Mastering Question

[Dalton_Sound]

When studying mastering techniques, and working on projects for class I have been told that it is a good idea to gain a clip to true peak at roughly -.3 in order to maximize the dynamic range used by the recording. While working this last week, I heard from another person that it is a good idea to gain your clip to -3 True Peak. Is there a guidline for this?

I like the idea of -.3dbFS, because it helps me meet my -16LUFS Mastered for iTunes goal without having to use as much compression, but without clipping. (This is coming up from a -23LUFS mix.)

Thanks!

[Bob Olhsson]

The .3 recommendation is to avoid clipping D to A converters, especially older ones with analog volume controls. If you are using lossy compression, -1.0 True Peak will minimize clipping the encoder filter banks. For streaming, -3 is probably a good idea. I'd run tests. The Sonnox codec plug-ins, (https://www.sonnox.com/plugin/codec-toolbox, https://www.sonnox.com/plugin/fraunhofer-pro-codec) Nugen MasterCheck pro (https://nugenaudio.com/mastercheck) or the MAAT (https://www.maat.digital/drm2/) let you check levels and monitor through codecs.

[BScout]

If your goal is to get MFiT qualified, you need to use Apple's MFiT tools and one of them will give you ISP/True Peak readouts for test conversions (it's the command line tool that can output to a text file the final stats for conversion.) These are test conversions as you do not submit converted files to Apple but the converter does oversample during encoding so True Peak clipping can happen even if your starting file is below clipping even by -.5 dBFs

Even if you do not clip, Apple still prefers a minimum of -1 dBFS true peak for delivery of the source file (though they won't reject it if it is less than that as long as it does not clip through the current encoder.) Part of MFiT is the concept that they may change/improve encoders down the line and (re)convert their entire MFiT catalogue. So if you are running precariously close to clipping with today's converter, who know what will happen later.

[Dalton_Sound]

Bob,

I will look into those plugins.
Thanks!

BScout,

That is very helpful information. I was confused by Apples information on their website for MFiT, and it looks like they haven’t updated their page or software for it at all recently. I also noticed that when putting my WAV into the MFiT Droplet, that it removes all of my metadata. Why would that be? Also, is there a better way to check your files than using AU Lab? Is that something that the codec plug-ins Bob mentioned dies?

[BScout]

Apple's tools are the only ones "qualified" but Sonnox Codec Toolbox and Nugen Mastercheck Pro will give you approximately the same sound in an "active means" so you can hear back while working with the audio. The Apple tools (other than AU lab plugin) are offline so you can only use it after making your final master audio. But make no mistake, like everything Apple, the only actual qualified tools are the Apple provided ones.

As for metadata, when you submit MFiT files, you are submitting uncompressed wave files with no metadata. You supply the metadata separately for data entry. Apple then converts the source wave files and then injects the metadata. Also the reason why the MFiT converter discards any metadata (a lot of which isn't actually embedded metadata in a wave files which has a limited metadata tag vs something like mp3 or aac.) You can always enter metadata after creating the MFiT files -- but remember these are only test files. You can do what you want with the aac files you've created by you do not submit converted to AAC files to Apple. You send them uncompressed wave files.

[JFreak]

i usually master to -1 or -2 or even -3 dBFS

[Bob Olhsson]

On a Mac, they all use the Apple code. On PC, they use the Fraunhofer pro encoder set to the same settings.