what exactly happens when i press "Audio Mixdown" (Cubase)
Home > Home Recording Forum > Mixing Techniques > what exactly happens when i press "Audio Mixdown" (Cubase)
Posted on Sep 19, 2007 05:40 am
crux
Typo Szar
Member Since: Jul 04, 2002
I dont know if anything even actually happens, but i just feel like when i go Export>Audio mixdown, its doing "something". This "something" is along the lines of using somekind of compression to turn the many trax of my song into one stereo file right?
My question comes from this, if it makes any sense, i got feel like, i do mixes that im pretty happy with and when i mix them down they dont sound the way i mixed them. A very obvious example is sometimes ill get the reverb for my vocal JUST the way i want it, only to find that my .wav file has alittle "too much" reverb.
I thought about it and thought that maybe to mix down a file digitally the computer has to figure out somekind of baseline, like say 0db and then it calculates things you do like eq cuts and boosts and prints a song based on what is more or less than that baseline. Ill an example like, if i boost only my bassdrum at 5k so i get a good click sound going ni cubase, when i export the file everything is very quiet, with that click sound dominating everything, so makes me think that the computer is saying "oh he wants this, that other stuff dont count as much".
Am i right>? or am i just blaming technology for my own short comings at mixing... i have to say im pretty disheartened at the moment, ill spend hours on a mix and the end result is always just short of what i want.
[ Back to Top ]
Sep 19, 2007 06:28 am It 'sums' your files into one file. Its the same as recording off your stereo outs. It encodes the sum of these outputs into a digital representation of the audio.
I've noticed that it does subtract a little something from your mixes too. Thinking about buying a HHB or the like CD recorder because of it.
McMerkinHobbyist musician,pro recorderMember
Since: May 15, 2007
Sep 19, 2007 09:34 am Audacity has a feature called Quick Mix that seems to do the same thing, I get the levles right where I want them and I try that feature and everything changes.
So I simply don't use that feature anymore.
But even summing two channels into one with very careful level placement still tends to change things enough to be noticable. Of course, when working with mono/stereo I tend to forget about the 3db punch the center will now be, so perhaps on my end it is more operator error than the program itself. Still, it IS frustrating.
cruxTypo SzarMember
Since: Jul 04, 2002
Sep 19, 2007 09:37 am
I knew i wasnt going crazy! hahaha
so is there anyway around this? are there anymore parameters i can set to get a "what you hear" type mixdown?
wat should i be considering when im mixing if i have the knowledge that these alterations will occur?
Sep 20, 2007 04:48 am hmm i've never really experamented with summing...a just assumed if you do the "mix export" feature on the daw it'd be the same as playing it back in real time....
if that's not the case, how about mixing down to another device? like DAT or even tape? as in run your stereo output buss to another separate device and recording from that.
Sep 20, 2007 05:07 am I think you can do a real time mixdown, I'm not sure that it makes a difference though.
WaltChief Cook and Bottle WasherMember
Since: May 10, 2002
Sep 20, 2007 08:03 am Very interesting thread!
I'm kinda with WYD here. I was under the impression that the algorithms utilized to create the playback would the same used in mixdowns with only a redirect to file. Now my ears or old, so it doesn't supprise me that some of you younger folks would here more than I do.
I will say this; I have always had the 'sensation' (the word used as my intrepretation and or explanation) that when playing back multiple channels from my HD24 through an analog mixer, using that mixer to sum the signals, the result was 'cleaner' than importing the same files into Cubase and using Cubase to sum the audio for playback.
A much larger difference to me is in the mastering process. I start with a 48k 32bit mixdown and mash it into compressed (L2 or other maximizer) 44.1k 16bit wav. In that process I notice a difinate dumbing down of the content. Obviously there are numerous things effecting that process.
Sep 20, 2007 08:05 am Quote:
I think you can do a real time mixdown, I'm not sure that it makes a difference though.
Well, the first thing that pops into my head is to play back the whole track, bouncing it down to a stereo track while playing back...if it doesn't mixdown "like you hear it", bouncing it down would probably work.
That said, I haven't used Cubase in years, so I dunno...
cruxTypo SzarMember
Since: Jul 04, 2002
Sep 20, 2007 10:26 am
I tried a realtime mixdown today, but i have dont my monitors to really check the outcome, but i can honestly say there is a big difference.
Im starting to think the normal mixdown is built for speed, so u can quickly pop out a demo or something, but the real time mixdown is the "Real deal" mix, this is just my own speculation though.
Sep 20, 2007 11:14 am how do you do the realtime mixdown in cubase?
WaltChief Cook and Bottle WasherMember
Since: May 10, 2002
Sep 20, 2007 01:49 pm Found this on another site. This is news to me, but it do make sense:
10. If you activate Real-Time Export, the export will happen
in real time, i.e. the process will take the same time as regular
playback.
Some VST plug-ins require this to have time to update correctly during
the mixdown – consult the plug-in manufacturers if uncertain.
Noize2uCzar of MidiAdministrator
Since: Apr 04, 2002
Sep 20, 2007 03:44 pm I was just gong to say. Sonar does real time mixdown so I would assume some other apps do it as well.
And ya, there is a bit of a difference, not a huge one in Sonar that I can tell though.
cruxTypo SzarMember
Since: Jul 04, 2002
Sep 20, 2007 09:46 pm
Wow, thats really shocking, ive been doing my mixdowns like this for years!
You can do realtime mixdowns by when pressing Audio Mixdown, look at the bottom left of the window, it should have a selection box for real time mixdown.
As far as i can tell on Cubase, it makes a big difference and not just in terms of VST plugins working properly, but i sware by my ears that the "summing" it does is somehow different from standard mixdown. all i know is that its defenitely not a "what you hear" mix with they im doing it now.
BeerHunterwww.TheLondonProject.caMember
Since: Feb 07, 2005
Sep 20, 2007 09:57 pm ...what exactly happens during a real-time mixdown in Sonar? I only mix and then export. I mean I'll bounce down tracks and **** but how can I take advantage of "real-time" mixdown? It is sounding like I'm missing something real cool here.
Sep 21, 2007 04:37 am wow Bob Katz got the second reply...
[quote]There is absolutely no difference! In fact, if you add "dub to an external digital recorder via AES/EBU" as a third alternative, there is absolutely no difference among the three choices. This is the biggest urban legend yet. IF YOU HEAR A BONIFIED DIFFERENCE while cutting these files in two different ways, then take the resulting files, they completely null, they are bitwise identical.
Done numerous tests on this precept with alternative files sent in to us by clients. Blind and non-blind.
BK[/quote]
cruxTypo SzarMember
Since: Jul 04, 2002
Sep 21, 2007 04:55 am Have i tapped on to something weird here?
From most of the posts seems like people either realized it sucked and moved on, or just never thought about it.
Im actually really frustrated that after all these YEARS i only started to pick up on it now. Even though i dont know how to rectify the problem yet, im still very sure that alot of the doubts i had about my mixing before might be directly linked to this problem.
EDIT: just saw WYD's post after i posted, i seriously sware on my ears that i hear a difference. And if there is no difference, how does he explain the fact that people all over the world even brought it up. Bits and bytes aside, something is happening, maybe its just an issue with working in teh digital realm vs analog, i dont know, but somehow my mixes arent coming out the way im doing them and its not just coz im not "there" yet. I admit im not a real sound engineer or even close to one, but then shouldn't my mixes sound the same summed or non-summed either way?
WaltChief Cook and Bottle WasherMember
Since: May 10, 2002
Sep 21, 2007 07:03 am I can think of one simple test that in theory may shed some light. Do two identical mix-downs with the exception of doing one 'real time' and the other 'standard'. Load the two mixes into your prog and phase reverse one of them. The result if identical should be silance. Obviously that still does not test the origional question; 'is there a difference between how these progs play back and render the music'. At this juncture, with folks claiming that timing of plugs are an issue, I would also do two identical mix downs to see if they differ per computer utilization variances as well.
Interesting.
Sep 21, 2007 03:48 pm walt: brilliant idea!
BeerHunterwww.TheLondonProject.caMember
Since: Feb 07, 2005
Sep 21, 2007 04:01 pm ... still trying to grasp what you guys are talking about...
I'll wager a guess and say that a real-time mixdown is listening to the playback while it bounces to a new track and standard just bounces like what I'm used to but.... help!
Noize2uCzar of MidiAdministrator
Since: Apr 04, 2002
Sep 21, 2007 04:09 pm BH, the standard mixdown/export will render in real time. I only use the quick mixdown when I need it right away.
And no, there really should be no difference at all between the two. Doing the quick one only takes a chance on having a pluggin jitter, that is all.
BeerHunterwww.TheLondonProject.caMember
Since: Feb 07, 2005
Sep 21, 2007 04:13 pm Cheers noise! I get it now. I thought I was missing something but it just sounds like different terminology between the two programs.
Noize2uCzar of MidiAdministrator
Since: Apr 04, 2002
Sep 21, 2007 09:04 pm Yep, that is all it is. Just different words to describe similar actions.
I think the thing with other programs is you must choose to do a real time mixdown as they default to a quick mix.
Sep 22, 2007 03:40 am in bob's other posts he mentions dooin' null tests and comming to the same conclusion....what kind of differences are you hearing between the two crux? volume stuff? reverbs popping out? tonal changes?
cruxTypo SzarMember
Since: Jul 04, 2002
Sep 22, 2007 05:24 am
Its tonal changes, i think thats how to describe it. Like an eq that i did on say cymbals will sound fine in the app, but mixed down theyll sound washy. A common problem i have is getting a good amount of bass, and after mixdown it sounds thin, on the same monitors that i was mixing with, not changing to a different system.
After reading the thread at GS i noticed its about pro tools, could this be a setback of cubase? They said that protools has "corrected" their math, could it be that steinberg hasnt?!
but yes, to me there is a big difference and even now with real-time i cant quite put my finger on it. But another example is when i was doing audio mixdown, i used a Classic Reverb for my vocal and it sounded just right in the app. and the settings were relatively low, but when summed, there was a very noticeable amount of verb on my vocal. My whole band agrees on that problem, that our effects arent processing the way we added them, so im quite sure its not just me
WaltChief Cook and Bottle WasherMember
Since: May 10, 2002
Sep 22, 2007 10:11 am First test in:
As a control test I used a single mono track of a guitar recorded with a microphone. I put the same track on two mono channels in the Cubase 4 mixer. I did a phase reversal on one of the tracks. As expected, I got silence during playback. Obviously I could mute either track and have playback.
The second test was to use two disimilar tracks, this time same recording of the guitar but using one track of a mic on the body and one track of a mic on the neck. I performed a mixdown without the switch for realtime mixdown activated, i.e. a 'quick' mixdown as it were and had it imported back into cubase as a seperate track. I did a phase reversal on the mixed down track and got silance on playback until any of the tracks were muted.
Obviously this does not take any 'plugs' into consideration, but proves that this mixdown is exactly what is heard via playback.
WaltChief Cook and Bottle WasherMember
Since: May 10, 2002
Sep 22, 2007 10:43 am Ok, second batch of tests:
Utilizing a fairly complex project using various plugs etc, I did two mixdowns; one quick and the other real time. Again I imported them back into cubase and did phase reversals on both tracks. Each mixdown was a seperate test. Neither test produced silence. The real time export produced a more muted playback indicating it is 'better' than the 'quick' mix.
Neither were descrete.
Crux, your ears are correct. Good listenin dude!
cruxTypo SzarMember
Since: Jul 04, 2002
Sep 22, 2007 12:04 pm
Wow, thanx for the help Walt, i barely put that much effort into my own problems. So for now i know.. im not crazy
BeerHunterwww.TheLondonProject.caMember
Since: Feb 07, 2005
Sep 22, 2007 12:18 pm Cheers Walt... excellent info
WaltChief Cook and Bottle WasherMember
Since: May 10, 2002
Sep 22, 2007 12:35 pm This was an odd 'journey' for me. We do have a shared interest, so it is not 'selfless' on my part. I realized that although I thought I heard a difference, it never occured to me to research it. I simply did what I could to get the best results I could e.g. changing mix, mastering, etc. It was your inquiry that made me sit back and think about how to test the system. It would now interest me to see if other progs tested under the same conditions yeild better results. These results could actuly give substance to arguements of one prog over another. It took no more than a half hour to perform the tests using existing projects.
This process did give me something tangable to work with. I will do 'real time' mixdowns for final mixes. I will no longer simply blame my psyche for differences that I precieve. I will continue to 'repair' mixdowns, but with greater confidence. It's all good...just not perfect. Hummmmm..... Like everyting else.
HuePinnipedal Czar (: 3= Member
Since: Apr 11, 2004
Sep 22, 2007 09:10 pm I'm obviously late with this but, nice one crux... there is a difference. I've never been able to put my own finger on it, but there is deffinately an audible difference.
I remember talking with Noize about the audible differences of mixing for cassette as apposed to disk, and the difference of mixes from back in the day, when a cassette was 'it'. I would say that the 'real time vs. export' difference could be 'heared' simmilarly. Everything is just a bit more... hmmmm...'evident', for lack of a better term.
I've noticed it, and I use Sonar.
Sep 22, 2007 11:07 pm Crux, you've got better ears than Bob Katz! :-P
Interesting stuff, thanks for the reuslts Walt.
Sep 22, 2007 11:34 pm Walt and I are out of phase from reality about the same number of degrees - probably about 270 - so I think we understand each other. Correct me if I'm off, Walt.
Excellent analysis, by the way.
Sep 23, 2007 03:30 am wow cool info, yeah i was aware that Katz was workin' specifically in PT, so the results could be different in other recording apps....
good investigating on you Walt!
and yeah, null tests can reveal alot on many different situations.
Sep 23, 2007 05:03 am sorry to bump again, but the thread on GS is getting rather interesting, and is very much relevant....for those who are interested, keep an eye on it....
cruxTypo SzarMember
Since: Jul 04, 2002
Sep 23, 2007 08:53 am
Yeah WYD said it, but i think i should give a +1 that Bob Katz was using PT, they always said PT was "better" and though i still love Cubase, im sure each app has its strengths and weaknesess and its starting to look like this is weakness shared by non PT apps.
So some guys r saying that maybe i should just run busses to an outboard recorder... that is interesting actually, i think it would yield more "what you hear" type mixdowns consiering its basically like sticking a recorder into the part of ur chain where ur monitors use to be, maybe even give a chance to add back some warmth and beef?
could TAPE be making a comeback for non PT users?!?!
Jul 28, 2008 03:30 pm Interesting thread and reading it I too was wondering this very question.
I hate to burst the bubble but there is NO difference whether you are exporting real time or offline.
The test is easy. Simply bounce the mix and import it back into your project and on that channel reverse the phase.....as already mentioned.
With both offline and real time I was presented with total silence, indicating no variation between the running mix and your exported file.
Im not sure how you did those tests Walt but perhaps you had some plugs on the master??
Who knows.
I think that any percieved discrepancies are entirely psychological.
But hey, its all part of the process in finalising your track.
Its always best to listen to your mix in different environments and then come back to it after some time, all the pros do this aswell - and perhaps there are some changes that you should make but its because you need to adjust the mix not because of some notion that your equipment isnt working right!
CptTrippsCzar of Turd PolishMember
Since: Jun 20, 2006
Jul 28, 2008 04:36 pm One thing to check, which is an option in Audition is to have it downsample and dither on mixdown. I disable all these functions until I am ready to burn to CD.
WaltChief Cook and Bottle WasherMember
Since: May 10, 2002
Jul 29, 2008 08:27 am EC,
You are confusing me a little with the term "Master", but when I did the tests with no plugs on any of the tracks I did get a perfect mixdown indicated by total silence. When I added plugs to the tracks, (and I don't remember which plugs were used) I no longer got silence. These tests were performed in Cubase in the mixing stage of a project. I did not perform any tests while mastering.
cruxTypo SzarMember
Since: Jul 04, 2002
Jul 29, 2008 10:02 am I'll have to second Walt here, i think the difference in the mixdowns is more than psychological, if not then y would there be two different functions anyway?
The fact is i think the difference comes into play when u add plugins to ur mix, anywhere on any track. Even in the cubase help section it says that the live mixdown function is designed becoz some plugins react differently, so ofcourse there is that difference. I'm thinking that those differences to the plugin reactions r quite great coz things like reverbs and such that ive used come out totally different between the two mixdowns.
Jul 29, 2008 03:59 pm Walt, i mean plugs on the master channel. If you had say an eq/comp/limiter on there then that would make a difference as your bounced audio would be running through that x2.
Also CTripps mentioned that you may have bounced down to 16bit or 24bit. To get a true representation of the live mix one must bounce to 32bit float stereo.
I read the Cubase help and youre right Crux, it does mention that some plugs need to bounce in real time to update correctly and certainly this could have an impact but I am guessing that this statement could be in regards to VSTs like the Virus TI that MUST export in real time.
It would be good to know what plugs you were using, because I use a lot of plugs and vsts and I got a total match.
Perhaps you may even of had a plug with a free running oscillator? That could make a difference also.
Anyhow, I suppose its horses for courses but Im pretty confident. like Bob Katz, that Cubase does export accurately and if there are any discrepancies its the fault of the plugin and not the host.
WaltChief Cook and Bottle WasherMember
Since: May 10, 2002
Jul 30, 2008 08:58 am Ok, Got you EC,
No, there were no plugs or master section involved. No dithering or sample rate alteration either.
As far as "fault"; my perception is that fault matters not. I haven't experienced a mix that has been effected enough to bother isolating the plug involved. I look at it as "nature of the beast". Theoreticly, per my perespective it could be "blamed" on hardware as well. I can imagine that some hardware may not react to the code used in mixdown the same as others.
My motivation was to find a way to test the assertion made by Crux that What You Hear is not What You Get. I found his assertion to be correct per tests. And yes, the deviation accured with the introduction of plugs. I do use a fairly wide array of plugs and doubt all have been tested by Steinburg for phase alterations in mix down. I don't know Mr. Katz and obviously don't know what constants he uses in his findings. It is simply suffice for me to know that there can be phase alterations in a mix down.
Regardless, perception is the only critical evaluation criteria per dealing with music. To date I have experienced no mix downs that I have found objectionable in relation to the audio heard pre-mix down.
