A couple of odd things about ReplayGain in DBPA CD Ripper? Problems?

[Max Dread]

Hi all

The ripping continues... And I've noticed a couple of things about RG that were not as suspected.

1. Hidden Tracks. One of the last CDs I ripped had a hidden track. So I created a new profile with the Hidden Track Removal DSP in the chain BEFORE the ReplayGain DSP (so RG would be calculated AFTER the removal of the silence). It worked fine. But out of curiosity, I then re-ripped to a new folder, but with the order of the DSPs switched the other way. The RG figures were exactly the same (both Track and Album). Not at all what I expected. Is that because I have misunderstood something?

2. FLAC v Mp3. The second thing.... I had a disc with one track with errors. I ended up ripping that one track in Burst Mode. I therefore had to re-do the RG. For FLAC, this went entirely as expected (looking at what the figures were previously for RG Album). But for the mp3s it changed quite a lot and not how I would have expected it to. Not only that, but the Track RG changed too. Any ideas what's going on here?

Here's what it looked like BEFORE redoing the RG:



And here's how it looked afterward:



Any help form you guys would be most appreciated.

Cheers

Max

[Spoon]

Re: A couple of odd things about ReplayGain in DBPA CD Ripper? Problems?

mp3 is lossy and it can include its own subtle changes, if these small changes worry you (0.01 db) you can calculate RG on the mp3 files after they are written with the Batch Converter and [Replaygain] utility codec (convert the mp3 files to this codec to redo the rg).

The hidden track only removes silence, as silence is not audio it is unlikely to affect the overall rg value, as the non-silent audio is still there in its entirety.

[Max Dread]

Re: A couple of odd things about ReplayGain in DBPA CD Ripper? Problems?

The hidden track only removes silence, as silence is not audio it is unlikely to affect the overall rg value, as the non-silent audio is still there in its entirety.

I think I understand.... How about if instead of silence it was just a long quiet/ambient section? And I did RG to the full track, and then RG to the track without the quiet/ambient section? Would I see different RG values then? Is it just because it is pure silence that nothing changes?

I have not got a thorough understanding of how RG works, so I might be getting confused. Please be gentle!

mp3 is lossy and it can include its own subtle changes, if these small changes worry you (0.01 db) you can calculate RG on the mp3 files after they are written with the Batch Converter and [Replaygain] utility codec (convert the mp3 files to this codec to redo the rg).

The GAIN changes don't worry me (I don't think!). But I am confused as to why the Track Gain changes when nothing else has changed at all. If the mp3s are the exact same files, then why would I get two different Track Gain values? What has changed to make the value change? Nothing that I can think of....

Track Peak values have changed drastically...

Album Gain has changed in a way I would expect, but why has the Album Peak increased by 50%.

Whilst on the subject, I'm a little hazy on what the Peak values actually tell you. If anyone can explain or link me to somewhere that explains it, that would be most helpful. Part of my confusion is that I though the scale was 0 to 1? If that is the case, why am I seeing values of 1.5? (Especially seeing as I applied exactly the same RG... Cannot see why it would have changed at all)?

Many thanks and sorry if I am misunderstanding stuff

Max

[Spoon]

Re: A couple of odd things about ReplayGain in DBPA CD Ripper? Problems?

1.5 is allowed, it is a decoded value which would be over the allowed range (0-1), so the replaygain application uses gain + peak to set the multiplier.

[Max Dread]

Re: A couple of odd things about ReplayGain in DBPA CD Ripper? Problems?

1.5 is allowed, it is a decoded value which would be over the allowed range (0-1), so the replaygain application uses gain + peak to set the multiplier.

I'm not sure I follow that tbh, but the main thing is that it sounds like it is not a problem.... That's good.

I'm still very keen to know how doing the same operation produces different results. IE - "If the mp3s are the exact same files, then why would I get two different Track Gain values? What has changed to make the value change? "

Cheers

[Spoon]

Re: A couple of odd things about ReplayGain in DBPA CD Ripper? Problems?

Replaygain as a DSP effect is calculated prior to the encoder, if the encoder (mp3) changes the signal then this accounts for the differences.