For those about to submit audio for mastering please read bullet points below.
Session Mixes Submission Requirements:
- All session mixes without pre/post fades.
- All session mixes to have a pre/post gap to the perceived audio for at least three seconds, including any floor noise from mixing system/recordings.
- Submit session mixes at the same sample rate and bit rate as the project if mixing in the box. If out the box submit at the highest possible rate from the final mix AD conversion.
- File format must be standard PCM i.e. bwf, wav, aif
- Do not apply any master bus processing to the session mixes – i.e. limiters, EQ or compression.
- All hard disks/pen drives and optical disk submissions as data PC/MAC ISO format. Only submit as CD audio if no other version or source exists.
If possible in addition submit:
- An instrumental version of the mix. (i.e. The same mix with all vocal parts muted.) Or full stem session mix break down (Breakdown sheet available on request.)
- A sequenced version of the final album/EP with fades/cross fades and track spacing.
If you’re interest in why these requirements are important please read on…
Fades and top/tail:
A session mixes supplied with fades; this is the most frequent error in submission. Easily done as when mixing the focus is upon get the emotive effect right, hence why we fade. But the issue then arises when processing in the mastering stage as a fade goes though the transition it will passed the threshold of any dynamic/tonal processing which will proportionally then be less/more effective.
These problems lead to a step in the fade as the dynamic is no longer effective and/or any tonal change will drop proportional in desired colouration. On a long fade these artefacts are both difficult and time consuming to correct effectively.
Always submit session mixes without fades but if there is an important transition supply a faded version as well for reference. A fade with the same intent can be applied to the mastered file without the unwanted artefacts above. This will obviously save valuable mastering time.
Top/Tailed or trimmed files cause similar problems. If a file has been cut tight to the start/end of the perceived audio it is likely, when mastered with a reduction in dynamic range of over 10 db, that the equivalent raised noise floor will now not sound like natural start/end but jump in/out as a step.
It's important to maintain the pre/post noise floor as to give the mastering engineer space to ramp the appropriate entrance/exit. Without these tails a session mix end often needs to be recreated with reverb, again a time consuming hence costly artefact to avoid.
The other issue is noise prints, without a few seconds of pre/post tail noise it's impossible to use a noise print reduction system if the apparent signal to noise becomes too much after mastering dynamic restriction and tonal balancing. This is often the best way to reduce broad band noise with minimal artefacts especially on quiet intro/exit sections.
The other great use for start/tail noise is track transitions. It's common, if not essential, to record 'silence' for live concert work. Thus help join musical pieces without going to digital black, which would make the feel of the transition unnatural. This same good practice can be applied to an album sequencing to give flow. It is in part these seemingly small details that can make a fantastic album, amazing!
File format and rates:
It’s important we maintain audio resolution throughout the recording/mixing and mastering process. We should only down sample or reduce bit depth at the final stage of mastering when creating the differing output formats. Equally we should always record and mix at the highest sample and bit rates that can be practically achieved.
In 2011 this would be a minimum of 24bit 48khz if not 32bit float 96 or 192kHz. There is no rational to the nonsense you sometimes read or hear that we should record at the same rates as CD to avoid conversion. This is truly crazy; the basic concept of dynamic range is we have in 144db at 24bit and at 96db at 16bit. If you’re recording at a higher bit depth you’ll capture more detail, mix more detail, and finally sum more detail. Those of you who have mixed at 16bit and move to 24bit are aware of the increase in definition obvious in the summing tracks in the box. We don’t even have to go into why a higher rate of sampling captures more detail!
It’s important we maintain the same sample rate and bit depth as the project making sure we do not mix down at a lower rate. This will introduce sample rate conversion, dithering or even worse truncation of the bit depth, all of which will lose important detail or introduce harmonic distortion. We’ll discuss this dither and truncation issues in another article.
It’s imperative we are aware of these issues and do not accidently convert a file. Even up sampling can introduce artefacts if not done correctly. All of this should be left for the mastering engineer to do when and if required. They’re best placed to make these decisions to avoid adverse effects on the quality of the project.
Master Buss Processing:
Any processing on the master bus is basically what the mastering engineer will do – apply processing to the stereo mix! So why put any on if you're going to have the mastering engineer use their skill, tools and good judgment to apply these effectively. Any processing applied to the master bus at the mix stage will only affect the possible outcome of the mastering process negatively, in various ways; One; the quality of the tools used will not equal that of the mastering house. Two; any compression/limiting/EQ applied will directly restrict the ability of the mastering engineer to achieve correct dynamic restriction or tonal correction of the track.
There is no rationale to apply processing but the skilled mix engineer may apply a quality master bus compressor (SSL/Neve/API) on the master buss achieving one or, at maximum, two db of full band downwards compression to add 'glue' to their mix. This is perfectly acceptable. A limiter/multiband is not. If you feel you need to apply a multiband to your mix to make it work you need to look again at the composite parts of the mix, as this is where the problem lies. Limiting your mix is just madness – this should only ever be done by the mastering house after the track has been correctly tonally and dynamically enhanced.
Audio CD submission is a last resort, apart from the fact that CD audio is at a low rate (16bit 44.1kHz) it is also likely the original audio has been dithered/down sampled or truncated to be delivered in this format. The other issue is error correction; CD audio optical format by design has error correction. Basically without going into to much detail if a disk has been burnt well, with care (I’ll discuss this in another article), it will contain errors, however, they are fully correctable. If the disk has been burnt by either under or over exposure and/or jitter has been introduced it’s very like the disk will have correctable errors but these are not guaranteed to be the same as the original (i.e. the error correction makes it up!) or even worse, non-correctable errors both leading to thinning or even glitches in the audio.
All this can be avoid by delivering audio as data, even on an optical disk, because it has 1:1 error correction the same as a hard disk/pen drive. It either works or it doesn’t. This for a mastering engineer is ideal, 100% or nothing.
Attention to detail is key in achieving the best possible outcome for any project. Every error in the chain degrades our goal of the best possible outcome for the audio.
Supplying instrumental versions has a dual edged bonus; firstly it can help the mastering engineer achieve easy adjustment of the vocal content of a session mix without having to apply corrective processing to the overall mix. By a simple sum/difference matrix we can phase invert the full mix against the instrumental, leaving the vocal elements only. Corrective processing can be applied to the vocal stem, which could just be as simple as gain change but more normally EQ or de-essing, to clean up the vocal in the mix. This is added back to the instrumental mix recombining with zero artefacts on the instrumentation but the vocal is corrected.
The second bonus is with the expansion of new media and digital TV there is a bigger call for music for Sync. It’s essential to have instrument versions for this process to allow music supervisors to place music as back ground for adverts or TV programs etc.
The above corrective concept can be taken a stage further with a ‘Stem’ breakdown. This is the composite parts of the mix, Kit/Bass/Guitar/Vocal/Backing Vocals/Synths delivered as stereo files which when combine make the original stereo mix. These facilitate the mastering engineer the ability to rebalancing and correct mix issues on any parts required. This is an excellent method for mixes which are less than ideal. Though time consuming, excellent results can be achieved.
Stems delivery is also something becoming more common place because of game formats, such as Guitar Hero or Rock Band, which are a growing money earning stream for artists and labels. These are easier to construct at the time of mix down and not months later when request. As we all know, mix recall can be full of issues when time has past, gear has changed, versions of software updated - something to bear in mind for your next commercial project.
One last particularly useful and time saving effort is the creation of an album sequencing example. This is the session mixes to be master formatted in sequence, top/tails, fades and cross fades applied as reference for the mastering engineer to use in the construction of the final mastered album. This can save considerable mastering time as the editing guide avoids the to and fro of reference copies and minor changes.
To concluded, as a mix engineer if you take these points on board you will end up producing a better quality end product in conjunction with your mastering engineer. Equally you’ll save mastering time, making budgets stretch further and provide the possibility of other revenue schemes for labels and artist inherently filtering back down to all the audio engineers in the chain.