What levels are YOU mixing down to for mastering???

[dpaterson]

Hello all.

I'm really getting the hang of this I reckon i.e. working on something that I personally think sounds wonderful (at least to my "palate" as it were). However: after getting what I believed to be a REAL satisfying mix I started experimenting (what's new???) and now have some questions or, rather, would like to know what others around here are doing in this regard and why.

What level are you mixing down to, from Mixcraft, when preparing for the mastering process???

I spent a lot of time researching this on the Internet and, as per usual, there are as many opinions on the "correct" level as there are people making music. So I'm just curious as to how my fellow Mixcraft producers and recording engineers (sounds good huh!!! LOL!!!) etc. etc. etc. are approaching this.

My thoughts at this time:

I don't think that the same headroom rule applies at the mixdown stage. The reason I say this is because if I follow my own standard input level / track level headroom rule of -6dB / -12.0 LU at mixdown: the mastering software has to do too much (in my opinion). I usually create three different masters to compare using IKM's Lurssen Mastering Console, iZotope's Ozone 8 Advanced, and IKM's TRackS-5 "The One". If I mixdown to -6dB / -12 LU (iZotope Insight's default for headroom monitoring) then in Lurssen Mastering Console I have to raise the input gain CONSIDERABLY and this just doesn't feel right to me. Ozone 8 Advanced doesn't really seem to care i.e. it just boosts the whole mix to oblivion (LOL!!!). T-RackS-5 "The One" does pretty much the same thing as Ozone 8 Advanced (to a slightly lesser degree is all). Now there's obviously a lesson in this i.e. the absolute best, incomparable, results I get are with Lurssen Mastering Console (note that in all cases I'm just using supplied presets). BUT: the said mix is definitely not as "loud" as with the other two. So what is "right" or "acceptable" (and please don't tell me "it's what I hear" i.e. once you've listened to these three different mixes about a hundred times they all start sounding the same I'm afraid and there has to be some type of technical preference / guideline to all of this). After all said and done: thus far the "best" result I get is to mixdown to 0dB / -10 LU (iZotope Insight's default for loudness monitoring) and push this through Lurssen Mastering Console. No input or output gain is then required and it just "does its thing". But whether or not "its thing" is compliant with certain standards I know not (hence this post).

Any tips, tricks, ideas, opinions (you get the picture) would be greatly appreciated.

Regards,

Dale.

[freightgod]

I have no idea what I'm doing at mixdown!

If I'm as happy with my mix as it sounds like you are, I'm there!

I look forward to a good fresh discussion on this topic though. I hope a standard emerges actually but I'm guessing there's a thousand nuanced answers to your question.

[dpaterson]

Hello.

If I'm as happy with my mix as it sounds like you are, I'm there!

Yip. For once I'm not complaining!!! LOL!!! Then again: this is right up there with the saying that "beauty is in the eye of the beholder"!!!.

What I think I'll do is post screenshots of the waves of the different resultant mixes and masters, given the exact same mix only mixdown level being variable, for the sake of interest along with why I think one sounds better than the other (they actually do indeed all sound different and now that I've given my ears a break from the weekend, and taken the volume down a notch, I can clearly hear what's my favorite and, oddly enough, it's NOT Lurssen Mastering Console this time around i.e. it obviously does depend on exactly what the project is so there's another variable).

One very interesting little nugget that I found was (as I understand it) this business of using peak levels is no more (or should not be) i.e. apparently "the world" is trying to get back to perceived average loudness as a measure. Now from my limited understanding: that's exactly the difference between these mixes that I'm mucking about with i.e. it's relatively easy to get mixes with no clipping (obviously) but the difference is in the loudness. And here it gets more interesting i.e. loudness is / was, as I understand it, based on RMS levels. HOWEVER: PERCEIVED loudness is apparently something totally different and is why we now have the LU scale (which, although not knowing much on the subject up until now, and by sheer luck, is what iZotope's Insight uses and what I've been basing all my headroom and mixdown levels upon) (I was getting real lazy at one stage and using Hornet's VU meter for everything but I discovered an issue with it that, so far as I know, they're busy trying to fix).

Ah well. Let's see what everyone comes up with.

Regards,

Dale.

[TrevsAudio]

I've done Ian Shepherd's mastering course as well as Mastering with T-Racks 5 from Groove 3. Both prefer starting with files at -6db in the peaks of the song. This obviously leaves headroom for stuff like EQ, compression and final limiting at the end of the chain.
BUT - it also depends on your genre and planned export platform. Personally, I don't like stuff with the crap squashed out of it and the waveform looking like a solid black line, but some of you metalheads may - enter, the loudness wars! I prefer some dynamic range, and also bear in mind that platforms like YouTube, Spotify, Itunes etc will bring your mix down if it exceeds their norm. And in this case a more dynamic song will be less affected by this.
Ian Shepherd has a free site to check these various levels here - https://www.loudnesspenalty.com/

My 0.02P/Cents worth

[TrevsAudio]

I've done Ian Shepherd's mastering course as well as Mastering with T-Racks 5 from Groove 3. Both prefer starting with files at -6db in the peaks of the song. This obviously leaves headroom for stuff like EQ, compression and final limiting at the end of the chain.
BUT - it also depends on your genre and planned export platform. Personally, I don't like stuff with the crap squashed out of it and the waveform looking like a solid black line, but some of you metalheads may - enter, the loudness wars! I prefer some dynamic range, and also bear in mind that platforms like YouTube, Spotify, Itunes etc will bring your mix down if it exceeds their norm. And in this case a more dynamic song will be less affected by this.
Ian Shepherd has a free site to check these various levels here - https://www.loudnesspenalty.com/

My 0.02P/Cents worth

[freightgod]

Conditions CASE A
1) just about allow -6db in the peaks of the song
2) allow platform to apply proprietary compression algorithms.

Tey gonna doed it anyway. Don't compound it.

[Ian Craig]

Hi,

I use iZotope RX6 and like to maintain as much dynamic range as is feasible in the shape of the resultant waveform (feasible meaning that the thing has to be audible and also of a reasonably pleasing volume without adjustment of volume levels when listening). I normalise files to -1.7dB due to the fact that it is suggested in various places that reprocessing to mp3 needs -1.5dB (at least) to leave headroom for the reprocessing not to cause clipping. I also work on stuff then go back and work on it again and the crushing distorting effects of limiting-maximisation are cumulative (an effect that my whispering improvised vocals have, which can be masked or fixed due primarily to the fact that they are buried in the sound and not the entirety of it). As regards mixing on active projects I don't even think about it except to avoid it getting too loud and distorting and so when it does I start turning everything down (often far too extremely) and merely boost the levels through normalisation in RX6 after it gets mixed without distortion. I have used Ozone 8 Advanced quite a lot for it's Spectral and stereo Imaging and have let it's maximiser do it's business a bit too much so I seem to have started to leave it out of the processing chain. I like active EQs currently the McDSP AE600 and the McDSP CompressorBank CB303 as it is very good for both compression and active EQing and also has definable filters AS WELL. My major problem is with screech frequencies around the 3khz range and sometimes I use the CB303 for that and sometimes the brilliant, simple and cheap HoRNet DynEQ (which has a great tutorial video on Youtube). I tested the output of both on a track using the spectrogram of RX6 to see the frequency volume curve alterations and found the HoRNet easier in getting precise results as it's relatively simple to understand what it's doing. They both produce a slightly unique result, either of which could be better depending on what you want for the track you're working on. For years, when played back on other equipment, my tracks suffered from horrible effects of booming the muddy range around 315hz, the hiss around 2.2Khz and the screech frequencies around 3-5khz. Unfortunately, recently when listening back to somethings from months earlier I noticed that I had effectively killed them and they sounded like inaudible flat rubbish, so it is important to keep reassessing things. This is clearly why professional albums have always been recorded, then mixed a month later in a different studio dedicated to mixing and are still months away from release. Therefore separation processes in multiple strands of getting albums out is vital. It will always be a hit or miss thing, i.e if you want to make brilliant memorable music from scratch you have to live in it at multiple different times and develop a relationship with each track and then master it into an album using EQ matching capabilities of things like Melda's MAutoDynamicEQ, though there may be simpler and possibly better EQ Matching tools that I can't recall offhand, in RX6 advanced for example, they may be better or worse, I'm not sure because I haven't used those functions enough, but if I was putting an album together I might have to use them, but usually I just use the spectrogram to visually analyse things and my ears. Sorry, I've rambled through all this, away from all the question about levels, "It's good to talk" though

[dpaterson]

Wow!!!

Much to process up above let me tell ya!!!

Just popped in here to post about a little trick I've just started using (not expecting to see the above) (which I will get to shortly).

I just realised that instead of using the standalone versions of all of mastering applications: load your (final mix) audio file into Mixcraft and load them as plugins and work from there. It seems to be better or, at very least, give you a bit more insight into, and control over, what these things are actually doing. Possibly the best mixes now (this morning anyway) have come as a result. Essentially: load the mastering plugin and then load a DECENT metering plugin after e.g. iZotope's Insight. This way you can monitor ALL of the levels coming out of the mastering plugin and adjust accordingly. Insight gives you more info. than the built in meters in the mastering plugins so far as I can tell. Well. Put another way. Lurssen Mastering Console doesn't give you much by way of metering so at least doing it this way I was able to work out exactly what to feed it at the input and how far I could push the output with regards to peaks and the three LU levels that Insight measures.

And, this morning anyway (these things change!!! LOL!!!), I have to say I've been experimenting with mixing down to -6dB / -10 LU (again) in preparation for the mastering software. To be honest: although the mastered track isn't as loud "on paper" (from just examining the waveforms) there doesn't seem to be much difference in the PERCEIVED loudness BUT the upshot of this being that all time and effort that I've spent trying to create the illusion of space and width (the reverb thingy again) is NOT getting lost in the mastered output. In this particular case: if mixed down to 0dB before mastering then the mastering software pushes the mix and many nuances seem to get lost (whether they're ACTUALLY lost or just perceived as being lost I know not). A good example in this instance is my kick drum (here BFD3 shines too I have to say). In the dry track you can actually HEAR the kick drum pedal (mallet) strike the kick drum (in this case vellum I'm sure) and you can hear the air in the drum. Push this too far in mastering and this all but dissapears (for technical or scientific reasons unbeknown to me that is).

Regards,

Dale.

[freightgod]

Wow I think it's great your just trying all sorts of stuff right now. That's when recording/mixing is at it's most fun. Sometimes. But be careful that what sounds like a best process evah "as of this morning" might not just be your ears liking a fresh sound.

I wonder also if the target medium doesn't make a difference also. For example, if I'm mixing for an album or extended ep where I can expect my adoring audience to listen through a few tracks at once, I can expect them to bring their (the adoring fans) listening devices to comfortable listening levels. I might give myself complete reign over dynamic range because I'm Paul Simon for crying out loud, etc...

But if I want my annual grammy award nominated single to stick out of the Pandora playlists, I might as well mix with full awareness of my consumer's listening device, which is probably 75/25 earbuds/uber.

I say mix as opposed to master because the older I get the more I'm a 'master-in-the-mix' believer. I'm open to experimenting with mastering plugins again, but I haven't used them in years.

Great discussion. Please summarize your what your proposing as a workflow once in a while, it'll be easier to follow all them words for us old folk.

[Ian Craig]

Sorry, the rest of my answers got erased when I hit a wrong key somewhere . So to be brief, try out Sonible's software. It's expensive but brilliant (and gets reduced periodically). I have their SmartEQ+Live on the master track every time now. Their stuff is the fastest way to sort most problems out. Look it up, try some out.

[dpaterson]

Good afternoon all.

My head still spinning a bit from Ian's post (had to take a break!!! LOL!!!). Still not done but will come back to where I left off with it and then post (if there's anything I can add that's worthwhile of course).

freightgod:

But be careful that what sounds like a best process evah "as of this morning" might not just be your ears liking a fresh sound.

and

I say mix as opposed to master because the older I get the more I'm a 'master-in-the-mix' believer.

Well you're "talking my language" when you say "master-in-the-mix". That's utopia but I do believe it's achievable. And the reason I say this is because with all of these tests something strange actually happened to me: I must have, oh, about thirty different mixed down and subsequently mastered tracks and had many of them open in Sound Forge (that's like my "end of job" tool for just inspecting waveforms, normalising if necessary, and some one or two other good uses). In order to attempt to compare some of these tracks I was simply alternating between each of them (easy by clicking the tabs for each open file at the bottom) and previewing them. One jumped out at me and when I looked at which file it was: to my absolute horror (and pleasant surprise too) it was a mix down straight from Mixcraft with NO mastering!!! Alright: then I carried on comparing the others and, well, that one sort of "got lost in translation" (probably as my hearing got more accustomed to the "pushed" stuff). But the point is: it is possible for sure I believe. And I am also reminded of something that Andrew Scheps has said time and time again in his videos and interviews: he stops with a mix when he has reached a point where (for that day anyway) there is nothing beneficial that he can add to or change in that mix. And as Ian noted too: the mastering is done in different studios and by different people (engineers) (actually have heard this before too) and there's good reasons for this obviously.

Now I actually have been wanting to post a post about which sampling rates to use and why e.g. are there benefits to, for example, recording at 48kHz, saving as 96kHz, doing everything in 96kHz, until you're ready to distribute the final product (all that 32-bit floating point of course). AND THEN I FOUND THIS to my absolute astonishment (and found loads of other links etc. to back it up):

The topic of choosing a sample rate, and downsampling is also sometimes an exercise in sanity. Here are the facts: 44.1 kHz, 88.2 kHz, and 176.4 kHz are sample rates for audio mediums. Think CDs. If your audio ends up on a CD, that’s what these are for. The sample rates of 48 kHz, 96 kHz, and 192 kHz are for video mediums like DVD’s, blu-rays, etc. The idea of “more is better” is inaccurate. It should be “more is different.”

A link to the entire page is here: https://theproaudiofiles.com/6-facts-of ... bit-depth/

Now not that I didn't know about 48kHz for DVD (video editing another little expensive hobby of mine) but had no idea about the rest.

So. Guess I'll be working at 88.2kHz from now on (even although my notebook ain't too shabby in the bigger scheme of things it battles at those high sampling rates depending on the plugins in use obviously) and, finally, mixing down to CD quality (I too am an "old timer" it would seem and CD quality is good enough for me i.e. it's what I know!!! LOL!!!). BUT "Houston we have a problem": I just noticed that mixing down to 88.2kHz is not an option in Mixcraft. Somebody care to tell us all why??? Never thought about it before but now that I do: it's the only software that I have on my system that doesn't allow you to select 88.2kHz as a sampling rate!!!

And it does make a difference as I've just done some tests i.e. mixing down from Mixcraft 96 - 96, 96 - 48, and 48 - 48. In the 96 - 48 and 48 - 48 files: there are peaks that are HIGHER than in the 96 - 96 files (obviously using the exact same project / clip with no changes).

One other thing in my tests (and notwithstanding the above and working at 96kHz througout in ignorance): I noticed that on my left channel I had some kick drum peaks that were skewed higher than the right channel. So using Sound Forge and something called "Wavehammer" (which really is just a compressor): I compressed those few peaks BEFORE sending the file through mastering software (so that BOTH tracks had the same peak value) and, for some reason, it seemed to make a difference. Why didn't I use a compressor to do this in Mixcraft??? Because I didn't want anything (and everything) ELSE compressed to begin with. And secondly: at mix down you don't KNOW where those peaks are relative to the rest of your audio. As to why it made a differnce in the mastering software??? Hazarding a guess I'd say it's the same reason why NOT eliminating spurious peaks like these would cause an issue if you were normalising audio (both tracks at the same time) i.e. the software would normalise both channels only by the amount specified and by the net change on the track with the highest peaks. In other words: if your highest peak was at 0dB on the left channel and your highest peak was at -1dB on your right channel then a normalise (and who knows if this applies to mastering software) would do nothing if you were normalising to 0dB. Worse still: if your RMS levels were nevertheless correct and you normalised only the right channel in this example then your whole audio would be skewed as the right channel would be brought up and could result in the overall RMS level being changed (which could most certainly affect the overall artistic intention). So. There's a little trick I just taught meself (been around a while you know!!! LOL!!!).

Now about that 88.2kHz issue in Mixcraft Greg???

Regards,

Dale.

[Acoustica Greg]

Hi,

I personally think 88200 is a silly number.

Greg

[dpaterson]

Hello.

Sorry. JUST posted a new thread asking about this (in a crass attempt of saving you from having to read through all of my personal musings and experimentation).

HOWEVER: now you have to read the contents of my previous post (toward the middle) i.e. turns out it's not really that silly a number!!!

Regards,

Dale.

[skarabee]

Hello
Just chime in to add my 2 cents. After years and years and several albums (pop, rock, reggae, acoustic, electro) I'm now in a "mix and master" mood. I also remember sometimes when a rough mix sound better thn the 14th attempt of a self made master...

My workflow now is to have some mastering plugins setups. Depending of the style and the sound I want to render, (and also the support) I have some plugins stacks with different brands and settings, sort of basic self-made presets.
I take care of headroom during the mix process (-6db is a good advice), and ONLY when mix is ready, I load the bunch of plug in the master channel and activate. This way I can still go back and forth between the pure mix and master channel without having to feed the "render/load/master/listen/reload" loop
Of course, you may need a solid machine to hold both mix and master in the same flow.

[dpaterson]

And (drum roll) enter MELDA again to save the day!!! LOL!!!

A thing called MRecorder!!!

Placed it on the master out track, with my interface set at 88.2kHz / 32-bit and, guess what, it "mixes down" my audio to, guess what, 88.2kHz / 32-bit IEEE Float!!!

I can only ASSUME it's doing the exact same thing as Mixcraft's mix down option???

Anyways. NOT an ideal solution as most folks don't have the luxury so, if I may be so bold as to ask, please remedy!!!

Regards,

Dale.