'Loudness' Measurement for LP-rips

2009-12-06 12:37:08

Hi!

Links given in thread Standardized 'Loudness' Measurement for CDs? made me to compare couple 30s samples listed here, one ripped from CD to mp3 CBR320 against two samples recorded from LP (one 24/44.1 and the other 24/96 (both downsampled to 16/44.1 using r8brain free (quality option: very high quality) because the TT DR Offline tool doesn't accept other WAV resolutions)

CD (mp3) Don Henley: Best Of ('95 (IIRC))
16/44.1              DR9

LP Don Henley: I can't Stand Still ('82)
24/44.1-16/44.1  DR11
24/96->16/44.1    DR12

Is this comparison valid?

Has anyone else done this type of comparison?


Juha

'Loudness' Measurement for LP-rips

Reply #1 – 2009-12-07 09:10:13

The answer to your question is extremely specific to the Pleasurize meter.

You need to be absolutely sure that the spectral content from is about the same, but most especially in the sub-bass (0-20hz) region. The TT meter can penalize normalized recordings with excessive content at bass frequencies. The numbers can also change randomly based on varying speed differences etc.

I think the meter itself is only accurate to about +- 3 db, to be completely honest. As you've just discovered, trivial changes to a file can cause the result to swing by 1, and there is easily 2db of potential inaccuracy on top of that. To be fair though, I don't think anything comparable will do do any better than +- 1-2db.

'Loudness' Measurement for LP-rips

Reply #2 – 2009-12-07 10:00:19

For what's (supposed to be) the same audio, ReplayGain, or even simply RMS, should do.

Yes, subsonic rumble could confuse simple RMS, but it would have to be loud enough to be clearly audible to make much difference to ReplayGain.

However, that really is an attempt at measuring "loudness" - whereas I think you're still really talking about dynamic range. I think you've already had the best suggestions for that in the other thread.

Though I think it was only implied, rather than expressly stated, that all such measures are inferior to actually listening to the thing! (Sorry that doesn't help - but it is the truth).

Cheers,
David.

'Loudness' Measurement for LP-rips

Reply #3 – 2009-12-07 17:42:11

Hmm... I had 12th order HP (Butt.) @ 10Hz in that 24/96 sample and ... I also did try, just now, by using the same order/type HP filter but @ 20Hz on that same sample ... quess what ... it shows now DR14.

BTW, ripping was done using software based RIAA EQ.

I noticed also that, when I measured one sample track without RIAA EQ in it, DR tool showed DR20 for it and then loaded exact same file with RIAA EQ added in it and it showed 'only' DR13 (which is ~right value)? Levels were ~equal.

Juha