replaygain debate, track vs album, random mix

[MrSinatra]

i often find that the things which are the most obvious to me, are the hardest to explain or justify.

here's my contention. i believe all files should have track and album RG tags. thats not debated between myself and my unnamed contrarian.

the debate is that i go on to say that album RG should ONLY be used (ie. applied on playback) when playing back an album in sequential track order. so, in other words, i would use track RG for random mixes, and i would also use it if i was doing a random mix of a single album. (i believe sbs "smartgain" operates the same way, correct?)

he however, contends that album RG is the "audiophile" RG, and should be used basically for all listening scenarios, except those that would be more like "Radio" which he seems to interpret as dance music. i think that the terms "audiophile" and "radio" have nothing to do with this debate.

he posted this (absurd, to me) link to back up his claim:

http://replaygain.hydrogenaudio.org/radio.html

now, to me, its not even debateable that track RG is right (and the only choice) for a whole collection in random mix. using album RG for that is, imo, absurd on its face. but its harder for me to explain my contention that one should use track RG when listening to a single album in random mix (shuffle) mode.

can someone please help me explain the obvious?

i have tried to explain that the changes in loudness from track to track when played out of order will not be what the artist intended when using album RG, but of course, the artist didn't intend random mix itself either. also, in an "absolute" sense, using album RG will make the first track the same difference in loudness from the fifth track, even though that difference was never meant to be back to back.

what say you all?

[garym]

Good questions. First, yes in my opinion, all tracks should have tags containing album and track RG values (whether one uses them or not currently).

I use "smart gain" in SB, as I believe that when playing a complete album I want to preserve the relative loudness among songs on the album. And of course, with random tracks it will use track gain (which to me seems appropriate). With regard to the "audiophile" issue related to always using album gain, I'd argue that once you are not listening to the entire album in the artist's (or producer's) desired order, all bets are off anyhow (Dark Side of the Moon anyone?). As an aside, I'd assume "audiophiles" would prefer not to have any volume adjustments made in the digital playback chain so wouldn't use album or track gain.

With regard to the only real debatable question I see (for me) in your question is the track gain for random play of an album. I guess in my case I never actually do this. I listen to an album (in track order) or I listen to random songs, which are always from multiple albums (even if from the same artist). This said, I tend to agree that once you go to random, track gain is the default choice.

One question though, and I don't know the answer. Does smartgain automatically use track gain when playing a single album but in random order? I would of guessed otherwise (that is SB sees a complete album in the playlist and chooses album gain, even if random). But again, don't know.

Another aside: The mp3 or AAC files I have (either only or from my FLAC files) I did use mp3tag to convert my itunes SOUNDCHECK values to the Album RG value (given that with SOUNDCHECK you can use only one value). So given that I could only have one, I went with Album Gain. I suppose this doesn't sound logical, given that with track value, I could use track gain for random play and simply turn OFF soundcheck on ipod for album play. But I didn't want to be bothered with that. And frankly, I rarely listen to IPOD anymore given my proliferation of SB players everywhere.

[jeebers]

Personally I use Album Gain for everything, even random play. I've tried track gain but I found it far more jarring in terms of the *perception* of differential volume.

I listen to a wide range of genres, so if I had a random mix where a delicate acoustic guitar track was followed by full-on four-to-the-floor techno, then using track-gain just sounds wrong to my ears. The guitar track *should* be quieter, therefore using album gain retains this volume differential, whereas track gain removes it.

[MrSinatra]

Good questions. First, yes in my opinion, all tracks should have tags containing album and track RG values (whether one uses them or not currently).

I use "smart gain" in SB, as I believe that when playing a complete album I want to preserve the relative loudness among songs on the album. And of course, with random tracks it will use track gain (which to me seems appropriate). With regard to the "audiophile" issue related to always using album gain, I'd argue that once you are not listening to the entire album in the artist's (or producer's) desired order, all bets are off anyhow (Dark Side of the Moon anyone?). As an aside, I'd assume "audiophiles" would prefer not to have any volume adjustments made in the digital playback chain so wouldn't use album or track gain.

yep, i use smartgain too. but he apparently nearly always uses album gain, even when doing a completely random mix. whats funny is DSOTM is EXACTLY the example i made, even going so far as to post the values and doing the math on differences.

i mean, on a "whole collection" random mix, what good is playing "speak to me" at -3.69 album RG when its track value is +15 something?

for me, the only real debate is could one use album RG for a single album in shuffle mode? i tend to think not, but its hard to explain why.

With regard to the only real debatable question I see (for me) in your question is the track gain for random play of an album. I guess in my case I never actually do this. I listen to an album (in track order) or I listen to random songs, which are always from multiple albums (even if from the same artist). This said, I tend to agree that once you go to random, track gain is the default choice.

exactly the same with me. i haven't listened to a single album in shuffle mode for over 20 some years. but as a thought experiment...

One question though, and I don't know the answer. Does smartgain automatically use track gain when playing a single album but in random order? I would of guessed otherwise (that is SB sees a complete album in the playlist and chooses album gain, even if random). But again, don't know.

i always assumed smartgain took the track numbers in the playlist into account, meaning it looked for sequential playback to enable album RG, but i have no evidence to back that up. surely an expert here knows exactly how smartgain functions?

Another aside: The mp3 or AAC files I have (either only or from my FLAC files) I did use mp3tag to convert my itunes SOUNDCHECK values to the Album RG value (given that with SOUNDCHECK you can use only one value). So given that I could only have one, I went with Album Gain. I suppose this doesn't sound logical, given that with track value, I could use track gain for random play and simply turn OFF soundcheck on ipod for album play. But I didn't want to be bothered with that. And frankly, I rarely listen to IPOD anymore given my proliferation of SB players everywhere.

if it were me, i would maintain both sets of tags. so, do RG tags on everything, (winamp does this well), and then let itunes add soundcheck tags as well. (u gotta do winamp first, b/c it deletes frames it doesn't understand)

[garym]

if it were me, i would maintain both sets of tags. so, do RG tags on everything, (winamp does this well), and then let itunes add soundcheck tags as well. (u gotta do winamp first, b/c it deletes frames it doesn't understand)

yep, I do this. I have RG tags (album and track) AND SOUNDCHECK values on all the mp3/aac tracks). RG only on FLAC tracks as these aren't used in itunes.

My point was that my SOUNDCHECK values are "based on" album RG values (done with mp3tag). And since I'm mostly playing songs in either SB players or foobar2000 on my computer, I'm always using the RG values. Soundcheck only comes into play on my ipods/iphones.

regarding your debate. I think I come down on the "use track gain when randomly playing a single album" even though I never actually do this (shuffle a single album that is).

[MrSinatra]

Personally I use Album Gain for everything, even random play. I've tried track gain but I found it far more jarring in terms of the *perception* of differential volume.

I listen to a wide range of genres, so if I had a random mix where a delicate acoustic guitar track was followed by full-on four-to-the-floor techno, then using track-gain just sounds wrong to my ears. The guitar track *should* be quieter, therefore using album gain retains this volume differential, whereas track gain removes it.

i understand what you are saying, but while i sometimes have that happen, i find using album RG in that scenario worse far more often, as in the 'speak to me' example above. (it would depend on the amount of variation or to put it another way, volatility in your collection between a tracks two RG values)

also, what you are saying is basically a subjective approach, rather than an academic philosophical approach, but for you, that only makes sense of course, (and i certainly wouldn't begrudge you it)... however, its not really objectively defensible.

[MrSinatra]

Another aside: The mp3 or AAC files I have (either only or from my FLAC files) I did use mp3tag to convert my itunes SOUNDCHECK values to the Album RG value (given that with SOUNDCHECK you can use only one value). So given that I could only have one, I went with Album Gain. I suppose this doesn't sound logical, given that with track value, I could use track gain for random play and simply turn OFF soundcheck on ipod for album play. But I didn't want to be bothered with that. And frankly, I rarely listen to IPOD anymore given my proliferation of SB players everywhere.

yep, I do this. I have RG tags (album and track) AND SOUNDCHECK values on all the mp3/aac tracks). RG only on FLAC tracks as these aren't used in itunes.

My point was that my SOUNDCHECK values are "based on" album RG values (done with mp3tag). And since I'm mostly playing songs in either SB players or foobar2000 on my computer, I'm always using the RG values. Soundcheck only comes into play on my ipods/iphones.

regarding your debate. I think I come down on the "use track gain when randomly playing a single album" even though I never actually do this (shuffle a single album that is).

so i'm confused... did you base RG tags on soundcheck, or soundcheck tags on RG? and how did you use mp3tag to translate one to the other?

and yeah, i use track gain too. i still can't get my head around the idea of doing multiple albums tracks shuffled using "album gain" and thinking that was somehow the "audiophile" choice.

[JJZolx]

"Audiophile" ReplayGain? No such thing. If you're _listening_ then you're also able to control volume, and you probably use it. I know I do. Anything other than listening while parked in front of the stereo system in a quiet room with the preamp volume control nearby, is just a form of background listening. Not that there's anything wrong with background listening. I suspect that at least 3/4 of my own listening is done like this, as I'm cooking, eating, doing housework, exercising, reading.

I enable SmartGain on synced players around the house, but I'm not sure I would notice the difference between it and TrackGain. I turn it off on the main system when I'm _listening_.

[garym]

so i'm confused... did you base RG tags on soundcheck, or soundcheck tags on RG? and how did you use mp3tag to translate one to the other?

and yeah, i use track gain too. i still can't get my head around the idea of doing multiple albums tracks shuffled using "album gain" and thinking that was somehow the "audiophile" choice.

I base Soundcheck on RG tags. I use an action in mp3tag that reads either album or track RG value and then replaces the SOUNDCHECK value (COMMENT ITUNNORM) with the value from RG tag. In my case the ALBUM RG tag. see:

http://forums.mp3tag.de/index.php?showtopic=9139&hl=soundcheck

This way, when I'm playing, say DSOTM, on my IPOD and SOUNDCHECK is turned on, I don't hear the change in volume that I'd get from one song to the next if SOUNDCHECK was using its normal track level numbers (instead of Soundcheck created FROM album RG values).

[MrSinatra]

ah, i get it Gary, that makes total sense since you can't have both types of tags in the soundcheck scheme.

"Audiophile" ReplayGain? No such thing. If you're _listening_ then you're also able to control volume, and you probably use it. I know I do. Anything other than listening while parked in front of the stereo system in a quiet room with the preamp volume control nearby, is just a form of background listening. Not that there's anything wrong with background listening. I suspect that at least 3/4 of my own listening is done like this, as I'm cooking, eating, doing housework, exercising, reading.

I enable SmartGain on synced players around the house, but I'm not sure I would notice the difference between it and TrackGain. I turn it off on the main system when I'm _listening_.

totally agree Jim, but why disable smartgain when intently listening? i find RG always suits me.

(and yeah, that link which seems like the official source of RG def has some strange ways of describing what they mean)

[garym]

ah, i get it Gary, that makes total sense since you can't have both types of tags in the soundcheck scheme.



totally agree Jim, but why disable smartgain when intently listening? i find RG always suits me.

(and yeah, that link which seems like the official source of RG def has some strange ways of describing what they mean)

not sure about Jim, but I assume this has something to do with the idea that to get "everything" from the digital stream, the volume should be set to 100% from the Touch/Transporter, etc. Even with the volume set to 100%, with a negative RG value being applied, the digital volume would be "reduced" from 100%. Smarter people than me around here can explain how this may not matter much as long as the digital volume is set to at least 80% or so (given the extra bits between 16 and 24, etc.).

[KarolG]

I use "Smart Gain" and few days ago I have discovered a new song in my library. This happened when I played random stuff. The new song was strange and I wondered WTF is it ?
Then I realised it is just "Outside the Wall" from "The Wall" by Pink Floyd played with its +20dB Track Gain instead of +1.32dB Album Gain

[Phil Leigh]

Smart Gain is perfect as far as I am concerned. I have never felt the need to turn it off, even for "critical listening". The RG for well-engineered albums/tracks will be <5dB anyway so, assuming you have volume on max, nothing bad is going to happen to replay quality.

[Nonreality]

Mr. Sinatra you are right.

And your friend? is wrong. It's really that simple. I've been a big fan of replay gain for years and it works as it should. On albums you want to keep the same sound levels that it's recorded as. When playing tracks from many albums you want an constant sound level that doesn't blow you away with drastic changes. He may be right in saying it's not perfect but when playing tracks who cares? I don't want the sound going up and down in huge changes. He likes it that way, more power to him. It's not the right way for listening to random tracks. The sound levels are not measured against other albums so I like to equalize them. It makes listening much better. I've heard this arguement from the audiophools before and it doesnt' hold water. Do what you enjoy and don't buy $1000 power cords either. I do use dbpoweramp for all my replay gain setting and always have. It shouldn't matter but I suppose you could use a crappy program that could mess things up.

[Nonreality]

Same here

Smart Gain is perfect as far as I am concerned. I have never felt the need to turn it off, even for "critical listening". The RG for well-engineered albums/tracks will be <5dB anyway so, assuming you have volume on max, nothing bad is going to happen to replay quality.

And I know you know a thing or two about music. I've never had a problem with the SB Smart gain. Works like it should.