JoshJ Posted March 30, 2016 Share Posted March 30, 2016 Hey guys, When bouncing stems (I include Send fx so tend to solo each track individually and bounce one by one) how important is it to set the signal chain up (track fader, bus fader & master output fader) to optimum levels, instead of just setting the master output fader to the optimal level, to get the overall bounced signal to around 0db? In other words, would I be loosing loudness on my stems if I left my track & bus faders fairly low and just adjusted the master output fader so the level is around 0db? Link to comment Share on other sites More sharing options...
triplets Posted March 30, 2016 Share Posted March 30, 2016 to get the overall bounced signal to around 0db? I don't think this should be your goal. Because if all your stems are around 0dB, your master output is gonna clip anyway and you would have no room to mix the stems. Louder doesn't necessarily mean better sounding. Link to comment Share on other sites More sharing options...
JoshJ Posted March 30, 2016 Author Share Posted March 30, 2016 Also guys, how much headroom do you recommend leaving on each track? Have just bounced a bunch of tracks with around 3db of headroom after the peak point of the signal. Is this adequate? Link to comment Share on other sites More sharing options...
triplets Posted March 30, 2016 Share Posted March 30, 2016 I have the individual tracks peaking around -3dBFS tops. I try to avoid clipping.. If you want to have headroom on your final mix for compression and mastering, make sure the stereo output track does not clip and the peaks are around -6 and -3 dbFS. That is if you care about your music's headroom. Some people don't and like it slammed to the tilt. Link to comment Share on other sites More sharing options...
btfnk Posted March 30, 2016 Share Posted March 30, 2016 I'm guessing you're printing stems to be sent to someone else for mixing? Or something along those lines. As long as the stems don't clip and you output the files as 24bit, you'll be fine. 24bit has so much resolution that the difference between them peaking at -15db and at -0.01db is effectively the same for this use. As long as nothing crosses 0 and clips there's no data lost. Intersample peaks don't exist in a digital file, so the whole needing headroom thing doesn't matter as long as the file can be turned down before it's output to analog. However just as there is no need for headroom, there is no reason not to leave it, it'll just save the step of turning things down later in the process, there's 144db of signal in an undithered 24bit audio file, leaving 10db unused at the top won't hurt your fidelity. Link to comment Share on other sites More sharing options...
JoshJ Posted March 30, 2016 Author Share Posted March 30, 2016 Thanks guys. Yes, I am bouncing stems to sent to a mixing engineer. I mixed the track first to create a close reference for him. So just to clarify; in future can I just leave all my faders how they are set in my original mix, and bounce all the stems at different volumes (some relatively low) and not loose out on much loudness in the final mix? Link to comment Share on other sites More sharing options...
JoshJ Posted March 30, 2016 Author Share Posted March 30, 2016 Also, when you refer to dBFS, im not quite sure what the implications of this are. I know it means digital full scale but, when you say -3dBFS do you just mean -3db on my Logic fader (ie, just beneath 0db clipping)? Link to comment Share on other sites More sharing options...
btfnk Posted March 30, 2016 Share Posted March 30, 2016 Yes you can leave the faders as they are in your mix. As long as no individual channels are clipping. The normalizing setting 'overload protection' can be helpful in making sure nothing clips, however if something does get lowered in volume by it, then it's relative volume will be off compared to the rest of your stems. That really shouldn't matter to your mixer though, as long as you send a bounce of your master as a reference mix also. The stem level will just be a quick starting point in their mix. Yes for what we're talking about you can pretty much just think of db = dbfs Link to comment Share on other sites More sharing options...
shivermetimbers Posted March 30, 2016 Share Posted March 30, 2016 This doesn't make any sense. You are having someone else do the mix, why send stems? You can also export the tracks as 32bit (float) and hopefully they can import them and adjust as necessary. Link to comment Share on other sites More sharing options...
btfnk Posted March 30, 2016 Share Posted March 30, 2016 There's plenty of reason to send stems instead of a project file, the most obvious being if they want to mix in a different DAW like pro tools. Nothing wrong at all with sending 32bit files which will also alleviate any fears about slight clipping. Link to comment Share on other sites More sharing options...
JoshJ Posted March 31, 2016 Author Share Posted March 31, 2016 Yes, i am sending to someone who will be mixing in pro tools. Is there much quality difference with 32bit float? Does exporting at 32bit actually increase the quality of the audio if the original file is say, 16 or 24 bit (as most of my samples probably are - not sure what bit rate virtual synths generate sound at?) ? Is there any compatibility requirement for the mix engineers DAW when importing 32bit float on the other end? Link to comment Share on other sites More sharing options...
shivermetimbers Posted March 31, 2016 Share Posted March 31, 2016 Once you bounce your tracks/stems, they become fixed in the 24bit format. If you happened to clip one of the files, it is too late to repair (unless you or they have sophisticated/expensive software editing tools). Using a 32bit (floating) file, protools can import and readjust the file back to a normal level with no clipping. So if you have a good sounding 24 bit file, there would be no reason to send it as a 32 bit FP file. In that case, there is no difference. Discuss it with your mix engineer for direct advice on what to send. Link to comment Share on other sites More sharing options...
MikeRobinson Posted March 31, 2016 Share Posted March 31, 2016 I suggest that you keep in mind what it really is: a file of (integer or floating-point) numbers. Millions of 'em. Floating-point numbers work like a scientific calculator. Integers have some fixed "maximum possible value." Simply make sure that each stem makes good and full use of the available range of numbers ... to capture detail ... without ever(!) bumping into that "maximum possible value" ceiling. (If that should happen, data would be irretrievably lost.) Don't try to mix 'em relative to one another: that's their job. Your job is to supply the data. The "downstream" operations that will now take place during mixing will start with this "data set." They can never add to the data you supply: they can only alter it and subtract from it. The levels of any track can be adjusted (normalized ...) individually, during that process, as long as the data, in each and every track, is there. They can throw-away some of the data that you supply if they choose to, and of course they will, but they can never create a satisfactory replacement for data that isn't there. As a rule of thumb, when you're using a computer to process any sort of data (music, images ...), "you can go down but not up." You can reduce something without losing detail, but you can't stretch it. ("A big image can be made smaller with no loss of perceived detail, but a small image can't be made bigger.") Supply them with data that they can reduce. Let them set the relative levels between your sources, such that they're always turning it "down," and not "up." By all means, ask the downstream engineer to give you his "wish list," and strive to make his wishes come true with what you send him. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.