volume normalization advice

[Partha]

Hi, I am creating an apple lossless 16 bit 44.1 khz library. all my hi-res purchased audio and ripped CDs (ripped with dbpoweramp into AIFF, no processing at all) will be converted into this lossless ALAC library.

I would like to use the default peak-to-peak volume normalization since it SEEMS TO ME to get rid of some pesky loudness in certain mid-90s CDs and hi-res files (like the R.E.M ones on hdtracks) without reducing the volume too much (I listen sometimes on an ipad to headphones).

I tested the replay gain volume normalization mode but it SEEMS TO ME to reduce the volume too much on the ipad.

However, i am adding replay gain and itunes album normalization tags separately for compatible amplifiers and players.

I have a backup of all the original files.

These are the steps I am following while converting an album:

1. Bit depth to 32 bit floating point.

2. Resample to 44.1 khz

3. ReplayGain (Album Gain & itunes Album Normalization).

4. Volume Normalize (default, peak to peak, maximum amplification x8 default)

5. Bit depth to 16-bit with triangular dither.

The files sound okay so far. Please tell me if I am doing this correctly or if i can improve on the settings.

[Spoon]

Re: volume normalization advice

You would have to put replaygain after volume normalize as it changes the volume after the replaygain values are calculated.

[Partha]

Re: volume normalization advice

Many thanks. I will start again.

[Partha]

Re: volume normalization advice

Does peak to peak reduce clipping? Will it reduce the volume of loudly mastered albums? Should I opt for adaptive instead? Please guide me. The goal is to reduce clipping and normalise volume in loud masters, like prince's 'chaos and disorder' or rolling stones' 'goats head soup' remaster.

[Spoon]

Re: volume normalization advice

If a track is mastered with clipping, nothing can reduce that, the track clips. Peak to Peak would never clip a signal which does not already clip.

[Partha]

Re: volume normalization advice

If a track is mastered with clipping, nothing can reduce that, the track clips. Peak to Peak would never clip a signal which does not already clip.

Okay, thanks. Are you okay with me sticking to the peak-to-peak default then? Or do yo want me to switch to another sort of volume normalization in the above steps?

[Spoon]

Re: volume normalization advice

It is your choice, Peak to Peak is ok, I personally use Replaygain applied volume normalization.

[Partha]

Re: volume normalization advice

It is your choice, Peak to Peak is ok, I personally use Replaygain applied volume normalization.

I have added replay gain, and also dynamic range compression (default) in the whole chain without properly understanding it. So now the chain reads...

1. Bit depth to 32 bit floating point.

2. dynamic range compression (default)

3. Resample to 44.1 khz

4. Volume Normalize (replay gain (track gain), default)

5. Replay gain (album gain and itunes album normalization)

5. Bit depth to 16-bit with triangular dither.

Is dynamic range compression in the right place and necessary?

[Spoon]

Re: volume normalization advice

You do not need dynamic range compression unless listen to music in very loud environments.

[Partha]

Re: volume normalization advice

Thanks. have removed it.

[Partha]

Re: volume normalization advice

NOt starting s separate thread for this. If I have to volume normalise (rpg) a bunch of mp3s from a hi-res source, do I still have to start with 32-bit floating point in the chain? for example, for a 24/192 file to lame V2, do I have to do this:

1. bit depth 32-inch floating point
2. resample 48 khz (is 44.1 better?)
3. volume normalise (replay gain)
4. bit depth 32-bit floating point (no dither)

what do you think?

[Spoon]

Re: volume normalization advice

1. bit depth 32-bit floating point
2. resample 44.1
3. volume normalise (replay gain)

Just those.

[Partha]

Re: volume normalization advice

1. bit depth 32-bit floating point
2. resample 44.1
3. volume normalise (replay gain)

Just those.

Thanks. Ok.

[ForSerious]

Re: volume normalization advice

So I've just been using the volume normalize DSP set to the EBU R128 setting. It has the problem of making songs with less full frequency usage sound much louder. (Like a song with just a guitar played after a standard rock song will sound much louder.)
Would the ReplayGain Apply DSP do the same thing and therefore have the same problem?

[timster67]

Re: volume normalization advice

So I've just been using the volume normalize DSP set to the EBU R128 setting. It has the problem of making songs with less full frequency usage sound much louder. (Like a song with just a guitar played after a standard rock song will sound much louder.)
Would the ReplayGain Apply DSP do the same thing and therefore have the same problem?

If you use Album gain rather than track gain, the same gain will be applied to all tracks in the album - so if a track is intended to be quieter, then it will remain so. I presume this is happening because you are using track gain? Or the track is a single?