View Full Version : Levels (revisited)

10-14-1999, 11:42 AM
For those who are mastering in PT... does the following procedure sound sensible? In the process of evening out levels for a 10 song record, I'm started by getting basic levels for each track. At this stage I want to approximate the levels of a standard CD and burn a CD-R for the client to listen to for EQ and dynamics control. I plan on getting a consistant average level for each tune after client evaluates each tune for EQ and dynamics. The process I'm using to get this inital CD build seems too time consuming. Basically, I'm setting levels for each tune, bouncing to disk, and burning.

Is there a quicker way? For example, if PT had a bounce-to-disk batch program I could set levels, select all 10 tracks, go away for an hour while the files were bounced, set Toast to burn, go away for another 1/2 an hour and I'd be finished with this pass.

Is my current approach the optimal aproach?

In the next phase I was planning on importing all of the songs into a single session and somehow getting a consistant average level for the entire record. Any tips on how to do this or on another approach is Greatly appreciated.


Camilo Orozco
10-14-1999, 11:56 AM
Hello Bluemt,
If you are using PC, you can use Steinberg's Wavelab, It has a Batch procesing function called "meta-normalizing", it analizes all your tracks and then it normalizes all of them to have the same RMS level. It does it VERY good, and you can leave the computer doing the job without being there.

Good luck
Camilo Orozco

10-15-1999, 12:11 AM
Thanks but I'm not using PC.

10-15-1999, 09:35 AM
There's a product called Barbabatch, I think that's the spelling. I'm looking into soon myself. It will do what you asked.

10-18-1999, 09:15 PM
Where can we find "barbabatch"?


10-18-1999, 11:57 PM
If you haven't already been there, try http://www.digido.com/homepage.html

This is the site for Digital Domain, with LOTS of pertinent info on mastering and setting proper digital levels, as well as arguments AGAINST normalization processing(!)

I've downloaded tons of info to read and learning a lot.

10-19-1999, 10:38 AM
Barbara Bach can be found at Titenhurst Park,Sunningdale, Berks where she lives with Ringo Starr.


10-20-1999, 08:36 AM

"Capital punishment turns the state into a murderer. But imprisonment
turns the state into a gay dungeon-master." Emo Philips

10-20-1999, 01:52 PM
jeez, you don't actually mean to leave a critical thing like that to a batching program, and what about the processing? How can you tell a program like barbabatch is actually going to do a good job of it, and not spoil the sound for you? I would never entrust material to something like that. Mastering is a delicate process in which you need to take great care about what you're doing. Unfortunately, protools is not the greatest program around for mastering, but I, too, am forced to use it. However, you can use Peak with added advantages:
(this can save you some time)
an example process:
1) where does the original come from?
-ideally it's already a digital file, hopefully 24bits (has never been 16...)-
import the original material the least destructive way possible
(if analogue, use good quality converters and sample it 24bits/44.1khz (-always ensure you are sampling the material with optimum levels and a good limiter-))
(if digital, import as is, and leave protools crunching it overnight (the tweakhead setting)
to 24bits
2) use Peak - set it up to work with TDM and DAE - In Peak you can create playlists and do faster than realtime processing, and it does it good
3) save the individual song files thru a UV22 process as AIFF or SDII (if two tracks are crossfading into eachother, do the trick in peak and save as one file, put the code there later)
4) import the files to masterlist

Otherwise, putting all the tracks in as a session in protools and processing there is not too hard- lay them out after eachother (maybe 2 by 2 on 4 tracks), create a masterfader, insert the good old L1, adjust levels using level automation. Like, not too hard, right.. After all, it's just a question about being careful with what you do with the material, not overprocessing if not needed, and so on..


10-20-1999, 01:53 PM
oh, and by the way, you don't actually use toast for production? (masterlist+adaptec jam is a good solution)


10-21-1999, 12:22 AM
um, yes I do use toast at the moment because it's all I have and the price was right...I'd like to beef up that part of my setup.

I don't know what Peak is or the tweakhead setting for bit-length conversion. But it sounds like what I've been envisioning. I don't expect the process to be automated but any tools out there that make it easier I welcome. In the last example you give, are you suggesting to leave the Master fader at unity and adjust the levels on the L1 for gain adjustment?

Thanks for your input.

10-21-1999, 03:44 PM
yes, I suggest leave the master at unity, I tend to set the L1 up at - 0.1 or -0.2dB and then take the threshold down til I get a decent RMS. You'd probably need to check the release time for distortion (set it a bit slower). Peak is "Bias Peak", currently at v 2.02 or something. Should be www.bias.com.
It's an "conversion/edit/sampler-support" program that has grown to be a little bit of everything in the digital sound domain.


10-22-1999, 11:03 AM
Thanks Benjamin. the next session scheduled will wrap this mastering project. Currently, the client is happy with the overall sound. I used your suggestion (fader levels/L1) in getting strong levels (comparable with the average CD)for each song individually. I haven't bounced each song to disk yet. In the next session I want to get an overall average level for each tune. As I think about the process I might use, it seems convoluted. What I want to do is quickly be able to a/b different songs/sections to check for consistancy. My original intention was to bounce each song to disk with current settings, import all tunes into a contiguous 2-track session, set memory location points, and compare the levels (using my ears) for average, consistant level. It seems to me that I can't make individual level changes for each song within this session because if I bounced-to-disk (after using L1 w/level automation)from this session I'd be creating a contiguous file when I would need individual tracks instead for creating ID #'s in Toast. Would you recommend doing this and then extracting the individual, finished tunes from this file or using the contiguous file for reference only and then go back to the invidual files, make gain change and then bounce-to-disk? hmmm.

Tim Walters
10-23-1999, 12:05 AM
Here's what I do. It may or may not work for anyone else.

1) Mix each song by bouncing to 24-bit split stereo *without any compression or limiting on the stereo mix*. If you or the client need a "hot" reference CD, you can also make a compressed/limited 16-bit copy, but don't use this for the final mastering session. You want to stay 24-bit as long as possible, and you don't want to compress or limit until you hear the song in the context of the whole album.

2) When all the songs are ready, import the 24-bit files into a new session. On the master fader, put your mastering processing chain: EQ, multiband or single-band compressor, L1 or Maxim last. Set dithering on the L1 or Maxim to 16 bits. Make all the relevant controls automatible. (You can create a template session that remembers the last two steps automatically.) Now adjust the EQ and/or compression for each song, and write it to all automation parameters. Then adjust the threshold of each song in L1 or Maxim to serve as the overall level control for the song. There may be some back and forth here as you decide it would be better to adjust compression rather than level or vice versa. Jump around between songs to compare the sound, and check the transitions between songs as well.

3) Once everything sounds exactly right, bounce each song to 16-bit. You now have all the songs ready to go, and you just import them into MLCD or Toast, adjust the spacing, and burn away. *Make sure you don't do any volume changes in MLCD or Toast.* If you change your mind about a level, go back and re-bounce that song.

10-23-1999, 12:17 AM
Here's what I do. It may or may not work for anyone else.

Tim, I do it almost exactly the same way... The only difference with me is that I do apply a bit of limiting on the stereo buss during the mix pass, but only very lightly. Musical preference - I like to mix through a limiter.


10-23-1999, 02:16 AM
Yes, I think Tims recomendations seem fine. I must say I am still in the thinking process about mastering in protools, I'd like to find a good way to do it as efficiently as in Sonic Solutions, but I still haven't gotten there. Soon, though..

Well, actually, I do have one thought, is there a way to play back stereo files thru stereo inputs in protools, so that I can stick stereo plugs straight into a stereobus for each track, without having to assign two individual tracks to aux busses? -That would at least help a bit.. That would leave me with 10 or so stereopairs with individual plugs and settings and no master bus.

Other than that, I think I need to get into plug-in automation. I guess it's about time I did..


10-23-1999, 07:23 PM
"jeez, you don't actually mean to leave a critical thing like that to a batching program, and what about the processing? How
can you tell a program like barbabatch is actually going to do a good job of it, and not spoil the sound for you?"

I'm only responding to bluemt's request for a way to do batch bounces, not batch mastering. Barbabatch will take a session and bounce all of your songs to disc for you so that you can do something else productive with your time. I saw it in use at other studios and can't actually vouch for the sound quality until I test it for myself, as always we all gotta trust our own ears or we got no business being in this business.

10-23-1999, 10:04 PM
<<is there a way to play back stereo files thru stereo inputs in protools, so that I can stick stereo plugs straight into a stereobus for each track, without having to assign two individual tracks to aux busses?>>

Stereo tracks is a request that has been going around for a while. Digidesign still has not implemented this and version 5 will not have this feature either. So yes, you do need to bus to an aux channel to apply stereo plug-ins.