Subject:matching volume and tonal qualities of diff songs

Posted by: Prateekdujari

Date:1/23/2012 3:46:07 PM

Hello. using batch converter in SF Pro10.0C. I've a bunch of songs from diff sources. When played they sounddifferent due to diff inherent volume levels and spectral content. Need to make them sound alike. So I figured I could use batch converter , and normalize the volumes first and then apply one of the presets of the Mastering-eq. Doing this sequence however still gives me diff volume levels from song to song? Should I first apply the Mastering-eq preset and then normalize the volumes?
Pl note my assumption is that the mastering-eq forces all the songs to the same spectral curve of whatever preset I choose.

Next I just learnt mastering exciter can be used to make the songs sound richer. I have some songs which dont sound rich. so what should the sequence be between normalizing volumes, mastering-eq to match tonal qualities and using the mastering exciter to enrich the sound of the songs?.

Subject:RE: matching volume and tonal qualities of diff songs

Reply by: Chienworks

Date:1/23/2012 4:25:55 PM

One misunderstanding and one misconception.

Normalize finds the loudest peak of the entire song, calculates how far that is below the desired level, and raises the entire song evenly by that amount. This means that songs with loud peaks may not be raised as much as those with less dynamics, and the non-peak areas of the more dynamic songs will sound quieter than the less dynamic sounds. You need to use some sort of dynamic normalization that detects the average loudness of the song. Sound Forge's RMS normalization can approach this, but not very well, and it won't handle cases where you're pushing the peaks beyond the upper limit gracefully. You might want to consider some mild compression instead of or in addition to normalizing.

A Mastering EQ can change the tonal quality of a song, but whatever change it makes to any given song it also makes to all the rest. That means all you've done is shift all the songs by the same amount of tonal change. You're not converging them all toward some middle ground. This is something you'll have to tweak individually for each song rather than batching.

Subject:RE: matching volume and tonal qualities of diff songs

Reply by: ChristoC

Date:1/23/2012 4:35:22 PM

What you are trying to do is all about human perception and hearing; there's no way that can be done with a simple batch converter or automated process. e.g. applying a random preset 'Mastering EQ' without any knowledge of the actual material it's being applied to can only end in sadness... the EQ has no 'intelligence' as such. Similarly the perception of 'loudness' has very little to do with measuring the peak or rms levels in the song, but everything to do with how the ear works.

Why do you want all your songs to sound alike? - isn't the joy of music in the diversity of sound?
Would you purchase many paintings then alter them to all look the same?

Professional Mastering Engineers can make songs sound rich, or even just very expensive..... which is why they earn their money!

Subject:RE: matching volume and tonal qualities of diff songs

Reply by: musicvid10

Date:1/23/2012 6:44:21 PM

I won't touch the question of making them "sound alike" WRT to spectral content and tonal emphasis; you're on your own there, but:

For normalizing loudness (not "volume") the new technology is very, very good. Download the trial of Nugen VisLM, read up on the basics of ITU-R BS1770/1 so you'll have an idea what you're doing, and normalize your songs to the same LUFS/LKFS. You'll be amazed how well it works.

Subject:RE: matching volume and tonal qualities of diff songs

Reply by: Prateekdujari

Date:1/24/2012 2:12:32 PM

I've my reasons to have all the songs sound alike, and pretty good ones for me. Maybe there is someone out there who likes a copy of the same painting in every room. by the same token I wear the same shoes day in and day out even though my socks change all the time.
to answer your question directly, even though it , I feel is irrelevant, I dont want to be running back and forth betwene my receiver and wherever I am in the room to adjust the volume, base and treble each time a new song starts playing in my playlist. Good enough reason?

I think my solution is to buy a eq spectral matching plug-in that apparently SF is missing. This plug in I've in mind works with SF, is called Repli-Q by Bias and claims to forece the spectral curve of a whole bunch of songs to all match the spectral curve of a reference song file I have. To me it seems doing that wil match the mids, bass and treble throughout my playlist. After that I can use some rms method to match the volumes to some degree at least. comments on this plan?

Subject:RE: matching volume and tonal qualities of diff songs

Reply by: Geoff_Wood

Date:1/24/2012 5:58:53 PM

What you are wanting to do is 'mastering'. I like to do it in CDA, individually per event (track). Is possible to preview without too heavy a plugin load, on a new fast cpu.

Makes it very easy to skip around the whole ablum to make it sound coherent, if that's what you want.

Now only if CDA supported VST directly, and paged in/out active plugins....

geoff