While I was randomly playing tracks from my library last night, I noticed that one album (the original LP is filed way away, the tape from that is long gone, I need to work from the digital copies I have) has weak volume. I need to boost the volume (normalize?) overall.
Now, I know Audacity can do that. Question being, what’s the way to do that with the least additional distortion or degradation? I can work from MP3, FLAC, or WAV, and I’m guessing FLAC would be the best source to work from? Adjust it manually, or…I just upgraded to v.2.3 so I have no idea if there are any new better options for me to use in there.
Or should I just turn up the volume (A LOT) on the player, instead of adding distortion by changing the file?
The “digital maximum” is 0dBFS (zero decibels full scale). If you try to go over 0dB you’ll get [u]clipping[/u] (distorted flat-topped waves). On the default Audacity waveform view the maximum is shown as 1.0.
Nothing bad happens if you get close to 0dB, only if you “try” to go over!
There are two ways to normalize (AKA “maximize”) with Audacity. If you run the Amplify effect, Audacity has already scanned your file and it will default to whatever change is required for normalized 0dB peaks. There is also a Normalize effect with a couple of additional features/options.
I can work from MP3, FLAC, or WAV, and I’m guessing FLAC would be the best source to work from?
Use WAV or FLAC. WAV & FLAC are lossless.
As you may know MP3 is lossy compression. When you open an MP3 in Audacity (or any “regular” audio editor) it gets decompressed. If you re-export to MP3 you are going through another generation of lossy compression and the "damage’ does accumulate. You may not hear any quality loss, but it’s something to be aware of. If you want MP3s, it’s best to compress ONCE as the last step after any editing/processing.
Note that perceived loudness does not correlate well with peaks so if you normalize all of your music all of your files won’t be equally-loud. And, there are a couple of reasons that a digitized LP (or tape) won’t’ be as loud as a digital original -
The process of cutting & playing-back an LP changes the peak-to-average ratio (AKA crest factor) without changing the sound of the dynamics. This often leads people to falsely-conclude that the record is more dynamic (less dynamically compressed) than the CD made from the same master.
And, most older records ARE more dynamic because they were made before the [u]Loudness War[/u] kicked-into high-gear. (From what I’ve read, in the vinyl days Motown Records “won” the loudness war with a process called “loud and clear”.)
Is that “loud and clear” process used in the newly produced vinyl too btw?
No. That time has passed. And, I suppose it was more of a philosophy or goal than a particular process. They had some custom analog equalizers & compressors that don’t exist anymore. I actually read something else about someone buying a piece of equipment… Maybe at a garage sale? When they figured-out what it was, it was an equalizer (used at Motown) with a switch so they could equalize while listening at normal speed and then switch to apply the “same” EQ with the EQ-frequencies cut in half for cutting the vinyl master at half-speed.
I also read Jeff Emerick’s book (the Engineer for The Beatles). At EMI Records, they were jealous of the bass Motown was getting on their records.
Of course, it’s all done digitally now. You can do things digitally that were not possible with analog. Usually the vinyl is cut from the same master that’s used for the CD & MP3. Of course, it sounds different, the wave shape changes (which can “hide” clipping if the master has been “pushed” that far) and with the increased crest factor it’s hard to tell if it’s the same master. Or sometimes there is additional processing or different mastering/processing. That information is no usually published so it’s not easy to find-out what was done on a particular recording/production.
Thanks, Doug.
So it looks like I can just run “normalize” on the FLAC files, and if the original problem is just that some underpaid audio engineer made the original “too softly” this will either bring it closer to norms or at least, as close as it is going to get without getting complicated or distorted. And then I can remake new MP3’s from those FLACs.
I haven’t taken a look at the problem files yet, I had just heard them and noted which albums they were while listening to the FLACs the other night. I know that for those albums which I’d taped, The Nak582 was carefully adjusted for 0db peaks. These days…
I swear the albums that Hollywood puts CD release, are produced by chimpanzees. The track cuts are simply mangled. Lead-in’s and comments after a song, often are cut off and attached to the adjacent song–where they are totally meaningless.
One more thing that I know I can fix in Audacity!
I know that for those albums which I’d taped, The Nak582 was carefully adjusted for 0db peaks.
That’s the analog level on the tape. There is no fixed-calibration between the analog-tape level and the digital level. Different tape machines have different voltage-outputs and different analog-to-digital converters have different sensitivities.
In a pro studio, everything is probably calibrated, but not for 0dB analog = 0dB digital… They would leave some headroom on the digital-side because tape can go over 0dB. Pros typically record at -12 to -18dB (at 24-bit/96kHz) and the levels are adjusted/boosted during mixing/editing/mastering.
With tape you wanted a “hot” signal to overcome tape noise and you could get-away with going occasionally into the red because tape soft-clips as it begins to saturate and the tape equalization further “smooths” the distortion. With digital there is no tape noise, but it hard-clips at 0dB. Digital recording levels are not critical as long as you avoid clipping. (There is quantization noise, but it’s more than 90dB down at 16-bits.)
I swear the albums that Hollywood puts CD release, are produced by chimpanzees. The track cuts are simply mangled. Lead-in’s and comments after a song, often are cut off and attached to the adjacent song–where they are totally meaningless.
One more thing that I know I can fix in Audacity!
Unfortunately, you can’t fix the “loudness war” compression & limiting. You can get expander and “unlimiter” plug-ins, but with all of the peaks the same and constant loudness throughout the recording it’s impossible to know the “original” peaks or original dynamics. Plus, they use compression (and limiting) with unknown parameters, they may use multi-band compression, and they usually use compression limiting on the individual tracks before mixing.
I realize that. When I mentioned the level recording LP’s on the Nak, I wasn’t mentioning the digital side. But on that, I also carefully adjusted gain so that the Audacity files wouldn’t clip, but would peak just below 0db.
That doesn’t affect “the gain wars” and the production issues that were made by Hollywood, quite true. I’m just saying that all of the easy pilot errors that could have been made up front, at my end, were anticipated and accounted for.
I’ll have to open up the one album that I noticed was really faint (should have written down the names but it was late) to see if it does indeed still have 0db peaks and was just studio recorded ‘low’, or if for some reason I bodged the process along the way. In which case, normalizing will be about all I can do for it. If it needs compression to keep a few peaks down while pushing the rest up, we’ll see.