timaeus222

All normalizing does is boost the volume of the piece until the loudest peak is at 0dB. It's nothing complicated. What you happen to be doing now is one "solution" that you think is solving all your problems, but really, it's covering them up. Think of it this way: you're doing a math problem, and you did the wrong work to get the right answer, and you figure that the wrong work is then correct (which it isn't, as for example, one step was mathematically incorrect, and another mathematically incorrect step gave you the correct answer by accident). A similar, musical way to say it is that you're writing music louder than 0dB, then using a compressor on the master track and doing volume edits while the compressor is still on; therefore, your edits don't actually correspond to the results you are getting unless you consider what the compressor did. So, I actually wouldn't normalize it at all. It essentially limits your flexibility. Besides, you can still have a spike in your song and it wouldn't normalize properly. Why don't you just use a limiter? That's what most producers use on their music anyways. If you want a "best" volume, as you call it, then it's more helpful to your learning if you use your ears to find a good volume. The "best" volume is definitely not always at 0dB.

Maybe if you wanted to test it, you could try using a spectroscope (like smexoscope) and looking at what happens when you do the compression with the built-in EQ turned on, on a pure sine wave or a pure square wave with the Sustain all the way up on the ADSR envelope. That way, you're testing it on a completely flat sound. If it looks a little different after applying the EQ, then it works without the side chain feature.

This is actually not mastered that loudly since the mixing is so murky. The less treble you perceive, the quieter something usually seems. Think about when you stood outside a room trying to hear music inside the room playing at a normal volume with the door already closed. Sounds dull, right? Something like this at 0:48 is much louder, so this ReMix wasn't mastered with the goal of being loud, and you don't have to flip tables over loudness.

Yep, that's precisely what the "secret" is. If you carefully select matching sounds that are balanced and EQ'd such that you can hear many of them, you essentially create a piece with good depth perception---in other words, when you listen closely, you can hear more detail if you focus on particular instruments. In two words, Layering and Fullness. Also, while I was writing my ReMix for the Vampire Variations II album, I had found that, like you had, just boosting audio past 0dB a bit does not sound quite the same as mixing to 0dB and then applying well-executed compression to bring it past 0dB a bit (of course, presuming a brickwall limiter is already there). Quoting myself: Just a sidenote, but I'm 100% sure that SnappleMan has mentioned the attenuation bit before, as I remember posting to tell him that it was an awesome post.

Alright then. Not everyone hears things the same way I do. I mean, I thought those seemed quite different in loudness, but ah well. I'll try choosing more drastic examples. How about these pairs? Try imprinting the loudness that you perceive, and holding onto that, replaying a little of it before hearing the other piece. Perhaps hearing the quieter piece will make it seem more drastic. I used to find volume increases to be more noticeable than volume decreases (now they're similar for me, so either direction is pretty much an equal change perception). https://soundcloud.com/zircon-1/level-bounce @ 0:49 being louder than https://soundcloud.com/chimpazilla/cloudhopping-yoshi-touch-go-by @ 0:29 This is a loudness difference from compression on the master track. @ 2:28 being louder than @ 3:19This is a loudness difference also from compression on the master track, but additionally from drum sample choice. https://soundcloud.com/isworks/celestia-terrarium-naked-by @ 0:21 being louder than http://omnipsyence.bandcamp.com/track/fall-in-free-fall-mix-ft-tera-catallo @ 0:03 This is a loudness difference from layering, and it's meant to show that even though "Terrarium" doesn't have explicit drums, it is fuller (and thus louder) than the "Fall In" ReMix, which does have drums. There's not much physics that needs to be incorporated here. It can be explained like so: Think of a "what if" situation, where there are no limiters on any two songs or pieces you compare. The one that would clip more is louder. That's pretty much it. How capable your limiter is to brickwall is what helps you to get to a higher loudness (but at the same time, boosting too loudly will "squash" the mix and automatically attenuate the highest treble frequencies very slightly... until the limiter's tolerance is reached. Then it'll just clip). Also, when you say hearing something by itself with no comparison gives little indication of loudness, it's technically true, but the length of time between comparisons doesn't have to be quite as recently as, say, a day or an hour. I hear differences in loudness based on comparing everything I've heard to the loudest song or piece I have ever heard, ever. To be specific, I find zircon's Level Bounce to be the loudest piece I will ever comfortably listen to, and when I hear other pieces, my ears will naturally tell me that it's louder or quieter. Maybe it's just something with my photographic memory that helps, but that's how it works with me.

It may just be a personal taste thing, but I think there should be less of the low-mids picking sound on the guitar. It somehow feels mechanical with those in, and it's also clashing with the pad later on. Also, I agree that the clap currently doesn't fit. It can, though, with some work. I think the drum patterns with two claps in a row disrupt the ambient vibe. If you try lowering the dry signal of the clap and slightly increasing the reverb predelay to make it sound more distant, that should help. Try listening to this one.

Yep. Well... that's partially true, but you're thinking of it in a way that can be described like this: Let's say you do a volume measurement of something and there's a single random spike. Everything is at -10dB except for one spike that's at 0dB. You're saying that that is not loud overall, but it sounds like what you're really saying is that the stuff is not loud on average. While that's true, you'd be putting the 0dB spike to the side. If you zoom into the 0dB spike, select that portion, and loop it, it IS going to be loud; it's just sustained over a very short period of time. Hence, it doesn't sound quiet---it sounds loud relative to the quiet -10dB audio, and if you try to normalize it, nothing happens. If you don't look at your waveform for your audio, you may not notice that, and perhaps get confused on how to get it louder without clipping. Also, what you seem to be talking about with sub bass is "perceived" loudness. That's different for everyone, because if you have a different audio system, you're not going to hear the same loudness... at least, with this definition. Yeah, you can't hear sub bass, but it doesn't mean it's not making a sound and flooding your limiter. This "perceived" loudness is not exactly just the plain amplitude of the final result; it's more like how the amplitude changes relative to the rest of the audio. Jumping from -10dB to -2dB isn't quite as loud a perception as -30dB to -4dB, even though -4dB is quieter than -2dB. Additionally, it's based on how well the frequency spectrum is occupied. You can have flooded mixing and it'll sound loud. You can have pumped, compressed dance music and it'll sound loud. But the "real" way to sound loud (in a controlled way) is to layer carefully, making sure most things are distinguishable, levels are balanced, and the rise and fall of the audio is done tastefully. The frequency spectrum should be quite fully occupied in loud parts, otherwise a few instruments are just boosted too loudly. EX: This is louder than this, despite the fact that they both hit a 0dB limiter in the end. Yes, there is the consideration of "what if there was no limiter", but the point is, the layering in "Darkness" is more detailed and carefully done than in "Fiberoptics", and thus "Darkness" is substantially louder. ---------------- Yes, layering IS supposed to make things sound louder. Constructive interference. It's normal. Semi-basic physics. Add two waves, and crests that line up get taller after they add. Simple as that. Crests that line up with troughs cancel out and the taller of the two dominates the final result---if the crest is taller, then the result is a crest with an amplitude equal to the difference in their original amplitude magnitudes, and vice versa. No one makes DAWs that work towards making it harder for you to accomplish what you want. That's why there are beta testers and bug fixes.

Yeah, I think it does sound better! Or maybe I'm imagining things on the BD cover =P

The "dark metal romp" thing kinda gave it away Anyways, I figure the drums sound pretty metal. =P

I would just keep panning like a normal orchestra, but remember that some instruments should be more upfront than others. Trumpets are way in the back because they're naturally loud, and violins/violas/cellos are in the front because they're leading. You can simulate that with reverb predelay and attack. Moseph would have more insight into this.

Hm... I wonder if you used a microphone that recorded in mono (I know mine does ), because I think I'm hearing a slightly weird phasing, perhaps phase cancellation, when the violin recordings harmonize (this is not a big deal xD). I don't really know much about what microphones are good, but I think phasing happens less when layering similar stereo recordings because you can then literally stand to the left or right of the microphone and the way the sound compresses the air should sound different enough depending on how (bent over a little, leaning a little to the left, etc.) and at what angle from the microphone you stand. When you electronically pan, I think of it as you're panning/moving the whole signal to a side, but when you're in a room, the "panning" is just where you're standing, and however wide the reverberations are is set by the room size. In other words, DAW panning is kinda like... uh... moving your walls. Sidenote: sound travels by compressing the air, and the air next to that is stretched a bit, similar to a compression wave on a slinkie. I think the Japanese pronunciation sounds good. I had a friend who was into singing Japanese music in high school and she pronounced it similarly. The performance itself had some great expression, and I think it sounded good overall! Nice mixing on the piano! It's far enough in the back that the fakeness doesn't come through as much, but not so far back that the harmony it provides can't be heard.

Okay, so it seems like the drums are just four kicks with a snare on the 2 and 4 and some very subtle hi hats. Some more drum fills would lessen the repetition. So something like this is what I mean. You can add in more explicit open hi hats, subdivide the hi hats, put a snare off the beat, etc. The rising stab at 0:56 is pretty upfront, so that's what I keep hearing over everything else. Also, this just keeps plowing through with few transitions; 2:02 and 2:12 felt like they needed cymbals, 2:26 was sudden, etc. You have sections without drums sometimes, but they just pop in and out and I can't tell when I'm supposed to expect them.

I dunno what the vocals are saying at 1:22, and a little much fuzz (read: 3000Hz) on the rhythm guitar, but other than that this is some good horror-like metal.

Yeah, I personally don't have an issue with doing songs note for note, but the OCR submission standards ask that you do more than a 100% note-for-note cover, so that it's both original enough (but not too original or it's just an original song) and personalized enough. By MIDI-rip (well, I said "essentially like", but...), I mean many if not all of the notes match exactly. For example, this was said to be conservative, but it was personalized just enough. This was personalized a lot, but not too much. This I think is right in the middle or so. If you don't want to change the note pitches, that's fine; you could alternatively change the harmonies, note rhythms, neighboring tones (extra notes in between the original notes), the mood (major/minor), or even the tempo (case-by-case).

Ah, but isn't there a button to turn the side chain on and off on the far left?

Bottom left.

The submissions process takes more than a month =P (more like 6 months, actually, but I personally don't mind; I just submit and let the judges do it whenever) The production sounds reserved, to me. By that I mean that things blend together a little too much, and the more important elements of the piece don't come out enough (kick, snare, lead, bass). As a result, the dynamics sound too consistent throughout, and to me the structure isn't quite as clear. By the way, as good as this is, this is too much of a cover without enough personalization IMO. Actually, this sounds essentially like a MIDI rip, and so it's too conservative. But hey, now you know. At least you have some nice live playing skills.

Yep, you're telling the compressor "only affect this frequency range when you modify this sound". So if there's a band pass covering 200-2000Hz, then only that range would be compressed. Just to be clear, if you were to put an EQ plugin doing a high pass at 2000Hz and then a separate compressor next in the chain (no built-in EQ used), and compare the effect of that to a compressor doing a high pass at 2000Hz with its built-in EQ, it's not the same thing.

Glad you're going to keep going. Actually, the sampling doesn't affect enjoyment (or mine); it's just a caveat in the site submission standards, so it's something that is looked into whenever a case like this comes up. Sampling too much may sound like copyright infringement to those "automatic song identifiers" on YouTube perhaps (), but it's more about the personalization you can put into writing something from mostly scratch, from my perspective.

There's a huge amount of remixes in the queue, so the judges kiiiinda need to get through the ones sent in earliest (usually) before getting to the one(s) you sent in.

If you're thinking of chords that are like triads and bigger, Archtop: Hollowbody Guitar can do that (though it's not specifically made for rock, even if you add distortion)

These instruments sound lifeless, basically. There's an overall loudness issue. The bass is rather boomy, some lead instruments are dry, and the whole ReMix sounds the same. I don't know how you first approached EQing this, but if you were like me when I first started, then you might have a whole bunch of EQ overboosts from just trying random things. What I'd recommend is saving a copy of your project file, then resetting the EQ in this new copy on everything, and thinking about what quality of what instrument you want to remain or come through the mix. Cut the EQ of what you don't want to hear as much and of what is clashing with the EQ that represents the quality of what you want to come through. Try boosting temporarily to identify where in the EQ the good qualities are, then decide what you want to do with them. If anything, cut more often than you boost (this is a really nice read). Boosting can introduce too much of a quality you don't want, but cutting can only make things sound hollow or thin, which is not going to give you much ear fatigue usually. Unless you just *know* that you have a reaaaaally good audio system, boosting might make things worse. The key is to find a good balance. EQ in context. Don't just solo each instrument and then EQ them individually all the time; EQ them when everything else is going on too so you can tell what's different in the final result. Eventually you might have to reduce the frequencies on a quality you actually like on an instrument to make things work better, but you may not have the desire to do that unless you EQ in context. This might help some more. Think about what you want to be loudest, next loudest, and so on. If you can't tell whether or not something is too loud, just pick a loudness (try -5dB for the loudest thing) for the loudest instrument, then change volumes in reference to that. Try starting really low in volume, then slowly raising the volume until you think it's just right. In most cases (other than solo pieces or pieces with few types of instruments), something's gotta lead, something's gotta hold the foundation in the bass usually, something's gotta keep the rhythm usually, and something occasionally should give a sense of musical harmony. Try reading this. Things will stack up and get closer to 0dB, but should not go over. That stacking is called constructive interference.

I think I've used that before (badly) for like a week, but I switched to FL Studio and am still sticking with FL. Based on this, Magix CAN use VSTs, so I don't know what's not working (by the way, just for perspective, I wouldn't trust it if someone said "this is high-quality" until you try it yourself and make your own judgment on it, as companies *are* aiming to sell their products ). I like German products, usually, but not always. Besides, Impact Soundworks makes some incredibly inexpensively-priced stuff, for example, but they compare very well with the more expensive libraries out there.

The EQ defines what frequencies are targeted in the compression (Q is the band token filtering width, by the way; high Q = thinner than low Q). I've never used Live's compressor, so I don't know how good it is, but I highly recommend Cytomic's The Glue. Never stopped using it since I got it over a year ago. The Glue Example: Dry Drums Compressed Drums (there's a parallel compression send, and individual tightness compression on the kick and snare) Compressed Drums (High Pass at 2000Hz on the built-in compressor EQ; small difference but it's there. Try focusing on the snare, which seems to have gotten a bit tighter) Basically they should feel stronger overall, but also a little bit more cohesive.

25% off everything at http://adsrsounds.com/ on July 17-23. Everything!