A response is necessary here from my part:
The amount of times I’ve responded with advice here and on IRC on the issue of mixing and mastering is staggering and starting to become a little overwhelming. I often ask myself ‘how can I possibly condense years worth of experience in trial and error plus loads of reading into one succinct forum post?’. Another common thought: “How to write the advice in a powerful way so it goes heeded rather than ignored or misunderstood?”. And anything I write will be dated. These considerations are daunting and leave me on the back foot trying to weigh into these debates. No explanation can be short. Part of me would like to write a very large book on the matter, but then I’d have to find the time to do that and place it as a higher priority in my life than it is now. And even then, I’ve got my biases, my strengths, my weaknesses, my characteristic state of development that would make such a book a faulty preposition. So what do I do?
As many of you know, I’m trying to start a little business on the side with mixing and mastering work. The standard practice in the industry is for engineers not to specifically give away all their crafty secrets to preserve their business interests (and it is just too time consuming publically). However, this is no fun for all you people who are getting into these problems for the first time and want to LEARN. So, I’m back to my original problem - how to help practically.
I will write out some guiding points, but please bear in mind these points won’t solve your problems like magic, nor will they replace years of work and experience on your part to obtain a great sound. Reading through these points will hopefully illustrate the trickiness of the situation.
Without further ado:
[b]
- Digital is ALWAYS worse than analog.[/b] I can spend a lot of time bogged down in this point, but it’s just better to trust me on it. The more parts of your recording / mixing / mastering process you can export out on quality analog gear the better. Good gear is expensive. It’s hard to use and sometimes obscure and confusing. The more research you do the better. As soon as a signal hits any form of chip that converts the sound to zeros and ones then any layering of digital edits will cause various compromises to the sound quality.
2. When in digital the higher resolution of the data you’re using the better. This equates to two aspects. One is the sample rate. For example I do all my recording / mixing / digi-mastering in 96khz (all plugins perform better that this rate too). The other part is the bit-depth which equates to the dynamic detailing of the sound. Anything less than 24bit is not good enough - I use 32bit float. If you’re not convinced do some controlled A-B tests and let me know otherwise.
3. The quality of your gear means a lot. If you’ve got cheap entry level stuff you’ll get that sort of sound. Any soundcard must have a ADDA chip that has at least 96khz / 24bit capability and preferably balanced Inputs and Outputs for professional grade recording and monitoring. The quality of your monitoring is a HUGE part of your ability to make things sound great. Entry level pro stuff (e.g Yamaha, Adam, Mackie) in the right settings is fine, but nothing lower. I prefer 8" bass cones as I believe the low end presents truer at that size. 2.1 setups with subs can ‘work’ if balanced right but often cause aural confusion at the crossover point. Having more than one set of ‘checking speakers’ is a must: I have a 3 way system: Nearfields (2 different ones), Big Hifi speakers in another room, and the Car.
4. The acoustics of your monitoring environment is essential to address. Block out high reflections with sheets/material on the walls. Larger rooms are better, bigger than 4x4m. Monitors need to be placed in a position for even reflections of low frequencies and off the back wall by at least 2ft if not 1m. Monitors need to be at ear level sitting, around 1.5ms apart and 1.5ms from your ears in a triangle facing toward your ears. Also preferable if they are on separating foam to stop desk resonance. The more soft furniture in your room the better. Run a test tone on the low end freuquencies to see where resonances are.
5. Start listening to a lot of stuff on the above setup, and listen widely. Take some time to seek out stuff that has been touched by golden production. You can find a lot of top notch analog records from about 1978 to 1984 in the pop genre but in just about every genre you’ll find nuggets here and there since. Classical and/or pure acoustic recordings are good to listen to as well to understand instrument voicing in terms of tonality and mix position. This is a task you should apply yourself to over a period of years, if not decades. It never stops, it’s like constant research. Find out what you like, and then study those aspects closely. Think about how you can subtly use EQ, compression and filtering to re-create those sounds. The more you listen the more you’ll cultivate your skills.
6. Remember GIGO always. Garbage In - Garbage Out. No use in polishing a turd if it’s always gonna be a turd. There are three parts to this. Firstly, recording sound requires a lot of TLC. Avoid clipping, choose the right mic for the right job, and use tube mics and preamps if you can. Secondly, a lot of us here are writing using VSTIs and digital samples - these can often have all the negative parts of digital audio, such as dullness, flatness, muddiness, harshness, alising etc etc. Refer these digital sounds back to acoustic sounds and you’ll realise you’ve got a lot of work to do to improve them with the given limitations (more on this later). Thirdly, keep your skills up as a composer and musician - poor arrangements often lead to difficult mixes (although this is hard to articulate). And if you’ve written sub par music that’s boring or annoying, why go to all the trouble to polish it? Soul first, presentation later.
7. Mixing well requires much hands on experience with both analytical mixing techniques and ‘trial and error’ experience. The more time tweaking the better you get. The number one piece of advice I give to beginners in mixing is that they’ve got too much low end and everything is ‘muddy’. Digital lends itself to this sort of conundrum because of the lack of stuff like tape saturation or lack of signal loss that you have in an analog system. Start relatively low-shelf EQing everything below 200hz reductively. If you’re adding: 99% chance you’re wrong and you have to address your set-up and listening cultivation. In EQ it’s nearly always wrong to add - reduce the parts of the sound to leave just the nice part that you want. Emphasis on ‘nice’ even if it’s highly textured distortion or static stuff. Then boost the output gain to bring the sound back up in the mix after EQing. Common nasty areas I find myself reducing time and time again are 500hz, 2000hz, and 200hz-300hz for mud. Every input sound needs it’s own EQ curve specific to it’s tonality and voicing. EQ on solo first, then re-adjust with the whole mix or at least with the kick and bass. There’s loads more I could write, but this will do for now.
8. Adding brightness to a sound is better done with either an Exciter or via Saturation. If you can’t do this via analog gear or via reel to reel tape then there are a handful of plugins that do an ok job for both channel and master purposes. A good compressor will lift the tone of a sound too, a bad one will just make it sound like sludge. Saturation will clean up your transient (spikey peak sounds) nicely as well.
9. Compression is neither good nor evil, just a tool. The more you play around with compressors and learn what they’re all about, sorting the quality ones from the crap ones, you get a sense of where you would be using them and how. Beginners tend to over-compress - I tend to now use subtly ratios of dynamic reduction depending on the input sound combined with saturation. Limiters (which are a type of hard compression) can be used for really spikey stuff like snappy snare drums - gentle usage after EQing can tighten things if need be, depending on your style. I will note that due to a lack of tape saturation in digital audio that sounds are generally more spikey and undesirably lively. If you apply your learning of compression with the aim of preserving the clarity of your input sounds you can eventually find workable compromises. EQ sounds different before and after a compressor - make sure you check your order of effects.
10. Get some technical knowledge up about sends (or buses) and Mid-Side processing. Both are pathways to group and order complex sound in neater ways. With sends a common thing I do is group certain sounds together for common processing (say all the vocals, or all the low sound, or all the drums, etc). Mid-Side processing is very usable for the final management of complex mix sounds where it is easier to control things like overall width and punch.
11. A ‘light and bright’ mix makes for easier mastering. Don’t worry too much if the overall mix doesn’t feel like it’s heavy enough, or punching enough - let the mastering take care of that. The less bass amplitude and mud crowding the low end of the mix the more headroom the mastering process has to work clearly on the grouped sound. If your mix is too bassy the first thing that hits the mastering limiter is the bass and it pushes the higher part of the sound into yucky squashed sounding territory. Some of the best mixes in the world are quite conservative on the bass (yes even dubstep for example) because once the song is aired on a big PA, the oversized subwoofers are doing the RIGHT amount of work to make is sound massiv. The nicer all your mix elements are the more the song can be pumped up really really loud on a sound system and ENJOYED. Play My Bloody Valentine up really loud and see how long you last without the pain… Overall I check my mixes on a spectrum analyzer - the more the average impression of the curve is like Pink Noise the better. No big low end. Generally flat from 50hz up to 7khz, rolling off from there up.
12. Assuming you’ve got your mix right, mastering is all about the detail of a signal chain and how much of the fidelity you can preserve in the process. Now this is where engineers can quibble over the detail quite a bit, but let me try and impress upon you a basic chain approach: -> Mono low end -> Reductive EQ -> Mid-Side processing -> Saturation -> Compression - > Limiting - RMS leveling meter. Limiting is the part where a lot of masters get ruined - people pump up the gain too much and end up making sludge out of their song. Read up on The Loudness War. Everyone has their own position on it. My recent masters for most electronic and pop music come out around -10dB RMS, but different clients ask for different levels depending on the style of the music. Classical can be right down around -18dB RMS for example. A lot of cinema audio is at -14dB RMS. Then we move onto the next point…
13. Dithering and downsampling is ok if done digitally right as the VERY LAST STEP, but can be bypassed if you master your work to tape. I won’t write too much on tape mastering here as it’s not an option for most people. But you must try to keep all your work at least at 44.1khz 24bit, preferably 96khz 32bit-float or higher, right through your entire production from recording to mixing to very end of mastering. Downsample first - from your higher sample rate to 44.1khz for general playback on the web/ipods/CDs etc. Dithering is a conversion process that adds mathematical noise to the conversion of high bit rate data to 16bit data. There are loads of dithering algorithms out there - I’ve generally used a shaped gaussian approach for final masters and a triangle algorithm for stuff that needs further editing on it (different engineers will argue otherwise). …The whole gist of tape is bypassing all this faulty computer processing with the added bonus of tape saturation sparkle and flow, recorded directly back into your soundcard at 44.1khz 16bit. And wha-la you have a finished file. If you’ve done everything perfectly as above then your Mp3 at 192kpbs or higher (or any other format) will sound like gold.
14. Re-assess, re-test, evolve and be open to new ideas. No point being a stuck in the mud- it’s important to evolve and improve both yourself, your process and your system. For good engineers this is a never ending process, and they never stay reliant on the same ‘all in one’ solutions or lazy quick fixes. Get advice: test your work on other people, both industry people and outsiders. If you be stubborn on any given point be prepared to face the repercussions and responsibilities for that. You reap what you sow. And if all else fails get a pro to do it. Pros are great at having ZERO EMOTIONAL ATTACHMENT to your music and can see the sonics for what they are in a reliable way, regardless of genre. If you music is good enough it’s well worth the financial investment.
Hope you all find this useful in some way. I’m happy to answer any particular questions.