list. Thanks in advance!
Edit: I uploaded a new mix. Not 100% sure if it's better, but I try...
Solvent on
I don't know where he got the scorpions, or how he got them into my mattress.
Mixing! Your primary synth is so loud and midrangey that it is eating everything else. I can hear the drums, but only just. I think there might be an interesting bassline hiding in there somewhere, but it only peeks it head out for a moment before being devoured by that synth.
Quick tips: leave a ton of headroom on each channel. At the mixing stage if it isnt loud enough, turn up your speakers instead of turning up channels. Work on getting everything into balance first before pushing things around in the mix. Read as much as you can about EQ for mixing. EQ isn't just to shape your sounds, its far more useful as a tool to make sure that different instruments aren't fighting for the same frequency range, or stacking up and making a puddle of mud in the mids.
Song sounds cool, I'd just like to hear what everything is doing!
Mixing! Your primary synth is so loud and midrangey that it is eating everything else. I can hear the drums, but only just. I think there might be an interesting bassline hiding in there somewhere, but it only peeks it head out for a moment before being devoured by that synth.
Quick tips: leave a ton of headroom on each channel. At the mixing stage if it isnt loud enough, turn up your speakers instead of turning up channels. Work on getting everything into balance first before pushing things around in the mix. Read as much as you can about EQ for mixing. EQ isn't just to shape your sounds, its far more useful as a tool to make sure that different instruments aren't fighting for the same frequency range, or stacking up and making a puddle of mud in the mids.
Song sounds cool, I'd just like to hear what everything is doing!
I wish I could agree with your advice harder. EQing instruments for the overall mix can take an average mix and make it way better. On our last album I know the other guitar player wasn't happy with how his guitar sounds solo. However, in the mix the guitars sound amazing. Balancing everything against each other can be time consuming but I found it to be well worth the effort. Instead of making the kick drum louder, I just EQed it and the instruments interfering with its frequency range and hey, you can hear the bass guitar and the kick drum and the toms.
The part I bolded. So important. If you don't have a metering plug in of some kind that shows frequency response, get one. There are a few good free ones and a lot of paid ones. Voxengo SPAN is the free one I used to use. It works really well.
So @nightmarenny, there's a fun cheat technique I use often, which is basically a cheat towards making music seem ultra-technical. I notice that in this piece there's a lot of long melodic lines, and that's sort of fine with strings, but you can cheat lines and seem like a badass pretty easily. A trick I like to use is basically to breathe through phrases, and end any instruments participation in the phrase when you run out of breath. Generally let's say about 6 quarter notes at 120 is about the max that I like to have an instrument. I took your first melody introduction and tried re-voicing it for fun.*(in linked PDF) Also it sounds like:Soundcloud link!
Try splitting your melody into sub-melodies, and it's basically a way to seem like you're some technical genius without much thought
Awesome thanks for the tip! It reminds my of a lot of advice I've read about guitar soloing. That you should keep the length and contour of your lines singable most of the time. Did you have anymore thoughts? I'd really like to improve but Im having a lot of trouble getting a sense for what I'm doing well and what Im doing poorly. The class I made this for is performance analysis and the ability to take it as a composition class is very new. There was one other person taking comp lessons last year and nobody else this year. The class is mostly made up of singers and instrumentalists so while they had very literate and thankfully complementary things to say they didn't have much for me to lock on to. I also posted this piece on reddit but Im afraid I can't seem to get much substantive commentary going. I've been working on an album of rock music and when I show the demos to my friends they also don't have much to say so Im pretty starved for criticism!
Thanks, I appreciate your input!
I will take a deeper dive into EQing straight away.
This is a pretty neat interactive chart for navigating frequency bands while mixing: Frequency Chart
It obviously doesn't include all manner of common instruments, nor can it account for any number of synth patches, but it's still a good resource.
Oooooh thanks for posting this. Im in the middle of recording said album and Im basically learning as Im going. We're still a ways away from mixing but I like getting a headstart!
Right now I've got a guy recording drums to guide tracks Im making on sibelius. Im trying to get those done and get a rig set up for recording guitar. Think we've settled on getting a high quality microamp and a good mic.
Some good stuff on EQ being discussed here - it's a topic where even a little goes a long ways.
@nightmarenny You've got some good concepts going on in the Metroidvania piece. There was a healthy build up on the four-note background motif in the first section, I heard something Zelda-esque around the 0:40 mark and 0:50 had the makings of something "brave." That said, the feeling of braveness was never fully achieved imho, as you did not spend enough time elaborating on the theme. If there was one thing to change about your piece, it'd be making the parts (esp. melodic ideas) more cohesive. Solid work though, rock on!
//
Here's my latest. This is a boss theme intended to accompany a crossbow-wielding, motorcycle-riding punk.
I only recently discovered this thread, so I'm going to give a brief overview of the stuff I'm working on, since it's a little, uh, different.
Basically, while I have a background in music composition, my current work is on generative music, or computational creativity, or lots of other names. The short version is that it's writing systems that generate music(Generally speaking on computers, but "generative music" doesn't necessarily involve computers). My work is mainly in integrating these systems into video games, and also connected making super low-computation systems (So that they can be in games without needing, like, all of the resources).
I made a little prototype system recently that basically tries to create rich and flowing voice leading, by being as lazy as it possibly can be. It's pretty standard in actually composing voice leading, but oddly there don't seem to be too many systems that use the approach. I put it up on Itch for people to play around with it.
It's surprisingly successful for how basic and simple it is.
Also, the sounds aren't, well, great, but I basically had to make an FM synth from scratch to get them to be reasonable, and I don't want to fiddle with the sounds too much out of fear of breaking everything.
So, what you're doing is probably way more advanced than I have any right to talk about, but could you generate that as a MIDI signal and then run it into any number of synth options? That is if the goal is in any way trying to make the sounds more realistic.
Like, the whole "hey a program wrote this" part is super cool. And I think I let that part get away from me when I first listened to it. Is this something that's done sort of "on demand" or does the software write the part and then stores what it writes for playback later?
So, what you're doing is probably way more advanced than I have any right to talk about, but could you generate that as a MIDI signal and then run it into any number of synth options? That is if the goal is in any way trying to make the sounds more realistic.
Like, the whole "hey a program wrote this" part is super cool. And I think I let that part get away from me when I first listened to it. Is this something that's done sort of "on demand" or does the software write the part and then stores what it writes for playback later?
It could use MIDI, and there are some systems that do stuff with external audio stuff. Sometimes MIDI, sometimes though patches in Ableton. I generally am trying to do real-time synthesis with what I'm doing because when I integrate the stuff into games, I want to be able to have a single package rather than having some external... thing that does the synthesis.
This project in its current state is also weird, because it was also for a final project for a class, and we were using a Java applet called Processing. My normal MO is to use PureData, which is a little easier to build stuff in, and I'm hoping some day to take the time to build a good FM Synth in it.
For this system, it's completely real-time. Normally for larger-scale stuff I often do sort of mixes, where it'll maybe generate bursts of stuff so that it can put phrases together, but it really depends on what the system is trying to do. This system does have some non-generated stuff. The bass lines are all taken from pre-existing progressions, but the rest of the voices are all reacting to the "chord" when they change notes.
The idea behind this system was basically was that a lot of generative systems focus on creating chord progressions, and linking chords together in cool ways, but leave the voicing to either hard-coded relationships, or with some logic that does something like holds the notes to the chord and moves voices until those notes are all filled. This system is... similar to the second type. Basically though what I wanted for this one was to have the voice leading be as lazy as it possibly could be. So each "Chord" is just the bass note and the mode that the chord would suggest. Then voices can either select a subset of notes that they let themselves go to(the complexity part of the little grid they all have), or at the "real" version of the complexity, they just can go to any note in the mode. Voices will basically only move if they randomly decide to, or if they absolutely have to to fit the mode.
So the difference from some of the more common solutions to chords is that the actual chords that are just a result of the notes that are played, rather than notes being picked as a result of the chords. It results in some cool voicings and chord things, because it's not constrained to any particular chord.
Im moving forward on my recording project and trying to set up to record guitar in my room. Im pursuing a bunch of possibilities still but my current plan is to use a tiny amp and SM57 and just record that. My room is more or less untreated right now but Im planning on doing treatment after doing some tests. If anyone here has more practiced ears Id love some second opinions on the sound. I need to get this as close to exactly right as I can since my songs have rather elaborate arrangements.
Anyway I also managed to find better midis for some old pieces I wrote last years.
I can't listen to your mix at work because it's blocked but I'm going to info dump some guitar recording information I've learned over the past however many years. It may help.
Re-amp your guitars. If you are already doing this, good for you. You're saving time. If not, record a clean guitar signal track into your DAW. Route that track into a bus with a guitar amp simulation VST effect to make recording not sound "weird". There are some free ones out there that should be fine for tracking. Then you can record and record until you get the guitar part 100% correct. Now the playing is done and you don't have to worry about it anymore. Now time to set up the amp and microphone. So, you now the clean guitar track output from your DAW and interface become the input on your guitar amp. Connect the guitar amp microphone back into your interface and create a new track on your DAW with this microphone as the input. Place the microphone where you want it, dial in the tone you want on the guitar amp. Make sure the clean guitar signal is the ONLY thing playing, solo it. Arm the microphone track in front of the amp to record. Now record that sweet sweet guitar. Don't like the end result of the guitar tone? Easy, just change up mic placement or amp settings and record again. This way, you only have to perform the part correctly once and you can re-amp again and again until you find the tone you like. You could also record a few different settings and blend them. Something like, one track with the SM57 pointed directly at the center of the speaker and another track with the mic slightly off center and mix the two sounds together.
Or, if you can afford it, there are some good guitar VSTs and cabinet responses out there right now. Demo a few of them and see if you can dial in a sound you think is better than the amp you are currently using. If it doesn't sound better to you, don't use it. I've found though that with a good amp sim VST, cabinet response, and patience; a good sound in the mix is achievable. I've done a record with Amplitube 4 and cabinet responses from 3 Sigma Audio and no one has called it out and anyone I've told has been very surprised. They may sound a bit unusual on their own, but in a full mix with all the other instruments going on it sounds completely natural.
For the untreated room part. If you're recording guitar, just throw a moving blanket or sleeping bag over the whole thing, making sure not to move the mic, and you should be fine.
I can't listen to your mix at work because it's blocked but I'm going to info dump some guitar recording information I've learned over the past however many years. It may help.
Re-amp your guitars. If you are already doing this, good for you. You're saving time. If not, record a clean guitar signal track into your DAW. Route that track into a bus with a guitar amp simulation VST effect to make recording not sound "weird". There are some free ones out there that should be fine for tracking. Then you can record and record until you get the guitar part 100% correct. Now the playing is done and you don't have to worry about it anymore. Now time to set up the amp and microphone. So, you now the clean guitar track output from your DAW and interface become the input on your guitar amp. Connect the guitar amp microphone back into your interface and create a new track on your DAW with this microphone as the input. Place the microphone where you want it, dial in the tone you want on the guitar amp. Make sure the clean guitar signal is the ONLY thing playing, solo it. Arm the microphone track in front of the amp to record. Now record that sweet sweet guitar. Don't like the end result of the guitar tone? Easy, just change up mic placement or amp settings and record again. This way, you only have to perform the part correctly once and you can re-amp again and again until you find the tone you like. You could also record a few different settings and blend them. Something like, one track with the SM57 pointed directly at the center of the speaker and another track with the mic slightly off center and mix the two sounds together.
Or, if you can afford it, there are some good guitar VSTs and cabinet responses out there right now. Demo a few of them and see if you can dial in a sound you think is better than the amp you are currently using. If it doesn't sound better to you, don't use it. I've found though that with a good amp sim VST, cabinet response, and patience; a good sound in the mix is achievable. I've done a record with Amplitube 4 and cabinet responses from 3 Sigma Audio and no one has called it out and anyone I've told has been very surprised. They may sound a bit unusual on their own, but in a full mix with all the other instruments going on it sounds completely natural.
For the untreated room part. If you're recording guitar, just throw a moving blanket or sleeping bag over the whole thing, making sure not to move the mic, and you should be fine.
Question for those with a lot of recording experience-
Ive been attemtping to start recording guitar for my album in my limited free time the last few weeks and Im havig a hard time. I was wondering how close to perfect you get your takes? Should I just try to get something close and do extensive cut/pasting to get it sounding where it needs to be?
This is going to depend on how much a not perfect performance is going to bug you for all time. There are older albums I've done where I can still hear very minor things and it drives me crazy. But it's hard to fix a recording that's 10 years old and has been released. If you're just doing this 100% for yourself and are okay with things being imperfect and you can let it slide, go for it. If it's something you plan to have other people listen to and receive feedback on, get it as perfect as you can. Because someone somewhere may hear it and point it out to you.
Please don't do cut and pasting. I know it's done a lot and this is just my opinion on the matter but please don't. I think you would be better off practicing more until you can play the part the way you intend it to sound. It's more work, but in the end you get an honest performance. Also, if you don't cut and paste perfectly it can sound really really bad. I've never been a fan.
Now, if you want to do punch ins instead that's something that is usually less frowned upon. At least with punch ins you still have to perform the part correctly X number of times instead of once. Say guitar player is playing a song and upon listening back to the take we hear one bum note or a weird pick scrape or something. I wouldn't ask the person to record the entire track over again. Instead we'd just queue up the track a bit before the flub, start the song playing, the guitar player plays along and sometime before the flub and then we move from playing the track to recording. Once we have recorded past the flub we stop the recording. Then listen back to the new take and verify it is correct. Then you blend the original take into the fix of the second take with fades and move on. Takes less time than recording the whole song again and you still get a more real performance. I think it sounds more natural as well as you maintain a waveform without breaks.
Again, all opinions here. If you want to ignore it, do so. Ultimately it's your project and as long as your pleased with it and you accomplished what you wanted, everyone else can get bent.
Yeah, Bolt's pretty much covered everything I would say.
But what I might add is that my typical process would be to segment the recording process into the most discrete phrases/sections, do my best to execute those components with minimal error, compile those sections into a "best of" full performance, then practice to that and eventually record the entire track in a single take that I might then make smaller punch-ins to "perfect".
Thanks peeps! I didnt even realize there was much difference between doing punch ins and cut/pasting which I suppose shows my noviceness right there haha.
One more related question. Ive been hiring a drummer to do my songs based on a scratch track so I didnt have to set up a recording space that could accomidate a drummer. So I had planned to just commit to the form of the song I sent him but Ive been considering making one song longer. No new sections. Just an extra verse and a longer chorus intro. Im wondering if I copy the chorus drum parts and put it at the beginning if that will sound as obviously shopped as the giitar will.
So, you're saying you take the drum tracks only and copying and pasting those and then performing everything else in full overtop of those drum tracks? Could be doable. Just make sure you allow the pasted tracks to crossfade correctly if needed and that you don't end up with something where you've made the drummer play something that's not physically possible like hitting 5 things at the same time.
Unless the tracks were originally played to a click track, I wish you the best of luck getting the timing of the paste 100% correct. That won't be much fun.
Posts
Well, the idea is that you can know what you're going to get by analyzing the functions, so if the result is unsatisfying, it's probably that I'm using the wrong functions.
Thank you. Like, I understand my music (I know what to expect from it), but I needed the opinion of someone who didn't. You say it's bad? I believe you. Although I am curious: what exactly makes it unlistenable? Too repetitive? Bad tempo?
I do think this "composition by math" approach has merit, I just need to some form of Machine Learning to actually learn the math of good music.
I don't know if anyone else in here uses Reaper, but there's a free extension for managing extensions. It's like a Linux package manager, but for Reaper. There's a bunch of useful utilities.
https://reapack.com/
That said, I can imagine tha what you've presented could be more enjoyable if some adjustments were made such that it yielded a more distinct melodic phrase, as it's currently more of a "complex tonal sequence", as opposed to a chord progression that yields a truly harmonic sense of movement and purpose.
Ah, you misunderstand. This music is not "dictated" by formulas, it is the formulas, so to speak. Every note, a note being a triplet of numbers Pitch, Length, Velocity, there is a sequence of numbers P, L and V such that Pitch of 1st note is P(1st), Length of 1st note is L(1st), Velocity of 1st note is V(1st), Pitch of 2nd note is P(2nd), and so on. P(n), L(n), V(n)? I choose that. I can literally express any piece of music with this notation, I just have to find P(n), L(n) and V(n) by hand right now.
So it's not the fact that this was created using formulas that you don't like, because every piece of music can be expressed with formulas, even the ones you like. What you don't like is the formulas I chose, and that's okay.
Don't worry, I'm not offended! I'm just slightly annoyed that I was misunderstood.
Ive also been working on demos for a rock album. They are pretty rough right now but what the heck
There's actually some pretty cool research by Dr. Darrell Conklin, at.. uh.. University of Basque County I want to say? He does a lot of work on models for music prediction(and a fair bit on music generation, but most of the stuff I've read from him has been predictive). And specifically, about finding ways to reduce the amount of data needed to correctly model different styles of music. There's one great paper called Multiple viewpoint systems for music prediction that's a little dated, but still good as a general idea of not just using Pitch, Duration, and Velocity, but also using contour, harmonic structures, etc. I sort of ganked the idea for a lot of my work.
EDIT: Also hey cool a music thread!
Just dropping by with a recent piece of mine, written for a Ace Attorney-inspired point-and-click. The composition is meant as a suspense theme, accompanying the initial discovery of crime scenes. Check it out
Thanks for the listen!
Watch my music videos
Here's another one I loaded up for the game. It's associated with the cross-examination segments. I'm looking for feedback on whether it properly conveys the mood - thanks for the listen!
Watch my music videos
So I'm very new to producing digital music, but I'm getting some ideas down in a DAW. I tend to like music (and compose music) that often has bass (as in bass guitar) carrying the melodic interest at times. I'm noticing that on headphones this sounds great, but playing it back through other speakers means you can barely hear what's happening. Does anyone have any tips for a newbie on this particular mixing problem?
http://newnations.bandcamp.com
The most basic rule to abide by is: don't mix bass-centric music in headphones (or any music, really) — you will always be deceived, no matter how much your headphones purport to be "flat-response" or whatever.
Beyond that, my basic process for reinforcing melodic bass is to work with stereo-field and EQ plug-ins that help ensure that the bass is centered and present, positioned in the front of the mix (thus minimal reverb).
Getting the bass to sound right is never easy.
A few different tips, actually!
One thing you'll have to think about is what the likely environment is that whatever you're working on might be played. If you know, for instance, that this is probably just going to be through computer speakers, then mix to computer speakers as your primary source, and then check against other, nicer setups. If this is going to be performed in a big ol' space with giant 8-channel speakers and a subwoofer the size of a car, then mix with a great speaker setup.
A fun thing to do is to try doing a mix on your crappiest setup, and then checking against the good setup, rather than working in a perfect setting and then going "Oh that sounds shitty anywhere else". Only do this when you're getting to the mastering part though, don't bother with it during composition or you might end up making choices that are shitty, but are covered up by the shittiness of the medium.
At a certain point, you'll run into problems. If you have $20 speakers and no sub, you're not going to hear a big ol' beefy bass, no matter what you do to it. If you think that the general listeners will be on that setup then you're going to have to take that into account at the compositional level, rather than the mixing level.
A thing that you might try is making sure that the bass is super clean. Don't muddy it up too much with reverb or weird balancing, don't abuse it with filters, just keep a clean bass line through the rest of the mix, and it should help it pop out more.
Also, on the best setup you still might run into a compositional issue with putting too much bass doing melodic stuff. Try popping it up an octave, even on the same instrument, and see if that clears it up. The lowest thing will still sound like the bass note, so as long as you don't have really muddy voicings, you should be able to pop the bass up a little and still make out that it's the bass. Also, try just transposing the entire piece and see if that helps. Maybe go up a third, or up a 5th. If your bass stuff is all at the bottom of the bass clef staff, it'll mud up really easily in a way that even just being above the mid-staff F won't.
Absolutely
This is really cool @Knob. Is there anywhere I buy or download it?
Also here is a knew piece I made for my performance analysis class. Hope you guys like it.
The whole record is on iTunes, Amazon Music, and Google Play Music. If you'd prefer a physical cd, they're available at http://jawbar.bigcartel.com
Awesome thank you I should have looked on itunes first I guess.
I just composed it. I can technically play the viola but the only instrument Im really passable at is the guitar. Its a synth version. Im hoping to get some players together and do a live recording next year.
Try splitting your melody into sub-melodies, and it's basically a way to seem like you're some technical genius without much thought
Edit: I uploaded a new mix. Not 100% sure if it's better, but I try...
http://newnations.bandcamp.com
Quick tips: leave a ton of headroom on each channel. At the mixing stage if it isnt loud enough, turn up your speakers instead of turning up channels. Work on getting everything into balance first before pushing things around in the mix. Read as much as you can about EQ for mixing. EQ isn't just to shape your sounds, its far more useful as a tool to make sure that different instruments aren't fighting for the same frequency range, or stacking up and making a puddle of mud in the mids.
Song sounds cool, I'd just like to hear what everything is doing!
I wish I could agree with your advice harder. EQing instruments for the overall mix can take an average mix and make it way better. On our last album I know the other guitar player wasn't happy with how his guitar sounds solo. However, in the mix the guitars sound amazing. Balancing everything against each other can be time consuming but I found it to be well worth the effort. Instead of making the kick drum louder, I just EQed it and the instruments interfering with its frequency range and hey, you can hear the bass guitar and the kick drum and the toms.
The part I bolded. So important. If you don't have a metering plug in of some kind that shows frequency response, get one. There are a few good free ones and a lot of paid ones. Voxengo SPAN is the free one I used to use. It works really well.
PSN : Bolthorn
I will take a deeper dive into EQing straight away.
http://newnations.bandcamp.com
This is a pretty neat interactive chart for navigating frequency bands while mixing: Frequency Chart
It obviously doesn't include all manner of common instruments, nor can it account for any number of synth patches, but it's still a good resource.
Awesome thanks for the tip! It reminds my of a lot of advice I've read about guitar soloing. That you should keep the length and contour of your lines singable most of the time. Did you have anymore thoughts? I'd really like to improve but Im having a lot of trouble getting a sense for what I'm doing well and what Im doing poorly. The class I made this for is performance analysis and the ability to take it as a composition class is very new. There was one other person taking comp lessons last year and nobody else this year. The class is mostly made up of singers and instrumentalists so while they had very literate and thankfully complementary things to say they didn't have much for me to lock on to. I also posted this piece on reddit but Im afraid I can't seem to get much substantive commentary going. I've been working on an album of rock music and when I show the demos to my friends they also don't have much to say so Im pretty starved for criticism!
Oooooh thanks for posting this. Im in the middle of recording said album and Im basically learning as Im going. We're still a ways away from mixing but I like getting a headstart!
Right now I've got a guy recording drums to guide tracks Im making on sibelius. Im trying to get those done and get a rig set up for recording guitar. Think we've settled on getting a high quality microamp and a good mic.
@nightmarenny You've got some good concepts going on in the Metroidvania piece. There was a healthy build up on the four-note background motif in the first section, I heard something Zelda-esque around the 0:40 mark and 0:50 had the makings of something "brave." That said, the feeling of braveness was never fully achieved imho, as you did not spend enough time elaborating on the theme. If there was one thing to change about your piece, it'd be making the parts (esp. melodic ideas) more cohesive. Solid work though, rock on!
//
Here's my latest. This is a boss theme intended to accompany a crossbow-wielding, motorcycle-riding punk.
I only recently discovered this thread, so I'm going to give a brief overview of the stuff I'm working on, since it's a little, uh, different.
Basically, while I have a background in music composition, my current work is on generative music, or computational creativity, or lots of other names. The short version is that it's writing systems that generate music(Generally speaking on computers, but "generative music" doesn't necessarily involve computers). My work is mainly in integrating these systems into video games, and also connected making super low-computation systems (So that they can be in games without needing, like, all of the resources).
I made a little prototype system recently that basically tries to create rich and flowing voice leading, by being as lazy as it possibly can be. It's pretty standard in actually composing voice leading, but oddly there don't seem to be too many systems that use the approach. I put it up on Itch for people to play around with it.
I also made a demo video of it running for a bit.
It's surprisingly successful for how basic and simple it is.
Also, the sounds aren't, well, great, but I basically had to make an FM synth from scratch to get them to be reasonable, and I don't want to fiddle with the sounds too much out of fear of breaking everything.
Like, the whole "hey a program wrote this" part is super cool. And I think I let that part get away from me when I first listened to it. Is this something that's done sort of "on demand" or does the software write the part and then stores what it writes for playback later?
PSN : Bolthorn
It could use MIDI, and there are some systems that do stuff with external audio stuff. Sometimes MIDI, sometimes though patches in Ableton. I generally am trying to do real-time synthesis with what I'm doing because when I integrate the stuff into games, I want to be able to have a single package rather than having some external... thing that does the synthesis.
This project in its current state is also weird, because it was also for a final project for a class, and we were using a Java applet called Processing. My normal MO is to use PureData, which is a little easier to build stuff in, and I'm hoping some day to take the time to build a good FM Synth in it.
For this system, it's completely real-time. Normally for larger-scale stuff I often do sort of mixes, where it'll maybe generate bursts of stuff so that it can put phrases together, but it really depends on what the system is trying to do. This system does have some non-generated stuff. The bass lines are all taken from pre-existing progressions, but the rest of the voices are all reacting to the "chord" when they change notes.
The idea behind this system was basically was that a lot of generative systems focus on creating chord progressions, and linking chords together in cool ways, but leave the voicing to either hard-coded relationships, or with some logic that does something like holds the notes to the chord and moves voices until those notes are all filled. This system is... similar to the second type. Basically though what I wanted for this one was to have the voice leading be as lazy as it possibly could be. So each "Chord" is just the bass note and the mode that the chord would suggest. Then voices can either select a subset of notes that they let themselves go to(the complexity part of the little grid they all have), or at the "real" version of the complexity, they just can go to any note in the mode. Voices will basically only move if they randomly decide to, or if they absolutely have to to fit the mode.
So the difference from some of the more common solutions to chords is that the actual chords that are just a result of the notes that are played, rather than notes being picked as a result of the chords. It results in some cool voicings and chord things, because it's not constrained to any particular chord.
Anyway I also managed to find better midis for some old pieces I wrote last years.
Re-amp your guitars. If you are already doing this, good for you. You're saving time. If not, record a clean guitar signal track into your DAW. Route that track into a bus with a guitar amp simulation VST effect to make recording not sound "weird". There are some free ones out there that should be fine for tracking. Then you can record and record until you get the guitar part 100% correct. Now the playing is done and you don't have to worry about it anymore. Now time to set up the amp and microphone. So, you now the clean guitar track output from your DAW and interface become the input on your guitar amp. Connect the guitar amp microphone back into your interface and create a new track on your DAW with this microphone as the input. Place the microphone where you want it, dial in the tone you want on the guitar amp. Make sure the clean guitar signal is the ONLY thing playing, solo it. Arm the microphone track in front of the amp to record. Now record that sweet sweet guitar. Don't like the end result of the guitar tone? Easy, just change up mic placement or amp settings and record again. This way, you only have to perform the part correctly once and you can re-amp again and again until you find the tone you like. You could also record a few different settings and blend them. Something like, one track with the SM57 pointed directly at the center of the speaker and another track with the mic slightly off center and mix the two sounds together.
Or, if you can afford it, there are some good guitar VSTs and cabinet responses out there right now. Demo a few of them and see if you can dial in a sound you think is better than the amp you are currently using. If it doesn't sound better to you, don't use it. I've found though that with a good amp sim VST, cabinet response, and patience; a good sound in the mix is achievable. I've done a record with Amplitube 4 and cabinet responses from 3 Sigma Audio and no one has called it out and anyone I've told has been very surprised. They may sound a bit unusual on their own, but in a full mix with all the other instruments going on it sounds completely natural.
For the untreated room part. If you're recording guitar, just throw a moving blanket or sleeping bag over the whole thing, making sure not to move the mic, and you should be fine.
PSN : Bolthorn
Sounds good thanks for the tip
Ive been attemtping to start recording guitar for my album in my limited free time the last few weeks and Im havig a hard time. I was wondering how close to perfect you get your takes? Should I just try to get something close and do extensive cut/pasting to get it sounding where it needs to be?
Please don't do cut and pasting. I know it's done a lot and this is just my opinion on the matter but please don't. I think you would be better off practicing more until you can play the part the way you intend it to sound. It's more work, but in the end you get an honest performance. Also, if you don't cut and paste perfectly it can sound really really bad. I've never been a fan.
Now, if you want to do punch ins instead that's something that is usually less frowned upon. At least with punch ins you still have to perform the part correctly X number of times instead of once. Say guitar player is playing a song and upon listening back to the take we hear one bum note or a weird pick scrape or something. I wouldn't ask the person to record the entire track over again. Instead we'd just queue up the track a bit before the flub, start the song playing, the guitar player plays along and sometime before the flub and then we move from playing the track to recording. Once we have recorded past the flub we stop the recording. Then listen back to the new take and verify it is correct. Then you blend the original take into the fix of the second take with fades and move on. Takes less time than recording the whole song again and you still get a more real performance. I think it sounds more natural as well as you maintain a waveform without breaks.
Again, all opinions here. If you want to ignore it, do so. Ultimately it's your project and as long as your pleased with it and you accomplished what you wanted, everyone else can get bent.
PSN : Bolthorn
But what I might add is that my typical process would be to segment the recording process into the most discrete phrases/sections, do my best to execute those components with minimal error, compile those sections into a "best of" full performance, then practice to that and eventually record the entire track in a single take that I might then make smaller punch-ins to "perfect".
One more related question. Ive been hiring a drummer to do my songs based on a scratch track so I didnt have to set up a recording space that could accomidate a drummer. So I had planned to just commit to the form of the song I sent him but Ive been considering making one song longer. No new sections. Just an extra verse and a longer chorus intro. Im wondering if I copy the chorus drum parts and put it at the beginning if that will sound as obviously shopped as the giitar will.
Unless the tracks were originally played to a click track, I wish you the best of luck getting the timing of the paste 100% correct. That won't be much fun.
PSN : Bolthorn