Replay Gain on ogg vorbis vs flac

[ultrarunner2019]

Hey guys;

I had been playing all of my music (classic Rock) on my Android phone with Poweramp from .flac files which I had either ripped from my CD's or purchased on iTunes (then converted to flac from aac - I know this is silly, but...)
I applied RG to all of the flac files before uploading them to the phone (was a Galaxy A51 running Android 13). The volume level was good on all tracks.

Now, I have a new phone (Pixel 6 Android 13) and decided to convert to ogg to save time (and space) uploading to the phone.
Once again, I applied RG to the files as I converted them using Foobar2000 to .ogg level 6.
I left the default RG analyze using set to EBU R 128 and did not apply any preamp gain.

I am now noticing that many of the tracks are not at the same level; some are too soft, and some too loud.

I have checked my settings in Poweramp (3.x build 957 purchased version) and they're all the same as I had before.

I have also listened to the tracks in Foobar2000 on my Windows 11 machine, and there is a noticeable change in volume between some tracks. That said, when I load some of the flac files into Foobar, the level is good on all tracks.

I'm thinking that perhaps because I applied RG in Foobar2000 during conversion from flac to ogg, that something went wrong; or that ogg files just don't play so nice with RG as flac do.

If I cannot fix the issue, I will probably just replace the ogg files on the phone with the flac ones - I did get the 256GB version of the Pixel 6 so I have enough space.

Any ideas?

Thanks
Ultrarunner

[andrewilley]

It's worth checking that PA has read the RG information correctly by long-pressing on the album art area in the Player Screen and selecting Info/Tags, you'll see the RG data near the top.

If it shows there, then the problem was probably in applying the RG data in the first place.

1) Why not just go back to the original AAC files and apply ReplayGain to them in foobar2000 ? As I think you know, you won't have improved them in any way at all by converting them to FLACs, and then later converting those FLACs to OGG just seems to be making even more problems for yourself. Although it won't necessarily make the audio any worse assuming you use lossless encoding, that is still basically just putting FLAC data into an OGG container.

2) If you really must convert to OGG, I suggest you do that first from the original AACs, and then do the ReplayGain update within foobar2000 as a batch operation on the OGGs afterwards. But again, easier to just leave them as AAC.

Oh, and check that your conversion progress isn't creating OPUS data in an OGG container, as there can be some case sensitivity issues with ReplayGain data in OPUS files:

Andre

[ultrarunner2019]

I just realized that when RG is applied during conversion in Foobar2000, the change is made to the music itself, but if RG is applied after conversion - using the scan and apply functions, then RG is applied in the tag. I think this may be what is happening for me. I did allow Foobar to apply RG during conversion to ogg.

The main reason I used flac instead of aac is that much of my music library was ripped using EAC from the CD's. I wanted a compressed, yet lossless format. I realize that converting aac from iTunes to flac isn't doing anything to improve the quality; it was just a habit I got into.

I guess I'll go through my entire library, and if I have an aac file, I'll just apply RG and send it to the phone for PA. If it's a .flac file, then I guess I'll convert to aac, apply RG and send to phone.

I could make things easy on myself for now by simply replacing the ogg files on the phone with the flac files on my PC. But eventually, I may run out of storage space, since the 256GB on the Pixel 6 is all that I'm ever going to have - there is no SD card slot on this phone.

 

[andrewilley]

There are several ways to increase the volume of an audio file.

During lossy conversion (such as compression) you could simply make the amplitude of each sampled value bigger by a set amount, so the highest sample value in the file becomes approx. 100%, and the volume of the entire content is raised by a similar percentage. That's how an audio editor would do it, and maximising volume like that is also available when ripping CDs, etc. It is destructive though (just like lossy compression) as there's no way back to the original sample data if you change your mind later.

With some formats, such as MP3, which work on small "frames" of data, you can apply an individual gain value to each frame. This does not change the actual encoded audio data, it merely tells the playing software to adjust the volume for that particular frame at playback time, and it should work automatically with all players as it's part of the music data. It can be reversed though, simply by changing that extra gain value back for each block.

Finally there's ReplayGain, which works by adding a single gain value into the header data of the file, which says to make the whole file louder or quieter by a set amount. The exact figure is worked out in advance and inserted into the file header - usually by software such as foobar2000 which scans through the whole song and works out the maximum peak that occurs anywhere in the content. This is also non-destructive, as simply removing the header tag reverts the file back to its exact original state. It is up to the player software to implement the gain adjustment at playback time though (which is often configurable) and not all players implement it.

7 hours ago, ultrarunner2019 said:

If it's a .flac file, then I guess I'll convert to aac, apply RG and send to phone.

If the FLAC was originally created from a lossy compressed AAC, all that will do is make the quality worse, it won't simply revert to the original AAC data. Far better to stick to the original AAC files if you still have them on your computer. FLAC is only non-lossy if it was created that way from the original PCM data.
 

As an aside, the availability of an SD Card slot is actually one of my key purchase decisions when buying a new phone. That's why I no longer buy Samsung's supposed "flagship" S-series models and went for a cheaper A-series version instead.

Andre

[ultrarunner2019]

Since I was advised - and actually have understood for a while - that converting from one lossy format to flac does nothing to improve the quality of the audio, and that converting from flac to ogg (another lossy format) causes some degradation, I have decided on a new plan:

First, I re-downloaded all of the tracks I purchased from iTunes. Most are in the unprotected m4a format, but a couple dozen are older downloads, which were protected in m4p files. Those are the ones I originally burned to CD's, then re-ripped into flac or ogg. Perhaps if I were to purchase a newer (ie; remastered) version of some of those tracks, I would get the m4a rather than the m4p files. I have some tracks I purchased from 7Digital in mp3 format @ 360K (or is it 340K? - not important here). I was never able to distinguish the difference between a Rock & Roll track I ripped from a CD into flac and one I purchased on iTunes. My player on my Android phone has always been Poweramp (purchased ver). On my PC, I use MusicBee.

I 'sync' tracks manually. Normall, after I purchase them from iTunes I would convert to .flac and send them to the phone after applying ReplayGain.

But now, to get back on track, I am simply applying RG to the AAC (m4a) files or mp3 files I purchased, and sending them directly to the phone. For the tracks I ripped from CD's into .flac, I will continue using .ogg Vorbis. I tried using the AAC converter in Foobar2000, but there was an error. When I looked it up, I read that updating the converter files might help, but in my case it did not. Since OGG is working fine, and Poweramp will read ogg or aac.

I think that the whole reason I started using .flac - even for aac files purchased on iTunes - is that I liked to create 'DJ mixes' with many tracks in Reaper DAW. My thinking was that if I wanted create a mix-down that I would want to have all of my tracks in lossless format.
I never even considered the fact that Reaper (and any other DAW) will happily open m4p, ogg, mp3, and a slew of other compressed formats. So, when I create a mix, I can just as easily use the aac files, then render the mix-down as anything I want. I seriously doubt that I will be able to detect any difference between a mix-down rendered to aac compared to flac.