• Techniques
  • Audio tracks with wide variation in RMS (p.2)
2016/02/12 15:11:49
robbyk
Jeff Evans
 
One good bit of advice I learned from a great mastering engineer is group all of the tracks together that you intend to put on te CD.  Choose one track out of all of them first and master that one.  Choose a track that has everything in it.  Everything going for it.  Max instruments etc.. Biggest sound etc..  Choose a final reference mastered level and master the track to that level.
 

 Yes this is exactly what I was doing by making the rough CD. Your perception is amazing.
 
Your advice is the key! Now I know to take the rowsin' song which is the full Robby K Band and get it's level right where I want it. Then match the others (mostly lonesome hippie cowboy singer solo songwriter movin' on the down road sing-alongs) to it. That is exactly the info I was looking for. Thanks big-time.
 
I have both mixes and small "m" masters of each song, so if there is substantial additional limiting to do, I want to go back to the masters for that and check with SPAN and the Klanghelm meter. Then I suspect if there is still a little discrepancy when all the small "m" mastered songs are lined up, I can use automation to match levels.
 
I turn 63 on Sunday and all the help is a wonderful b-day gift :)
 
Now to get to it.
2016/02/12 15:34:30
robbyk
I should further explain, my little "m" masters are simply placeholders for now. I enjoy learning and workin' on this; gives me something to look forward to each day as a break in the action.
 
The big "M" mastering is in Danny Danzi's hands (God willing) when I am done with the mixes of all the songs.
2016/02/13 02:19:40
Danny Danzi
A few ways to do this. Depending on the situation, all the methods above are perfect. Me personally, I go for simple. For example, I have a few strange methods that I use because I am NEVER happy mastering in one program. Something is always missing from one, some creature comfort or meter missing from another....I so need to team up with someone and create the ultimate mastering program!
 
Anyway, when I get done mastering songs, the last step is the limiter and sample rate conversion. When I'm in Sonar, I load up all the tracks like Bassjoker and Sven told you, use the fader on each channel to make them all sound the same volume with the one before, then I run a limiter on each song so that the limiter literally treats each song properly instead of one limiter doing everything. This way I can literally mess with individual volumes as well as thresholds and attack/release on each limiter.
 
When I'm done I can burn to CD right in Sonar or I can export each track individually and I'll be nice and level. I do watch my meters and try to get things as close as possible. However, certain songs will react differently depending on what instruments were recorded. You can get a clean song WAY louder than a song with distorted guitars even if the meters tell me things are the same.
 
So in THAT case, the meters will not help me if my ears tell me different. I just about never rely on the meters other than to tell me I've clipped something or gone above -0.3 or -1 when they are my targets. I sincerely think people watch this stuff way too closely. That's just me and in no way is that a slight on anyone. I don't have any Grammy's or anything, so maybe I'm not one to listen to.
 
I like doing things by ear and I don't like to hard limit unless I have to. Keep in mind, these songs are already mastered that I'm working with here in this scenario...so I've already manually leveled the peaks and compressed/eq'd etc. Hope this helps, Robbie! Looking forward to working with you when you are ready. :)
 
-Danny
2016/02/13 03:02:19
mettelus
I had a friend who had similar issue and sent me 30 of his songs to take a look at. Noise reduction was one thing, but after that step I realized most of his material fell into one of two bins so made an Ozone preset to handle each scenario but I still had the issue of the variances in what he was sending me. The simplest solution for his material (also a "just to play in the car" type of situation) was to normalize each file to -3dB prior to the mastering pass with Ozone so that the Ozone settings required the least amount of tweaking for each. In hindsight I found the normalization pass effective for the quickest workflow. For his stuff, the workflow I ended up using was:
  1. Noise Reduction and normalize to -3dB. Saved this to new file (destructive edit).
  2. Added Ozone and loaded appropriate preset for the genre.
  3. As the presets matured, I ended up spot-checking obvious portions of each song to adjust as needed (I used 4-bands for both presets).
I am not sure if the above will be applicable to you, but wanted to throw it out there. My friend's material had very consistent frequency content and dynamic ranges used, which made the situation a lot simpler for me.
 
This is definitely not a high detail/quality situation (as his material needed remixing anyway), but more a means to get consistent levels for the material he had.
2016/02/13 11:08:18
robbyk
Danny Danzi
 
Anyway, when I get done mastering songs, the last step is the limiter and sample rate conversion. When I'm in Sonar, I load up all the tracks like Bassjoker and Sven told you, use the fader on each channel to make them all sound the same volume with the one before, then I run a limiter on each song so that the limiter literally treats each song properly instead of one limiter doing everything. This way I can literally mess with individual volumes as well as thresholds and attack/release on each limiter.
 
-Danny


Well I am blessed for one thing, I am just an ol' geezer in the snow and I don't have clients depending on me for this kind of thing. I am more than happy leaving it with the pros :)
 
However, my mission was accomplished by all the helpful replies above, all of them! Thanks much!
 
I now can burn a leveled CD so that I can listen properly (usually in the car). Now I can evaluate my mixes over time.
 
Why I like foolin' around with Ozone and with this kind of thing (again, very small "M" mastering) is it often exposes a weakness or flaw in my mix that for whatever reason, I did not pick up. If I find something that works well be it EQ, compression, exciters, etc., then I have cause to ask why and go back and work on the mix some more. In the mix, I tend to know what I am doing. It only took 10 years, but I think I am getting there :)
 
So thanks to all once again and Danny, yes, Monday I forge on and I will have good material to send you, hopefully in the summer!
2016/02/13 12:01:35
Danny Danzi
mettelus
I had a friend who had similar issue and sent me 30 of his songs to take a look at. Noise reduction was one thing, but after that step I realized most of his material fell into one of two bins so made an Ozone preset to handle each scenario but I still had the issue of the variances in what he was sending me. The simplest solution for his material (also a "just to play in the car" type of situation) was to normalize each file to -3dB




Just for your head, I just wanted to explain why normalizing might fail. I'm glad it worked for you but here's the problem....
 
You're only bringing the hottest point of each file to -3. So if you are averaging -12 and have one peak that goes to -4, though the entire file will raise a bit, only that peak at -4 will get raised to -3. It is pretty rare to have a peak that much higher in a file, so I'm explaining that to you in a worst case scenario and am exaggerating a bit. But that's what normalizing can do in a negative way for some files and why you have to be careful while using it. Only the highest peaks turn out to be the target range. This leaves you with volume inconsistencies still and possible rogue peaks choosing your destiny. So just be careful. Loads of peaks in a file, consider another method instead of normalizing. :)
 
-Danny
2016/02/13 15:51:36
mettelus
+1, I definitely agree with you on this and my original thinking was more to make "consistency" for the compressor thresholds in the presets. In this case (which is admittedly quite unique), identical hardware and even tone was used end-to-end so variations were not as extreme as one might expect.
 
Ironically, this was a situation that actually backfired on me... originally I had a couple files sent to me that I did this with in the hopes I could get him interested in using a DAW to help him mix (each one needed serious mix help). I ended up getting the gambit of his wife's favorites, and that is as far as it went.
2016/02/13 16:38:09
Jeff Evans
Normalisation is not a process used in mastering at all for the reasons Danny mentioned.  I don't hard limit a whole track either often.  It is often just a few rogue peaks that are creating the issue so I tend to look at the whole track and there will often be a level where most peaks reside to and then the rogue peaks are obvious and they can be sorted individually.
 
I still feel a decent editor is hard to beat for this surgical type of work.  I find fixing a few things here and there in an editor is way better than applying a whole process over a track when it is just not necessary.
 
Personally I feel mastering is not a seat of the pants thing either.  It is one area where exact metering is almost a necessity.  I use three types of metering too.  RMS or VU metering, LUFS and DR range metering.  They are all important.  You will get perfectly consistent masters that way.  Not with your ears alone.  Every decent mastering studio I have been into has extensive metering and the great mastering engineers all use it.
 
A track may be 1 dB lower than the rest and your ears may not pick it up. You may suspect something and not be sure.  An accurate VU reading will show it to you in a second though.  Looking at the bottom end of your on mix on a spectrum analyser can also show up some interesting stuff too.
2016/02/13 17:10:03
gswitz
I use an EBU128 meter to measure average loudness.
2016/02/13 20:17:59
batsbrew
there is a touchy-feely aspect of deciding which songs sound louder or softer compared to the rest....
 
it's hard to explain, and even harder to show clearly why it is that way.
 
it's good practice,
to watch the meters.
 
but at some point, you should be make gut decisions,
not based on any vu or led
 
so, it may be that in those collection of songs,
there is clearly one that should be the loudest,
and you create the song order, based on building up to the loudest song,
and gently sloping away from it, in order of song.
 
point is, have a strategy
© 2024 APG vNext Commercial Version 5.1

Use My Existing Forum Account

Use My Social Media Account