GAIN vs Amplify vs Compressors (Vinyl vs DVD)

Audacity 3 Help Forum Windows
1

[Background: I’m on a kick to make digital copies of my entire vinyl collection, about 2,700 albums. I’ve already done about 700 CD’s. I’ve almost exclusively used Audacity over the past 3-years and am still working on it. I get as many 12 vinyl albums done per week, so it’s going to take a while :slight_smile:

I keep the output from the process as VBR MP3, depending on the track, is circa 250kbps. I’m doing this so I can play my entire collection on random play by genre, year, artists etc. and create playlists of favorites. The MP3 files are stored on a NAS, backed up by another NAS, which is also backed up online. The music is streamed via a separate Logitech Media Server(LMS).]

Problem: I use a logitech media server to stream music to every room in my house. The vinyl albums/tracks I’ve recorded don’t play as loud as the CD tracks I’ve ripped. I’ve read about the recording levels in Audacity and pretty much follow the recommendations for levels. When I play back random tracks the CD tracks are always just a bit louder, no matter what type of music.

There seems to be a couple of solutions to this.

  1. Use audacity AMPLIFY effect at the end of the editing to boost an album to as “loud” as possible. Easy enough on clean albums, but more time consuming on “noisy” albums. However, this doesn’t seem to still get the “loudness”.

  2. Start using compression - all sorts of reasons I don’t want to. Not least it adds another layer of processing and issues.

  3. Use the Squeezebox/LMS gain control to try to achieve a balanced loudness. It’s described here. http://wiki.slimdevices.com/index.php/Automatic_Volume_Adjustment

**Question:**Does Audacity set the gain volume adjustment metadata tags from your audio files?

Is there a good article that discusses the general issue of vinyl loudness vs CD loudness, or how best to achieve a balance?

Any recommendations?

++Mark.
website: https://ctproduced.com
discogs collection: https://www.discogs.com/user/4MC/collection

2

about 2,700 albums.

Holy smokes! That’s a lot, especially if you are “cleaning up” the recordings.


I’ve already done about 700 CD’s. I’ve almost exclusively used Audacity over the past 3-years and am still working on it.

I Hope you didn’t use Audacity to make digital-to-analog-to-digital recording… Normally CDs are digitally “ripped” with a CD ripping application. Audacity doesn’t do that.) Ripping is faster and if all goes well, you can get a bit-perfect digital copy and the metadata gets filled-in automatically

I keep the output from the process as VBR MP3, depending on the track, is circa 250kbps.

You might consider keeping a lossless FLAC archive, since you’re going to all that work.

  1. Use the Squeezebox/LMS gain control to try to achieve a balanced loudness. It’s described here. > http://wiki.slimdevices.com/index.php/A > … Adjustment

That should work reasonably well. The files do have to be pre-scanned to find the “loudness”. From quickly reading the link, I’m not sure if SqueezeCenter will write the tags, or if that has to be done by another application in advance.

The ReplayGain algorithms make a simple-single harmless volume adjustment for each song (before the song starts). They don’t use dynamic compression or automatic volume control or anything that will “damage” the sound.

There are variations such as MP3Gain and WaveGain that make “permanent” adjustments to the files so they are compatible with everything. MP3Gain has a limitation that MP3 can only be adjusted losslessly in 1.5dB steps so if you’re going to make “permanent” changes it’s better to use WaveGain on the WAV file before making an MP3.

There is also an unofficial “ReplayGain” plug-in for Audacity that can make “permanent” volume adjustments (it doesn’t write ReplayGain tags). (I think I heard a rumor that volume matching be officially added to Audacity soon.)

These algorithms are much better than simple normalization, but two different people may not agree when two songs sound “equally loud”, and you’ll run into some quiet-sounding songs with a few high-peaks that can’t be adjusted-up enough without clipping.

A couple of things are going-on with vinyl. Most of your albums are (probably) older and made before the Loudness War really kicked-into high-gear (and when they only had limited analog compression tools). They were using compression and from what I’ve read, Motown “won” the loudness war back in the vinyl days with a process they called “Loud and Clear”.

And even if the CD is made from the same master, the vinyl cutting & playback process introduces phase-shifts that end-up making some peaks higher and some peaks lower. These short-term peaks don’t affect the sound of the dynamics, but it means that a normalized (“maximized”) digitized vinyl copy will have a lower average volume the digital original, even if they both have 0dB peaks. (People who do “simplistic” dynamics measurements will say the vinyl is less compressed.) This effect is more pronounced with more-compressed masters.

  1. Start using compression - all sorts of reasons I don’t want to. Not least it adds another layer of processing and issues.

The “extra layer of processing” usually isn’t an issue, although you can get side-effects if you over-do it or use the wrong settings. The dynamic compression itself can be undesirable (if you enjoy dynamic music with quiet parts and loud parts).

3

Thanks for the thoughtful answer.

No, I ripped all the CD’s straight to digital :slight_smile:

For the vinyl albums, I record using Audacity, using a PS Audio Phone Nuwave Convertor to do the analog to digital, it connects to a computer using USB. Audacity reads the USB using WASPI under Windows 10. Once both sides of the album are recorded I post process them with an Audacity script/macro, the first step of which saves the entire recording as a WAV file, then goes on to do offset/normalize/click removal etc. Pretty much exactly what I found in an audacity how-to 3-years ago.

Once thats all finished, I then set about equalizing the audio and setting up to export as MP3. This step can be quick or really tiresome.

I’ve got a lot better at getting the volume right, and have asked the same question over on the LMS forum to see if I can come up with a better solution than I have now.

I also remembered that I can increase the gain from the Nuwave Converter to the PC, it’s currently set at 48db, 7.0mv but from memory that produced way too much clipping on some recordings. I’ll try that next time.

4

the first step of which saves the entire recording as a WAV file, then goes on to do offset/normalize/click removal etc. Pretty much exactly what I found in an audacity how-to 3-years ago.

[u]Wave Corrector[/u] is an automatic de-clicker and it’s now FREE. You might want to give it a try. I have very limited experience with it. (I’ve used Wave Repair for a long time, but it’s manual and VERY time consuming. On the plus side, it only ‘touches’ the audio where you identify a defect.)

Normalization (including optional offset removal) is harmless, but you can get artifacts (side effects) with click removal so it’s usually a good idea to listen carefully and/or save an unmolested backup/archive.

I also remembered that I can increase the gain from the Nuwave Converter to the PC, it’s currently set at 48db, 7.0mv but from memory that produced way too much clipping on some recordings. I’ll try that next time.

Digital recording levels are not critical at all as long as you avoid clipping so it’s common practice to leave some headroom and shoot-for peaks around -6dB.

Nothing bad happens when you get close to 0dB and you can go quite a bit lower, but analog-to-digital converters (and digital-to-analog converters, and regular WAV files) are hard limited to 0dB and they will clip if you “try” to go over. Pros recording live (with less predictable levels) leave even more headroom.

You can safely normalize or otherwise adjust the levels digitally after recording.

Since you have so many records, you might remember analog tape… With tape you wanted a hot signal to overcome tape noise and it was common to go occasionally “into the red”. But with digital, no tape noise! :wink:

And on the other end, tape starts to saturate and soft-clip if you go over 0dB whereas digital hard-clips at exactly 0dB.

5

Hey Doug, I’ve just been reading through this thread and your replies again and want to just say THANK YOU.

So much of what I do now was informed by your replies.