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 (Mono Fold-Down 45 & Stereo 115 from same flac source get differing r (Read 1287 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

ReplayGain (Mono Fold-Down 45 & Stereo 115 from same flac source get differing r

Hi there,

as the subject entry says I have a question regarding ReplayGain. To manage RG I use foobar 1.3.14 on Linux Ubuntu via Wine.
Beforehand I had converted a flac original (Peter, Paul and Mary - Blowin' in the wind (Bob Dylan Cover, 1963); 44,1kHz, 16bit CD standard) to opus mono (Fold-Down), 45kbps as I do with all roundabout pre 1968 master recordings. But since the stereo mix seemed really nicely done on this I also created a Stereo 115 kbps file and will deicde later on which one to keep.

I did a single-track RG Scan on both and it came out -4,73 for the mono version and -6,05 for stereo. Now it at least feels to me as if the stereo version is still louder even with RG taken into account on playback via VLC.

Any thoughts on this?

This isn't new to me since my whole music library has RG Info attached and still the mono recordings seem a bit more quiet. I am thinking of the pre White Album Beatles mostly. Now I do scan songs as part of one album if the source is the same, but I don't think that would account for as much of an effect as I believe to be noticing.

I thought it may be due to older records having been mastered differently as a trend of the time or early vinyl limitations // more background noise // so as not to have as much dynamical range as contemporary music (I am aware of the whole brickwall thing), which I would experience as the music not being played as loud in the peaks.


Being in the beginning stages of a physics bachelor study I am curious how RG is determined. I know it is more sophisticated than mere peak determination. Something like integrating the function's mean square and scaling accordingly?

Kind regards

Markus

Re: ReplayGain (Mono Fold-Down 45 & Stereo 115 from same flac source get differing r

Reply #1 –
Quote
I did a single-track RG Scan on both and it came out -4,73 for the mono version and -6,05 for stereo.
I assume that has something to do with the mix-down.   In general you have to reduce the levels to prevent clipping when summing the left & right channels.

Quote
Now it at least feels to me as if the stereo version is still louder even with RG taken into account on playback via VLC.
That could be just perception.

Quote
Now I do scan songs as part of one album if the source is the same, but I don't think that would account for as much of an effect as I believe to be noticing.
Album Gain maintains the relative loudness between songs so the quiet songs remain relatively quieter and the loud songs remain relatively louder, while adjusting the overall/average volume to your other songs/albums.

You can also run into limitations where boosting a quiet song to match the target volume would push the peaks into clipping.   If you've got ReplayGain configured to avoid clipping, the quiet songs may not reach the target volume.

Quote
I thought it may be due to older records having been mastered differently as a trend of the time or early vinyl limitations // more background noise // so as not to have as much dynamical range as contemporary music (I am aware of the whole brickwall thing), which I would experience as the music not being played as loud in the peaks.
If ReplayGain has "room to work" it should work equally well on digitized vinyl.    But, because of the all-pass filtering (phase shifting due to RIAA & inverse RIAA EQ and the conversion to/from mechanical storage), vinyl will generally have a higher peak-to-average ratio compared to a digital recording with equal-sounding dynamics.    So again, the ReplayGain adjustment may be limited by the peaks and you may not achieve the target volume.

ReplayGain isn't perfect...  I don't think it can be perfect...  If you asked two people to adjust two songs for the same loudness you'd probably get two different results, especially if they are different styles of music, or if one of the songs has especially-loud or especially-quiet parts and the other has fairly constant volume through the song. 


Quote
Being in the beginning stages of a physics bachelor study I am curious how RG is determined. I know it is more sophisticated than mere peak determination. Something like integrating the function's mean square and scaling accordingly?
Did you Google?   It takes an average (short-term RMS readings, I believe) and it takes the Equal Loudness Curves into account.    Because the equal-loudness curves are used,  the volume is matched at a pre-selected loudness (89dB SPL).

I'm pretty sure if you dig you can find the actual ReplayGain algorithm as well as the UBU R128 algorithm.

Quote
45kbps as I do with all roundabout pre 1968 master recordings.
Why so low?   Do you really need to save space?
Quote
I also created a Stereo 115 kbps file and will deicde later on which one to keep.
If you use Joint Stereo you don't need twice the bitrate for the same quality as mono.    Joint Stereo takes advantage of the fact that much of the sound is common to left & right, and that common information only needs to be compressed & stored once.  That part of the process (stereo to M/S) is lossless and completely reversible.   Those early Beatles Stereo recordings with no "center" would be an exception, but from what I recall, the White Album wasn't mixed that way .