Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: ReplayGain - FLAC native method vs. Foobar2000 method. (Read 4416 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

ReplayGain - FLAC native method vs. Foobar2000 method.

My question is, while foobar2000 changed its method of calculating RG to the EBU R128 thing couple of years ago, did the FLAC native RG scanning adopt the same reference or is it still based on the old method?

Is there any real advantage with R128 against FLAC native RG scanning?

Reason I ask is because I am going to test some Linux tools to see if EAC and foobar2000 are fully replaceable with the combination of the following applications: foobnix (as the player), whipper (as the cd ripper w/ AR support) and Perl Audio Converter to substitute the foobar2000 converter. At this point, I know these Windows only applications can run under WINE, but I'm not really interested in this.
This is why I ask, is R128 that important or of any concern... my whole library has a RG under this reference.

Thank you for your answers.

Re: ReplayGain - FLAC native method vs. Foobar2000 method.

Reply #1
Well, i don't what method flac reference encoder use in their RG scanner, r128 is new algorithm compare to old RG, many HA users already moved on to r128. I'll try to contact xiph to find out what method they used...
qaac -cvbr 0 -he

Re: ReplayGain - FLAC native method vs. Foobar2000 method.

Reply #2
No need to ask.  FLAC uses essentially the original replaygain (2.0) algorithm, tweaked slightly for practicality and speed.  It won't give the same answers as an R128 scanner, even hacked to approximately match the 89 dB loudness.

EBU R128 isn't a magical new concept, despite what you might read in gushing blogs all over the web,, but it is generally an improvement on the original replaygain specification.  Perhaps more importantly, it is becoming the standard and is commonly used to generate REPLAYGAIN tags that we're all used to.  In that context, it is usually normalised to -18 LUFS rather than -23 LUFS, so that it gives very similar results to the original replaygain method.

Re: ReplayGain - FLAC native method vs. Foobar2000 method.

Reply #3
I don't know which is "better", but I'd guess it depends on the program material and the listener...   Is R128 better 51% of the time?  Is the original RG better 90% of the time?   I don't know.... but the important thing is that R128 is a recognized standard.   (Well, it's not that important to me, but it's important to certain companies & organizations.)

Re: ReplayGain - FLAC native method vs. Foobar2000 method.

Reply #4
Well, i don't what method flac reference encoder use in their RG scanner, r128 is new algorithm compare to old RG, many HA users already moved on to r128. I'll try to contact xiph to find out what method they used...

Thank you for your effort. Perhaps they plan to move FLAC to R128 on the next release without breaking the binary?

Re: ReplayGain - FLAC native method vs. Foobar2000 method.

Reply #5
I don't know which is "better", but I'd guess it depends on the program material and the listener...  Is R128 better 51% of the time?  Is the original RG better 90% of the time?  I don't know.... but the important thing is that R128 is a recognized standard.  (Well, it's not that important to me, but it's important to certain companies & organizations.)

Don't take my word for it. From the wiki, with refs: "BS.1770 has been shown in several academic studies to be equally or more effective than the RG1 algorithm in modelling human loudness perception on music program".  The most obvious difference in R128 is gating, although replaygain had some complex processing aimed at achieving the same thing. In practice, there is usually little difference.  Individual perceptions of the results may vary, with the aim being very much to match the human perception of the loudness rather than any absolute measure. In isolated cases there are differences of several dB in results for the same track, and apparently it tends to be caused by different frequency weightings (R128 has a high-pass and fairly flat high frequency response) rather than the gating.

It is worth noting that the replaygain 2.0 specification specifies the BS.1770-3 standard rather than the original algorithms, hence it is essentially EBU R128 with a different reference level. I haven't seen any intentions to update Flac any time soon.