Sound quality and recording

This may be a bit offtopic, so I'm sorry if MG isn't the place for this questions. It turns out I'm not only new to modular but to music production too, and there're some basics I still don't get. I hope someone can point me in the right direction.

The issue I'm facing is about sound quality in general. When I record my music it doesn't sound as clean or proffesional as I would like. My setup is quite simple at the moment, this is what I'm doing:

• One MI Plaits sequenced by a Minibrute 2, the out of the Plaits connected to the master input of the Minibrute so it gets mixed with the main out.
• The main output of the Minibrute goes into a desktop mixer.
• I also have one Drumbrute connected to that mixer.
• The main output of the mixer is then connected to the computer line-in for recording.
• In the computer I'm using QuickTime player for audio recording and I tried adding some compression with the Audicity software too.

I think this are some of the problems, but I'm not completely sure and I don't know how to fix them:

• I'm using a cheap mixer and it may be adding some background noise. Well, this is easy to fix, get a better mixer, but... is it my main problem?
• Using the computer line-in may be adding even more background noise. Do I need some kind of controller o external gear to properly record my music?
• QuickTime player doesn't sound to be the best option for audio recording, should I use something more like Ableton? is there anything simpler? I've never used Ableton and it looks pretty complex.
• Audio compression and equalization: I really don't get these two, I tried adding compression because it seems like everyone is doing it, but what is it? how does it work and how does it affect sound? Regarding the equalization: how can I get it right? how can I know if the mix is well equalized? how can I mke sure it sounds right in different headphones or speakers? and What gear or software should I use for equalization and compression?

There may be more problems with my recording process and I'd appreciate if you guys could point me in the right direction. Also feel free to suggest articles, videos, books or anything else that you think could help me understand the basics of sound recording.

Thanks!

OK, let's answer some of this, sort of out of order...

Not only will a cheap mixer cause noise, there are definite sonic differences between something cheap and something that costs more. And the difference there comes from component quality. Cheap stuff (like Ammoon, Alto, Harbinger, et al) cuts corners on components, with the result being looser tolerances, which sort of cascades as your signal path goes through the board. One sub-par component is bad enough...now consider what a couple dozen of them in an audio chain will cumulatively do. Plus, certain mixers have a very specific sound quality to them, most notably the English-designed/made ones. This is what makes a pre-Behringer Midas desk so desirable...but not so much a post-Behringer one, as these don't have the same rounded "English tone" anymore.

Computer line-ins aren't the right thing to use, nope. The culprit here is noise at your A-D conversion stage. This is due to the A-D on a typical sound card (or sound card on the motherboard, depending) being typically unshielded from in-case electronic noise, plus the fact that that connection is going to be a consumer-level (-10 dB instead of +4) line-in and it's also unbalanced, which tends to allow more electronic crud into your signal chain. The line-out isn't as problematic, but to get really good results on recording, you need an outboard interface that's +4 dB, takes either XLR or 1/4" TRS balanced lines, and has a proper ground. And one other point: everything in a recording setup should be star-grounded. By this, I mean that everything you use needs to have a ground that is the same as all other devices, usually done by grounding everything to a single ground point (hence the name). By doing this, you can lower noise and help avoid ground loop issues.

QuickTime is not only the wrong tool, it's also VERY out of date. Use a proper DAW. You already have Audacity, so try recording in that instead. I actually multitrack in Ableton 10.0.6...but I chop loops and clips and also do my final editing and normalizing in Audacity. It works better for that, while Ableton works beautifully on multitracking, track comping, and so on. Ableton is also not the only choice; you might look at Bitwig, which is similar but has some of its aspects more streamlined than Ableton Live.

Now for the last pile of questions...first, EQ. Technically, there's three types: parametric, graphic, and program. Parametric is the type where you can specify the frequency per band, the level at that frequency, etc; you often see these on mixing desks in some form or another. Graphic EQs are the ones with fixed frequency bands with level controls, and tend to see more use in live applications for room correction, but can also be useful for similar purposes in the studio. And program EQs are things such as Pultecs, where you have specific boost/cut stages with their own tailored frequencies, often also working on the overtone spectrum of the selected frequency. This last bit is very typical of the Pultec EQP-1A's low end cut/boost control, where the 'boost' also works on the overtones of the selected frequency, but the 'cut' acts like a normal shelf, with the -3 dB point at the frequency.

For the most part, a program EQ is the only EQ you should boost levels on. All other equalizers should be used to subtract from what's present in the raw signal unless you're using the EQ as an effect in some way. The reason for this is that it's easier to compensate for lower levels of something in a mix than it is to correct levels of some type that're too hot. For example, let's say that one track has a band in the lower-mids that's sticking out, a sonic 'lump' as it were. It would be easier to isolate the 'lump's' frequency and reduce that on that one track than to bring everything else up in various levels and bands to even out the 'lump'. But with a program EQ, what's being done is more akin to "sculpting your mix's tone color"; accordingly, most of the time you'll see program EQs on the final mixbus to do those timbral adjustments.

However, tinkering with EQ without a good monitoring chain...flat, unforgiving response from as low as is feasible in the bass all the way up to the ultrasonic...is basically pointless. It's like trying to read a map, but you've forgotten to put on the reading glasses you need...ergo, you're probably going to get lost. Never skimp on monitors...unless, of course, you're trying to check your mix on a more "real-world" equivalent, in which case you need to incorporate those "everyday" monitors alongside the other, more precise ones. And this, btw, is how you check your mix; if you need to know how something sounds on, say, a typical set of computer speakers, by all means use some of those after you've done your mix on the mains. But if something needs fixing as a result, do that work back on the mains again. Motown studios always had a pair of 6"x9" car speakers in some cobbled-together wooden boxes in their studios specifically because Berry Gordy wanted to know how their stuff sounded in your typical car...and of course, Motown stuff sounds great in the car because of this "check". Headphones, however, are not something you mix in unless you're specifically mixing for headphones.

As for compression, there are again several types. Limiters basically "smash" everything above their threshold level and hold the dynamic limit right there. More typical compressors have various (and often adjustable) settings for how aggressively the compression happens as the desired level is approached, plus what sort of degree of compression (ie: ratio) is needed. And program compressors, like their EQ counterparts, are more for riding gain and "gluing together" a mix while used on the mixbus. As for the right way to use these, first keep in mind that anything over 4:1 ratio winds up behaving and sounding like limiting, especially with a hard "knee" (that "aggression" setting) at the threshold level. To get a compressor to behave transparently, use lower compression ratios and softer "knee" settings, which will then allow the compressor to compress over-level signals enough to fix level problems but not to make the track in question sound like its being "mashed". Unless, of course, that's what you want, since compressors are also useful for adding distortion and overload character to sounds that can use a beef-up.

Program compressors, though...those are a bit different. In their case, you use compression to get the overall stereo level on the mixbus to "float" around the desired track's loudness without exceeding 0 dB. So the meters on a program compressor might be floating at around -3 to -5 dB, and you'll use the makeup gain to bring that result's level up to where you need it to be post-compression. These are a good bit trickier to use well; like anything else in music worth doing, they require practice.

As for what to use...that's up to you, and what sort of sound you're going for. A good place to start, though, would be KVR Audio (https://www.kvraudio.com/), which has a trove of free plugins. You should, over time, be able to find the ones that work for your music and workflow...but again, this takes time, because in this process you're actually tailoring your DAW to be your bespoke recording "instrument".

Hopefully some of that is of use...

Thanks a lot Lugia, it is very useful and I'm sure I'll be re-reading it a few more times while I do my research on the topic.

There are entire professions devoted to recording and mixing (I'm in one of them). I don't think a couple of paragraphs are going to do it in order to get you where you need to be. I'd recommend taking the time to learn some theory and practical knowledge. The most important part is experience... and you only get that from trying and making mistakes. I made several with EQ and compression when I first tried home recording back in 1991.

My first recommendation would be to buy a decent audio interface. I'd shoot for something in the $300 range for a beginner. This will dramatically improve your results once you've mastered gain-staging. I would also invest in a DAW. Reaper is surprisingly affordable at around $50US for non-professional use. The great thing about a DAW is that you can EQ, compress, etc. inside of the DAW... save all your settings... and then revise your mix later on. You'll also find it valuable for learning and practicing.

A quality pair of headphones and if you can afford them a quality pair of monitors will help. Bear in mind that when using speakers, the room that the speakers are in will ultimately impact what you're hearing with your ears.

Hey thanks Ronin1973,

I've been looking around and it looks like something like the Focusrite Scarlett 18i8 would be a good place to start, what do you think?

And if I understood correctly I could plug my gear into the Audio Interface and then to the computer via USB, I wouldn't need the mixer at all. This way I could even record multiple tracks at once using something like Reaper, which looks great by the way, thanks for the suggestion.

Then, having the sound in different tracks would also be easier to equalize, right? let's say I record the mix using Audacity and all I have is one audio file, no tracks and so, could I still equalize it properly? I guess it's more difficult and the results wouldn't be as good.

Anyway, do the same concepts regarding equalization and compression apply when playing live without a DAW? do I need equalization modules and so on? or is it something completely different?

The Scarlett would be a good choice for recording. Here are some issues you may encounter: all audio interfaces introduce delay between the time the sound hits the interface until its recorded and played back through the DAW. This is called latency or sometimes referred to as lag. Most mid-range and higher USB interfaces offer the ability to mix the direct sound (your synth) and audio coming from the computer so you can monitor without lag. As long as you're not trying to listen to what you're recording through plug-ins IN your DAW this is an acceptable solution. You can get interfaces with very little delay (practically zero) but the cost goes up substantially.

The other issue is that Eurorack synth level versus line level (the level of operation in mixers and like gear). Synth level is hotter than line level. It's very likely that you can just adjust the input volume on each Scarlett input to compensate. Do your homework if this will work with the Scarlett. Ask around.

If you record each part on its own track (you'll need two inputs for tracks in stereo), you will have them isolated. You can process isolated tracks separately and is the best way to mix/edit your music. If you record everything as one stereo file you are basically stuck with what you have. Any effects (EQ, compression, reverb) is applied to everything in the mix... which is something that you probably don't want.

If you're playing live without a DAW, you'll have to decide if you want/need EQ, compression etc. You'll also need some way to provide them. I don't see too much of this inside of most peoples' racks. But they'll often use an external mixer that has EQ built in. There are several types of EQs and a reason to use each of them if the material requires it. Without getting into a big philosophical, technical diatribe, I wouldn't worry about it in the beginning. If it sounds okay, it is okay. Having EQ, compression will make your live mix better. But if you don't know how to use them properly (for now) it's more trouble than it's worth. The exception would effects like reverb, delay, chorus, distortion, etc. They are very important in defining your synth sounds and shouldn't be overlooked.

Next chapter... SYNC. aka how to lock your Eurorack tempo to your DAW's tempo for overdubbing more parts into the DAW for later mixing/editing. Do a little research and report back. :)

There's a lot to learn from your starting point and no easy way to explain it without you doing a lot of hands-on by yourself. You may... and I say may... want to subscribe to an online course regarding mixing and DAWs since there's so much to explain. You'll also get a better understanding about signal flow between your modules and your DAW.

Thanks Ronin1973, all that information has been super useful!

So far I've been able to sync my synth with Reaper. In fact it was easier than I expecte, both the Minibrute 2s and the Drumbrute Impact have USB port and all I had to do is select the device in the Reaper settings and send clock signal out from the DAW.

I've also been experimenting with recording my songs in multiple tracks, one track at a time as I don't have any audio interface yet, but it has allowed me to test things out and make sure I understand why I need the audio interface and what I need it to do.

I also have more doubts and questions though, how do you guys do the multi-track recording? do you group multiple instruments in one track or do you strictly use one track per instrument? I find myself easily mixing more than 5 tracks at a time, would I need to get an audio interface with that many inputs? or should I be doing overdubbing with the DAW?

In my mind I would be doing everything with the modular synth and once ready, I would use the DAW for recording all at once in multiple tracks, so I can do the final equalisation in the computer. But I'm not sure that's doable with an audio interface that only has a couple of inputs.

I'm also facing an issue I forgot to mention in the original post. It is about the volume of my recordings, when I listen any of them, all sound way lower than any other audio so on my phone I would need to go full volume to listen it properly. I'm already 'normalizing' the audio in Audacity, I'd expect it to bring the sound to the max volume, but it doesn't seem to work that way, what am I missing?

How I multitrack is pretty much how I've always done it, even back in the 2" tape days. I'll have separate instruments on each track that comprise the basic parts of a piece, recorded in one pass via an Orion32 (if external to the DAW) or direct in Live (if I use internal sources), or both as needs be. Once all of that's down, then I'll start doing overdubs while at the same time starting to work out what processing to layer onto the initial tracks. At this point is where we diverge from tape technique, though. In beginning the mixdown process, I try and combine related tracks as submixed stereo stems...say, all percussives on one group, bass and pads on another, "ear candy" bits on a third, and so on. By submixing and then subprocessing these stems in this way, you actually have quite a bit of control over the main mix with a minimum of faders in play and a minimum of CPU load because, once the stems are tracked, you can turn off all of the processors you used on the individual channels in the stem, plus your submix is now under the control of a single stereo fader pair and, when needed, you only need to add processing across the two tracks of the stem.

Stems are sort of a given these days with a lot of producers and engineers, but it wasn't that long ago that they were a rare thing, doable only when you have the massive budget needed for extra multitrack machines, tape, etc. With all of that tech out of the way, though, you can generate stems whenever you like and however complex you can deal with inside your DAW of choice. Then, once your stems and solo tracks are ready to go, the mix gets easier...you're not juggling a couple dozen faders all at once. Also, DO automate things such as levels, etc within your stems so that that 2-channel result is exactly the way you need it.

Mind you, this tends to take buttloads of practice to get used to envisioning how your mix should work prior to even mixing it. Ability comes with time and diligence.

Now, as for normalization...that process raises your overall track levels relative to the highest level. So, if it takes +8.5 dB of change so that the loudest peak comes up to your normalization threshold (and never, EVER normalize to 0 dB...always leave "excursion room" of 0.5 to 1 dB below 0 dB in case something gets raised by dithering, codec artifacting, etc), everything in the track gets raised by that amount. This doesn't equal apparent loudness, though. It just means that your peak level is where it goes and everything else went with it.

To increase apparent loudness, you need to use some form of compression. So, let's say you're cutting a track and your peak levels are hitting -1 dB, but your overall level outside of those peaks is about -12 dB. That's a pretty wide dynamic swing between peak and average, so what you'll want to do is to compress that 11 dB swing down to, say, 4.5 dB. Once properly compressed, your peaks should be back at that -1 dB level, but your average will now be raised by 6.5 dB, ergo the apparent loudness of the track is higher. Ultimately, you could even "brickwall limit" a mix so that everything is squashed into 1-2 dB of swing (or less, if you're some kind of sadist), but when you lose your dynamic swing, the track will just sound like a loud band of sound with no variance.

Mind you, all of this is for nothing if you don't have adequate monitoring while tracking and mixing. Especially the latter. Trying to get a good result with a pair of computer crackerboxes is akin to trying to read an important document without the aid of reading glasses if you're blind as a bat at close distances. You literally will not have any idea of what you're doing outside of certain inferences about what the end-result will be on everyone else's listening platforms.

And one other point along these lines: once you have your rough mix set up, you'll want to put your last set of processors on the DAW's mixbus, with the program compression last. That way, any changes to the signal levels caused by equalization, enhancers, stereo imagers, etc will still get dealt with properly just prior to being tracked or rendered.

Optimally, I prefer to break out of digital for initial mixing of stems and then for controlling stem levels; I simply like having the faders in hand for tiny adjustments. And doing this sort of mixing in analog on a quiet system with 24-bit audio (even at slower sample rates) still puts any noise and garbage signals way down in the mix where they'll disappear into the Least Significant Bit ranges when the track is reencoded at 16 bits for CD and other distribution methods. That is, if I want that; sometimes certain noises and noise amounts can actually add a bit of a presence.

Congratulations on getting the sync going. I looked over the 2S patchbay and there's also sync (clock) out. This will be useful if you have to share sync with other modules. If that's the plan, then look into clock dividers/multipliers. I use the Temps Utile for this as it can divide, multiply, sequence gates, Euclidean, etc. Sometimes you will want modules to be triggered at a different rate than your master clock and not all modules can divide/multiply on their own. Also check into the "reset" functionality as a useful way to always keep modules in sync. This will be useful if you need to reset a secondary sequencer to its beginning position or reset an LFO to the beginning of its cycle.

At the moment I have a 2 input audio interface (Scarlett 2i4). It's not very practical for recording multiple parts at once. I'll be upgrading to an 8 input interface by the end of the year. Another option is the Expert Sleepers ES8. It has 6 inputs at synth level and can output USB audio to your computer or ADAT audio to an interface that supports an ADAT input. You'll have to decide what's practical... especially if you're using a computer that can't stack multiple interfaces. Again, read, read, read...