EBU R128 "For Dummies" Questions

[Dexter]

I've been reading posts on various websites that have conflicting information about using the EBU R128 option in dBpoweramp. I think this might be because it was only available under "ReplayGain (Apply)" at first. I just need some confirmation that my understanding of this feature is correct, so I make the right choices going forward. I rip to FLAC.

1) If I chose the "ReplayGain" DSP option, (not the "ReplayGain-Apply" option) ..... regardless of whether I choose the new "EBU R128" setting or the older "ReplayGain" setting for gain calculation, the ripped audio files themselves are NOT modified or affected in any way. The calculated gain values are written into the tags - and my audio file is a pure lossless rip.

2) EBU R128 was adopted as the loudness calculation standard for television & radio - and is now expected to become the standard gain calculation for ripping audio files as well, eventually (possibly?) replacing the older "ReplayGain" calculation - because EBU R128 is considered better/more accurate. (And is this true for the US, as well as Europe & the rest of the world?)

3) The recommended LUFS setting for EBU R128 is -23. The reason why dBpoweramp is set at a default of -18 is so that it will better match the files that a user might have already ripped using the "ReplayGain" calculation method prior to the advent of EBU R128.

4) If one decides to switch from the ReplayGain calculation to EBU R128, then -18 is the preferred LUFS setting.

5) If one is planning to begin ripping their CD library ..... or plans to re-rip their all of their CDs ..... what LUFS setting should they use? -23?

6) Clip-Prevention does not affect the audio files either. It merely adjusts the ReplayGain or EBU R128 calculated values written into the tag - on music where the normal values would increase the volume so much that it would cause clipping.

Is my understanding of these features correct?

Thanks!

[garym]

Re: EBU R128 "For Dummies" Questions

1. correct
2. Yes
3. Yes
4. Yes
5. either is fine. I've read that the -23 was actually excessive for the home user and somehow relates more to the broadcast use of audio. But I'm no expert on this aspect of things.
6. Correct

[Dexter]

Re: EBU R128 "For Dummies" Questions

Thank you!!

Now I just need to determine which LUFS setting to use. I don't have a large percentage of my CDs ripped yet and really need to ramp that effort up. They are ripped in a variety of formats: Lame MP3 (at V2, V1 & V0 - before I discovered FLAC) ... and FLAC at various settings. So I don't have a problem re-ripping as I get to each artist.
.

I'm tempted to stick with -18. Opinions welcome!

[garym]

Re: EBU R128 "For Dummies" Questions

Thank you!!

Now I just need to determine which LUFS setting to use. I don't have a large percentage of my CDs ripped yet and really need to ramp that effort up. They are ripped in a variety of formats: Lame MP3 (at V2, V1 & V0 - before I discovered FLAC) ... and FLAC at various settings. So I don't have a problem re-ripping as I get to each artist.
.

I'm tempted to stick with -18. Opinions welcome!

I can tell you that I use -18. Even at that level (matching RG), many users in the past felt like it reduced volume too low. If one uses -23, it would be even lower. In some ways it really doesn't matter as long as everything in your library is the same. One thing that moved me to -18 is that foobar2000's implementation of EBU R128 in their RG application uses -18 and it is not configurable (and I sometimes use foobar2000 to "redo" my RG tags. I think even dbpa now comes with the setting at -18 by default (but I could be wrong about that).

[mville]

Re: EBU R128 "For Dummies" Questions

Now I just need to determine which LUFS setting to use. I don't have a large percentage of my CDs ripped yet and really need to ramp that effort up. They are ripped in a variety of formats: Lame MP3 (at V2, V1 & V0 - before I discovered FLAC) ... and FLAC at various settings. So I don't have a problem re-ripping as I get to each artist.
.

I'm tempted to stick with -18. Opinions welcome!

I am also using -18 LUFS, which IS the current ReplayGain default.

[schmidj]

Re: EBU R128 "For Dummies" Questions

If you want to educate yourself about LUFS settings, you might find it interesting to Google "EBU R128 for music". There is a lot of interesting discussion going on. -12, -14, -16, -16.5, -18. I am sure this subject will come up at the AES meeting here in New York in a couple of weeks.

Frankly, a 2dB change is close to inaudible, although A/B/X testing of two identical pieces of music with "A" 2 dB louder than "B" is likely to give "A" the preference - hence the loudness wars.

One major issue with setting it too low, like the -23/24 setting used in TV is that portable devices limit the available gain feeding the headphone jack to avoid hearing damage. (I think this is also an EU requirement, not sure.) Many players seem to do this by just not providing extra gain, or by warning you if you try turning the headphone gain up beyond some point, rather than actually monitoring the material being played for loudness or peak level. If you use a level that is low, like -23, on your audio files, you may well find that your portable device won't let you play it loud enough to be listenable.

I've never been a fan of the loudness wars that afflicted the pop music world and commercial radio. Many years ago I was the chief engineer of a commercial radio station. The program director dictated the settings of the processing. I was also the chief of a non-commercial college station, where I had the key to the processor lock. We weren't quite as loud, but much more listenable.

Back in the 1970's a network-owned FM radio station in New York had a tremendously excessive amount of limiting, it "pumped" horridly. Somewhere I have an off-air tape of them playing a then popular ragtime hit, I think it was "The Sting" which had a very long "stop" in the middle of the tune. The processor would keep increasing the gain during the stop until the playback hiss (probably from a tape cartridge) was pretty close to 100 percent modulation before the stop ended. And they used this ugly processing for some years...

Anyway, I use -18 both for my rips and for the level setting of the location recordings I produce. Seems to be a good conservative compromise.

One final reason to use -18 as opposed to -14 or so with dBPoweramp, is that a surprising number of recordings did and do have quite a large dynamic range. dBPoweramp will not increase the loudness above the point where it would put the loudest parts of the track into clipping, and it does not include a limiter or other "active" processing in its replay gain plug in. If you use too high (low number) a setting. you'll still find inconsistent playback levels, as the heavily compressed stuff will play back "loud" at the setting you chose, but the wide-dynamic range stuff which would have clipped will play back softly as dBPoweramp was unable to set the loudness to the setting you chose.

When producing my location recordings, I use editing software that includes a (pretty transparent) limiter to set the loudness to -18 LKFS and limit any peaks to -1 dBFS. I then listen to the results to make sure the limiting is inaudible (no pumping, etc).

[Dexter]

Re: EBU R128 "For Dummies" Questions

Strong arguments for using the -18 value - and that's what I will do. (I did take schmidj's Google suggestion and a lot of it was way over my head - but I got enough to convince me to go with -18.) If you hear or read about anything coming from the AES, post it - it might be interesting.