Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[ADDiCT]

NEW: Reads and writes track replay gain and peak values to MP3 id3 tags that is compatible with several other players.

How does this functionality interact with mp3gain tags that are already inside the files? A large part of my collection is already "mp3gained", and i'd hate to lose the tag information. What happens on audio analyze? More info about that topic is appreciated!

[gummbah]

10. NEW: Reads and writes track replay gain and peak values to MP3 id3 tags that is compatible with several other players.

Thanks, have been waiting for this a long time!
Are album gain values also compatible now or just replay gain?
I also noticed that automatic audio analysis on import still calculates replay gain values, even if they are already calculated by another player (and recognised in MC).

[ADDiCT]

10. NEW: Reads and writes track replay gain and peak values to MP3 id3 tags that is compatible with several other players.

Still waiting for info on that...

mp3gain writes tags, and i believe the values it writes are the ones it gathered during the analyze phase. If this is correct, and MC reads these tags, then the playback volume in MC will be way off, because MC would use the value before the mp3gain volume changes. And i don't even want to think about what would happen when MC overwrites the mp3gain tags. (;

More details about how MC handles the tags would be most appreciated.

[gummbah]

Still waiting for info on that...

mp3gain writes tags, and i believe the values it writes are the ones it gathered during the analyze phase. If this is correct, and MC reads these tags, then the playback volume in MC will be way off, because MC would use the value before the mp3gain volume changes. And i don't even want to think about what would happen when MC overwrites the mp3gain tags. (;

More details about how MC handles the tags would be most appreciated.

As far as I know mp3gain works differently from other replay gain analysers. It does not only analyse and write tags, but actually changes the mp3 files. The tags written by mp3gain are just there for the undo function (to actually change back the mp3 files). More info is on their webpage. Therefore I don't think the recent change in MC's replay gain writing will have any consequences when using mp3gain.

[jack wallstreet]

As far as I know mp3gain works differently from other replay gain analysers. It does not only analyse and write tags, but actually changes the mp3 files. The tags written by mp3gain are just there for the undo function (to actually change back the mp3 files). More info is on their webpage. Therefore I don't think the recent change in MC's replay gain writing will have any consequences when using mp3gain.

I concur.  I analyze all files using Mp3Gain and then do an MC analyze.  MC analyze typically adjusts all my files on MC by the same (about) amount due to the fact that MP3Gain has already leveled them using a very similar process (they are set to use a different base volume so there is a constant adjustment applied by MC). I use MP3Gain so I can play my files anywhere on any system (even if it doesn't support volume leveling).  I use MC replay gain to control my ipods.  However, I have not found a discernable difference between having the Ipod volume leveling function on or off (which seems to make sense since the basic files are adjusted with Mp3Gain)

[ADDiCT]

I'd still like to hear some words in the replay gain functionality from the MC devs.

Yes, mp3gain adjusts the volume by changing the mp3 file. But if MC reads the tags that mp3gain creates, there could be a problem. After proccessing the files with mp3gain, the tags will not contain the current volume level (the one proccessed with mp3gain), but the original volume level of the files. In previous versions of MC that was not a problem, because replay gain was a tag that was written in the "comment" field.

A mp3 track proccessed with mp3gain has the following tags (i believe these are "ape tags"):
MP3GAIN_MINMAX, MP3GAIN_ALBUM_MINMAX, MP3GAIN_UNDO, REPLAYGAIN_TRACK_GAIN, REPLAYGAIN_TRACK_PEAK, REPLAYGAIN_ALBUM_GAIN, REPLAYGAIN_ALBUM_PEAK

The question is: which of these tags is MC reading, and writing? I'd also like to know if MC's analysis does track gain or album gain. Track gain, for example, is pretty much useless for classical recordings, because "silent" tracks (which can be parts of a larger piece) would be much louder than the rest of an album when using track gain. The iPod's replay gain functionality (called "Sound Check") does track gain only, btw (which makes it useless for me).

P.S.: sorry for discussing the topic in the beta board, maybe it would sense to split the "replay gain"-related posts to a seperate thread.

[Alex B]

I have run a few tests and could answer some replay gain questions...

EDIT

Thanks Jim for gathering the replies into a new thread.

[JimH]

Here's the old thread: http://yabb.jriver.com/interact/index.php?topic=45465.0 (now locked)

[gummbah]

I have done a few tests.

Analysing files in MC and playing them in foobar: Track gain is recognised, but album gain is not.
Analying files in foobar and playing them in MC: Both track gain and album gain are not recognised

There is also a difference in of about 6 dB between replay gain values calculated by foobar and replay gain values calculated by MC.

[Alex B]

How does this functionality interact with mp3gain tags that are already inside the files? A large part of my collection is already "mp3gained", and i'd hate to lose the tag information. What happens on audio analyze? More info about that topic is appreciated!

If you mean the APE tags that the MP3Gain application writes MC does not care about them. It does not read them and it does not alter them. This has not changed.

During Audio Analysis MC measures and calculates the replay gain, track peak and BPM values. If tag writing is enabled it writes these three values and a calculated "Intensity" value to ID3v2 tags if the file format is MP3.

Starting from the v. 12.0.459 MC writes the Replay Gain and Peak values to ID3v2 TXXX tags instead of multi-frame COMM tags. The new system is compatible with some devices and some other software players.

MC can still read the old format, but on next tag write it will use the new format and erase the old tags.

If a file contains both tag styles MC reads only the new TXXX tags and erases the COMM tags on next tag write.

[Alex B]

Thanks, have been waiting for this a long time!
Are album gain values also compatible now or just replay gain?

Apperently only track replay gain is implemented. MC's analyzer does not measure the album gain and peak values similarly like some other programs do. A separate module calculates the album gain values dynamically on the database level. This approach has several advantages, but unfortunately hardware devices and some other programs are not as clever as MC and need hard coded values.

I also noticed that automatic audio analysis on import still calculates replay gain values, even if they are already calculated by another player (and recognised in MC).

Probably the files don't contain the BPM and Intensity tags. You can disable the automatic analyzer if you don't use the BPM and Intensity values. The Volume Leveling feature does not need them.

[Alex B]

mp3gain writes tags, and i believe the values it writes are the ones it gathered during the analyze phase. If this is correct, and MC reads these tags, then the playback volume in MC will be way off, because MC would use the value before the mp3gain volume changes. And i don't even want to think about what would happen when MC overwrites the mp3gain tags.

MP3Gain writes correct values to APE tags. If it changes the file's volume level it adjusts the Replay Gain and Peak values accordingly. It saves the undo info separately.

However, this is irrelevant because MC does not read APE tags in MP3 files. MC can read only ID3 tags.

[AoXoMoXoA]

Mp3gain writes correct values to APE tags. If it changes the file's volume level it adjusts the Replay Gain and Peak values accordingly. It saves the undo info separately.

However, this is irrelevant because MC does not read APE tags in MP3 files. MC can read only ID3 tags.

I have used MP3Gain on some files that had already been analyzed in MC for Replay Gain.
After running MP3Gain I again used MC's audio analysis on them, this time MC came up with different Replay Gain values.
To me this indicates that MP3Gain's ajustment does actually impact MC's replay gain values.

** I might add that this was not recently, so if things have been changed it may no longer be the case

[Alex B]

Analysing files in MC and playing them in foobar: Track gain is recognised, but album gain is not.

MC does not write TXXX style album gain tags to MP3 files.

Analying files in foobar and playing them in MC: Both track gain and album gain are not recognised

This works for me. MC reads the track gain values and dynamically calculates the album gain values if the album files are imported and correctly tagged.

Are you using the ID3.v2.4 setting in foobar2000? MC does not handle APE tags. If the values are in ID3v2.4 tags you may need to update MC's database: right-click > Library Tools > Update Library (from tags).

There is also a difference in of about 6 dB between replay gain values calculated by foobar and replay gain values calculated by MC.

The reference value is 83 dB in MC and 89 dB in foobar. MC's tag reader/writer corrects this automatically. The value in the file tag is compatible with programs and devices that use the 89 dB reference value.

[Alex B]

I have used MP3Gain on some files that had already been analyzed in MC for Replay Gain.
After running MP3Gain I again used MC's audio analysis on them, this time MC came up with different Replay Gain values.
To me this indicates that MP3Gain's ajustment does actually impact MC's replay gain values.

Actually it is the other way around. The measured values inside MC do not change automatically.
You need to make MC aware of the external audio data change. You need to reanalyze the files in MC or somehow import corrected Replay Gain and Peak values.

[Alex B]

10. NEW: Reads and writes track replay gain and peak values to MP3 id3 tags that is compatible with several other players.

Does this also apply to FLAC files using vorbis comments?

No. FLAC, OGG, APE, MPC and WMA have not changed. I don't know if JRiver has plans to change the other decoder plugins too.

[jimn]

Actually it is the other way around. The measured values inside MC do not change automatically.

This is correct. The regain values are stored in MC specific format just as before. You need to do a "Update Library from Tags" or "Update Tags from Library" for MC to read or write to the TXXX tags. Though MC may automatically "update"  new files, or files that have changed (like if you write out a replay gain analysis in another program).

Also as mention, MC will make the 6db adjustments when writing/reading to these tags.

Let me know if the implementation of these tags need adjustment.

Album values aren't implemented. MC has replay gain for album but not peak, so it would require work in the analyzer.

JimN

[Alex B]

AoXoMoXoA means audio data that has changed. MP3Gain can do that. It changes the scale value inside the actual MP3 audio data.

If compatible metadata (i.e. in ID3v2 format) has changed accrodingly, the "Update Library (from tags)" tool can update the values that are stored inside MC's library.

Since MP3Gain uses APE tags instead of ID3 tags the files need to be reanalyzed unless some other program is used for copying the values to ID3v2 tags.

MP3Gain is designed to adjust MP3 files directly. The APE tag format was chosen because the developer wanted to store the undo tag data in a proprietary format that would not be altered by ID3 tag writers.

[Alex B]

Album values aren't implemented. MC has replay gain for album but not peak, so it would require work in the analyzer.

Would it be too difficult to just pick the highest track peak value and store it in a new library field? This would be the correct album peak value too. MC alreadry reads all track gain values when it dynamically calculates the album gain value.

[jimn]

Would it be too difficult to just pick the highest track peak value and store it in a new library field? This would be the correct album peak value too. MC alreadry reads all track gain values when it dynamically calculates the album gain value.

I can't imagine it would be very difficult; it's just that I know nothing about the analyzer right now. What is the value of album gain and peak anyway? Why wouldn't you just adjust each track level according to the reference db?

[gummbah]

I can't imagine it would be very difficult; it's just that I know nothing about the analyzer right now. What is the value of album gain and peak anyway? Why wouldn't you just adjust each track level according to the reference db?

I prefer album gain over track gain. In particular in cases were there is a large volume difference between songs on an album. Introduction tracks or little soundbites between songs are often much softer then the regular songs. With track gain these songs sound much too loud compared to the rest of the album. Also for classical albums and live albums track gain gives different volumes for different songs, while album gain keeps the volume differences as they should be. 

[gummbah]

The reference value is 83 dB in MC and 89 dB in foobar. MC's tag reader/writer corrects this automatically. The value in the file tag is compatible with programs and devices that use the 89 dB reference value.

I am still a bit confused about this. I have updated my library using update library from tags. Foobar and MC still give different values.
Not sure what exactly you mean with "mc's tag reader/writer corrects this automatically".

[gummbah]

MC does not write TXXX style album gain tags to MP3 files.

This works for me. MC reads the track gain values and dynamically calculates the album gain values if the album files are imported and correctly tagged.

Are you using the ID3.v2.4 setting in foobar2000? MC does not handle APE tags. If the values are in ID3v2.4 tags you may need to update MC's database: right-click > Library Tools > Update Library (from tags).

I have foobar set up to write Id3V2 and Id3V1 tags. I have also checked the ID3v2 writer compatibility mode.
But MC does not recognised the values written by foobar.

[Alex B]

I am still a bit confused about this. I have updated my library using update library from tags. Foobar and MC still give different values.
Not sure what exactly you mean with "mc's tag reader/writer corrects this automatically".

On a file tag read the value is reduced by -6 dB so that the value inside the library becomes compatible with the system that uses a 83 dB reference value.

On a file tag write the value is changed by adding +6 dB to what you have inside the library.

I.e. if the value inside library is -10 dB it will be -4 dB in the "TXXX: REPLAYGAIN_TRACK_GAIN" ID3v2 file tag (and vice versa).

I have foobar set up to write Id3V2 and Id3V1 tags. I have also checked the ID3v2 writer compatibility mode. But MC does not recognised the values written by foobar.

Foobar's compatibility mode creates ID3v2.3 tags with UTF-16 character encoding. This format is not very well supported in MC12 despite the name foobar uses. Apparently MC12 does not recognize the replay gain tags if foobar's compatibility mode was used.

[llafriel]

I prefer album gain over track gain. In particular in cases were there is a large volume difference between songs on an album. Introduction tracks or little soundbites between songs are often much softer then the regular songs. With track gain these songs sound much too loud compared to the rest of the album. Also for classical albums and live albums track gain gives different volumes for different songs, while album gain keeps the volume differences as they should be. 

I mostly listen to single tracks so I use track gain. It would be cool if MC could recognize if you are listening to an album or single tracks and switch mode by it self. Like the option "Use gapless for sequential album tracks," only for replaygain.

[Robo983]

This is great! I hadn't looked for an update for a couple of weeks and glad to see this implemented. I do not use anything but MC so the discussion about the other ones won't affect me but would like a some pointers on how to "re-analyze" my library so I can start using this on my Rockbox MP3 players and IPOD.

What is the best way to re-analyze to get the new tags written just for the MP3 gain? I do not see a re-analyze option. Since MC already analyzed all I get is no files need to be analyzed. Do I need to do something special to delete the old replay gain tags?

[gummbah]

This is great! I hadn't looked for an update for a couple of weeks and glad to see this implemented. I do not use anything but MC so the discussion about the other ones won't affect me but would like a some pointers on how to "re-analyze" my library so I can start using this on my Rockbox MP3 players and IPOD.

What is the best way to re-analyze to get the new tags written just for the MP3 gain? I do not see a re-analyze option. Since MC already analyzed all I get is no files need to be analyzed. Do I need to do something special to delete the old replay gain tags?

You can re-analyse just by going to audio analysis, and uncheck skip analysed files.

[Alex B]

Actually, there is no need to re-analyse the files. It is enough to save the tags once.

Select the files > right-click > Library Tools > Update Tags (from library)

[flac.rules]

Ok, maybe I'm a bit slow, but just to clarify:

1. MC now reads replay-gain-info added by foobar2000 the same way foobar2000 does, and uses the same reference (89 db), including album gain?

2. MC writes RG-info the same way foobar 2000 does, with the same reference (89 db, meaning that the two will write the same value to the same song). But does NOT apply album gain?

3. MC writes id3v2.3 ISO-8859-1 when writing tags?

Have i understood everything correctly?

[Alex B]

1. The short answer is "practically yes".

The long answer is... hmm... long and rather techinical (but you asked for it )

1.a) The ID3v2 replay gain tags are interchangeable. The tagged value is based on the 89 dB reference level in both programs. MC still uses the safer 83 dB reference value internally (you can compensate this in the playback volume leveling options if you want).

1.b) The album gain value in MC is calculated from the track gain values by a separate module so the resulting album gain value is not exactly the same as you would get by analyzing the album directly as a single track (for album gain) and as separate tracks (for track gain) - foobar uses this method.

1.c) The used MP3 decoder is different. The peak values can be slightly different because MC measures the peaks after producing decoded integer PCM output. Foobar decodes to floating point and can measure occasional peak values that go over the full scale. This can happen when a loud audio track is compressed using a lossy encoder (like MP3). The orignal lossless PCM signal cannot go over full scale. These artificial peaks are usually very short - like 4 samples at max, which is about 1/1000 s. MC's decoder simply truncates them. In foobar you can preserve these peaks by using Replay Gain with clipping prevention. In my experience either system produces audibly identical output. These random peaks are simply too brief for the human auditory system. It is not possible to evaluate the difference.


2. Yes. (MC can write the album gain value if tag writing is enabled in the field's options, but it does not save it in a "foobar compatible" format. The saved tag is usable only in MC. In general saving this value is unnecessary because MC can calculate it on the fly.)


3. Generally yes. MC can read ID3v2.4 UTF-8 tags, but nornally it writes ID3v2.3 ISO-8859-1 tags (this behavior was chosen for compatibility reasons). If I have understood correctly MC can write ID3v2.4 UTF-8 tags when needed, but I have never seen that happening (my tags contain only standard latin characters).

[flac.rules]

I'm slightly confused about  answer 1a) and 2. More spesific the album-gain part. Is it to be understood that album gain is written to tags in the "old style" (as a comment)? And when you are taking about calculating on the fly, you mean setting adjustment to "automatic based on current playlist"?

Thanks for a very thourough and good answer by the way

[Alex B]

I'm slightly confused about  answer 1a) and 2. More spesific the album-gain part. Is it to be understood that album gain is written to tags in the "old style" (as a comment)? And when you are taking about calculating on the fly, you mean setting adjustment to "automatic based on current playlist"?

MC calculates the album gain value automatically if an imported album contains track gain values or when the track gain values are calculated by MC's analyzer. The value is saved in a library field. An album gain value in the file tag is useless for MC.

You can set MC to save the automatically calculated value in the file tags, but that is not enabled by default. (The setting is in the library field specific options: Tools > Options > Library & Folders > Standard library fields...)

When enabled it is indeed written to a "comment frame with description" if the file format is MP3.

The volume leveling feature can use the album gain value but its options are not related to the tag creation process.

Thanks for a very thourough and good answer by the way

You're welcome.

[gummbah]

So are there any intensions to change the way album gain is calculated and written in tags to make it similar to the way Foobar does?
If I understand correctly that would be the last step to make replay gain values fully interchangable with other players, am I correct?
Cheers

[Robo983]

Actually, there is no need to re-analyse the files. It is enough to save the tags once.

Select the files > right-click > Library Tools > Update Tags (from library)

I just upgraded to MC462 to take advantage of the replay gain changes on my Rockbox h320 and h120s and also my Sansa Clip. Instead of re-analyzing all of my files I did an Update Tags (From Library). The date modified shows 3/25/2008.

I first sync'd my clip and noticed only the files I re-analyzed showed In Queue to transfer. Since it doesn't have replay gain I didn't think much of it. Now syncing my h320 and h120 the files that show as date modified on the desktop of 3/25/2008 but on the h320 as 11/24/2007 are only showing On Device not In Queue to transfer with the new tag info.

Is this a MC462 bug or do I need to do a manual drag an drop?

[JimH]

Try build 465, just posted at the top of this board.

462 was doing auto analysis in the background and changing the date of the files.  465 should do analysis without changing the date.

[Robo983]

I think this may need a new thread but I searched on Sync Trigger and HH Sync to see if their was an explanation of what causes a file to be re-downloaded to a HH. Forgot to check the Wiki.

I loaded MC465 and did another Update Tags (from Library) now my date modified shows for all my audio as 3/27/2008 on the PC. My HH files show dates much older than that. Agian the only files that are In Queue are the ones a I actually did a re-analyze on.

Should the Date Modified field trigger a re-sync of the newer file over the older file on my HH? If it is not supposed to work this way I can do forced sync to the updated files on the PC but I guess I need to know what causes a sync to the updated file for the next time.

[Robo983]

Should the Date Modified field trigger a re-sync of the newer file over the older file on my HH? If it is not supposed to work this way I can do forced sync to the updated files on the PC but I guess I need to know what causes a sync to the updated file for the next time.

I got my answer over on the HH forum. Date modified is only for IPODs. Since that is the case I think it would be nice to update my non-IPOD handheld tags with the Update Tags (from Library) feature so that I would not have to transfer them. Is that possible?

http://yabb.jriver.com/interact/index.php?topic=45809.0