Replay Gain - dbPoweramp vs. Foobar

[Guinness]

Why are the replay gain values different in dbPoweramp and foobar? I ripped a cd using dbPoweramp cd ripper and used the DSP effect, replay gain. I also ripped the cd without the DSP replay gain and then used foobar to add the replay gain tag values. Below are my results. I'm curious has to why the values are different.

dbPoweramp - replay gain

https://docs.google.com/file/d/0B4tJ56bWRc1OOEZrUnZnZ1NOQVk/edit?usp=sharing

Foobar - replay gain

https://docs.google.com/file/d/0B4tJ56bWRc1OYVBEa1dBc2NZdmM/edit?usp=sharing

Sorry for the links to the images but I can't seem to figure out how to post the image.

Thank you!

[dbfan]

Re: Replay Gain - dbPoweramp vs. Foobar

Perhaps foobar switched to the ebu r128 calculation? You should use is method as a way forward in dbpoweramp as the results are more accurate.

[Guinness]

Re: Replay Gain - dbPoweramp vs. Foobar

********* - thank you for the reply.

I just want to make sure I understand your reply. Are you saying to use dbpoweramp's RG as they are more accurate than the way foobar calculates RG?

One more question. This is how I'm applying replay gain. First I'll rip the cd to flac using dbpoweramp without the replay gain DSP. Then I'll use dbpoweramp batch converter and convert flac to flac and apply the replay gain that way. Does this method work. For example, if I apply replay gain while ripping the cd and some tracks have errors (inaccurate rips) then I want to go back and re-rip to see if I can get an accurate result. If I get an accurate result then my replay gain album values will be different on those re-ripped tracks. Therefore, I would like to rip all tracks first and once I have all accurate rips I can then apply replay gain using batch converter.

[Spoon]

Re: Replay Gain - dbPoweramp vs. Foobar

I meant that EBU R128 is more accurate than the older replaygain.

I would use RG whilst ripping and for those discs where you re-rip, only then convert FLAC >> FLAC to recalculate.

[Guinness]

Re: Replay Gain - dbPoweramp vs. Foobar

Spoon - I'm a novice with replay gain so are you saying dBpoweramp uses EBU R128? If so, then when I add the replay gain DSP it is the EBU R128?

Also, I did some digging and read on the foobar's changelog that they use libebur128. I have no clue, libebur128 and EBU r128 are probably one in the same?

[Spoon]

Re: Replay Gain - dbPoweramp vs. Foobar

They are the same.

dBpoweramp has a checkbox to enable EBUR128 on the replaygain page (if you have R14.4)

[garym]

Re: Replay Gain - dbPoweramp vs. Foobar

The ReplayGain utility in foobar2000 does now automatically use the EBU R128 algorithm. However, note that the lufs target in foobar2000 is -18 rather than the -23 in dbpa. Based on what I've read, foobar2000 implementation used the -18 rather than -23 so that the result would be closer to the old RG method. Certainly in my limited testing, there can be a large difference between RG and RG with/ R128 method in dbpa. The foobar method has the two closer (because of the -18 target).

I personally am torn as to what to do with my own collection. Based on what I've read, the R128 method is likely a better volume normalization method, so I'd like to use this going forward. But in the meantime I'll have thousands of files with very different RG tag levels (new stuff with R128 -23 and old stuff with RG and occasional FB edited stuff at r128 -18. One possibility is to use R128 in dbpa but set to -18 target. Eventually I'll redo all my FLAC files with RG based on R128, but I'll wait until Spoon updates the RG codec so I can simply "convert to" RG with R128 rather than converting from FLAC to FLAC to get the new RG calc.

[mville]

Re: Replay Gain - dbPoweramp vs. Foobar

The ReplayGain utility in foobar2000 does now automatically use the EBU R128 algorithm. However, note that the lufs target in foobar2000 is -18 rather than the -23 in dbpa. Based on what I've read, foobar2000 implementation used the -18 rather than -23 so that the result would be closer to the old RG method.

garym, is the -18 lufs target gain fixed in foobar2000?

[garym]

Re: Replay Gain - dbPoweramp vs. Foobar

garym, is the -18 lufs target gain fixed in foobar2000?

Yes. I've not discovered any way of modifying this target in foobar2000. I think they did it this way because the ReplayGain option in foobar2000 was replaced with the EBU R128 method and they wanted the effect to be not dissimilar from the old RG method outcome in terms of loudness.

[2Bdecided]

Re: Replay Gain - dbPoweramp vs. Foobar

I'm sorry to jump in so late on this, but I think writing ReplayGain tags with a reference level of -23LUFs by default is a really bad idea.

IMO what foobar2k has done is correct, and what dBpa has done is incorrect.

The ReplayGain tags state the gain correction required to reach a pre-defined target loudness. The original ReplayGain standard included a loudness calculation, but you can swap that for another loudness calculation (e.g. EBU R128) without breaking anything as long as you maintain the same target loudness.

If you start writing tags with a different target loudness, but still using the same ReplayGain tag format, you break lots of things...

1. With ReplayGain enabled, files with the new tags play back at a different level from files with the old tags in all existing players - hence a standard that's supposed to remove annoying loudness differences actually creates them!
2. Some players (especially DAPs) design their audio pipelines carefully to manage levels and loudness (not quite the same thing) and expect a certain loudness when the user enables Replay Gain. If the tags have been written to target a completely different loudness, all the assumptions made in the design of that audio pipeline will be wrong. It should still work OK, but either the SNR, headroom, loudness matching, or volume of the player may not be as good as it was designed to be.
3. Some players have a ReplayGain-related pre-amp with some calibration points, and most players have some kind of ReplayGain-related pre-amp. The calibration points are now wrong, and the range of the pre-amp may now be inappropriate.

What ReplayGain defines...

http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=85536&view=findpost&p=736023

Good and bad ways of incorporating EBU R128 into ReplayGain...

http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=89841&view=findpost&p=841277

Spoon has explained that certain users (e.g. radio stations) want to use EBU R128 at -23LUFs via ReplayGain tags. The correct way to do this is to scan in the normal way, write the tags in the normal way (which means a -18 LUFs target), and to set the player pre-amp to -5dB.

I can see why, when providing a product to users, it makes commercial sense to meet their demands. However, using a -23LUFs level by default is breaking a decade-old standard by default.

Cheers,
David.

[rumpel]

Re: Replay Gain - dbPoweramp vs. Foobar

Hi,

being a long term EAC and foobar user, I just switched to dBpa, also because of the integrated ReplayGain scanning.

I agree with 2Bdecided in that it is a bad choice to use a standard level of -23 LUFS. Not only for the reasons 2Bdecided mentioned, but also because if users switch between "normal" RG and EBU R128 with standard settings, they will get drastically (~5 dB) different output levels. dBpa help says "typically the calculated gain values would be lower than standard ReplayGain". I think you should drop "typically" here as I doubt there is any real life example where the difference between classic and EBU R128 calculation with the same target level is > 5 dB.

Actually, until today I had been thinking that the RG target level was standardized, which I think is crucial for compatible tags. Obviously it is not, unfortunately.

Cheers,
rumpel

[Spoon]

Re: Replay Gain - dbPoweramp vs. Foobar

R11 of the DSP effects has corrected this, it is in the testing section of this forum.

[rumpel]

Re: Replay Gain - dbPoweramp vs. Foobar

Thanks for the info, I think that is a good decision. (R11 DSP info here)

[fschillizzi]

Re: Replay Gain - dbPoweramp vs. Foobar

Hi Spoon!

I ripped some CDs (not a lot) with dbpoweramp 14.4 and DSP effects 9.
I ripped in FLAC encoder and only use DSP effect ReplayGain (not ReplayGain apply), the original (track gain and album gain), not EBU R 128.

Now I am seeing a correction in the DSP 11.

Are my files loseless? or Do I have to rerip these CDs with DSP 11?

[Spoon]

Re: Replay Gain - dbPoweramp vs. Foobar

FLAC is lossless, visit codec central and install the [Replay Gain] utility codec, then use Batch Converter, select all your music and set the encoder to [Replay Gain] this will just recalculate the gain and write to the tags.