"Tagged" EQ PreSets - Possible?

[KindOfBlues71]

It's closed minded to realize "adding every single idea can make the software worse, not just better, so everything needs to be carefully challenged and weighed"?

It seems closed minded to me to dismiss the project as closed minded just because you don't agree with certain choices it has made.

It also seems you didn't read what I wrote. I specifically said "Rockbox seems closed-minded at times," not Rockbox is always closed-minded about everything. And I was referring to Febs' comment about the canned response regarding the intent of the EQ in Rockbox.

Every idea should be challenged. Period. Socrates once said "The unexamined life is not worth living." Simply accepting an idea, or simply dismissing it, would be closed minded. But to challenge it, seek its weaknesses and strengths, and once fully examined, to make a decision, is quite opposite. But to call it closed minded just because you think the decision is wrong is still foolishness.

Telling people to use a DSP on their music tracks because the Rockbox EQ isn't intended for that purpose doesn't seem like one of those fully examined decisions you just talked about. But I'm foolish, so what the heck do I know?

[TexasRockbox]

Not all music is recorded flat, mastered from flat tapes, or even mastered from tape at all.  Some older music was recorded and mastered with EQ applied so as to reduce problems with cartridge tracking and also to allow for more music (timewise) in the grooves.  Some older late 60's and 70's master tapes were encoded with Haeco-CSG and CDs have been mastered using those tapes.  There are many reasons to apply EQ -- even if just for one's own "taste".

An EQ tag would be nice, but I do try to correct obvious (in my view) problems before encoding.

[Chronon]

I would agree with Llorean that as it currently exists (with only persistent, absolute settings) the EQ is more suited to compensating for variations in hardware performance.

I would strongly argue a five-band parametric equalizer with "only persistent, absolute settings" (will disagree with your word choice later) is no better or worse suited to compensating for variations in hardware performance.  I would argue, in fact, that (not to be too pedantic on word choice (but words not only express, but shape, thoughts)) that the method of setting the features of a tool is not what determines its functionality ("suited"), but rather what determines its convenience.

I have seen no compelling argument yet that the Rockbox EQ is "suited" or "more suited", much less "intended", to solely compensate for inadequacies in hardware.  This appears (to me) to be a step-back defense of the previous proposition that the EQ's purpose was to compensate for hardware.  Febs did a solid job attacking the "purpose" argument, but now the same argument appears to me to have been resurrected simply by casting either the s/purpose/suitability/ or s/purpose/intention/ spell.

EDIT:  Rewrote last sentence.

EDIT 2:  hmm, how to describe?  trim of overly pedantic disagreement.

I made no assertion about the impossibility of using the EQ (as it currently exists) for any purpose whatsoever.  I was expressing the idea that the current EQ has settings that persist (once I set them they remain at that setting until changed) and are specified by an absolute (not relative) value, and as such it seems much less suited to accounting for eccentricities of individual files.  This addresses how we choose to specify the parameters for the software EQ (and their persistence), not some kind of assumed purpose on the part of the user or the firmware itself. 

I should have said that the EQ as it currently stands is better suited to applying persistent, steady state alterations to the sound.  I sloppily conflated hardware performance with persistent, steady state behavior.

Hopefully, my rephrasing will allow you to understand what I meant.  It sounds like much of your post points to the same ideas that I mentioned.  For example, you mentioned that the settings might auto-revert, but this is just alternate wording for temporary settings.

[soap]

Chronon, I'm going to be busy as can be the next 36 hours, so please don't take my lack of reply as an ignore.
That being said - perhaps I need to reread your latest reply - but I think there is still a difference of opinion here outside straight semantics - and I think you missed my point, but perhaps I am the one missing your point - will get back to you.
 

[Chronon]

Okay.  I will be a bit busy here too.  I'll see if I can try to make the comparison a bit clearer for now.

I had in mind a settings file that can be stored alongside a file that would apply relative changes (i.e. the EQ band gains would be delta values from their current settings) whose effect would revert at the end of playback of that file or directory (temporary).  These properties sound consistent with what I read in your post about your proposed .eq files. 

My suggestion that the EQ is currently more suited to steady state modifications only reflects on the current way that its parameters are stored and handled, not on how a user is permitted to use the EQ.  The kernel of my thinking on this led me to consider two types of settings with different notions of scope, persistence, etc.

[KindOfBlues71]

I'm not sure about the directory structure for other players, but where would Rockbox store the per-track EQ file when using, for example, the iPod's default directory structure? Would it make more sense to have all the EQ files in a central folder within the Rockbox directory, with each EQ file containing directory paths to the tracks/albums the specific settings should be applied to? Alternatively, one could have the option to choose where to save the EQ file to, similar to how Playlists and Bookmarking is handled.

-KindOfBlues71

[Llorean]

This offers less flexibility (files are harder to move, for one thing, you can't rearrange your collection) than simply putting them beside the original file.

There's no reason to add limitations to the system just because iTunes has a silly structure. We can't, or shouldn't, spend forever working around it, and sometimes it's better just to say "they'll have to figure out which files are their songs for this one." If there was a graphical UI for creating these individual EQ addition files, it could put them in the right place anyway.

I would take Chronon's idea one step further, personally, and apply all EQ modifiers in the path.

If you had /music/rock/ACDC/Back in Black/song.mp3 and there was a rock.eq in the rock folder, and a Back in Black EQ in the album folder, they should both be "added" to the static EQ, because the Rock one is there for your preferences for the genre, and the album one is there to correct for oddities in mastering the Album. The offsets should be cumulative.

[Old Schooler]

Hello All,

I am glad to see all of the response to my original post about this subject, though I didn't mean to stir everything up.

However, I do agree essentially that EQs and their settings are "supposed" to be used for hardware deficiencies, but that is sort of what caused some, well allot, of my variances in some of my music as far as the way I transferred it to MP3 format.

Coming off of Cd's, or digital sources were no problem. But coming from LPs and tape, I tried to "rejuvenate" some of them and through using a set of headphones that didn't really have good enough response for that purpose, some of my recordings are notably weighted in certain areas.

The topic was mentioned about redoing them, which would be allot of work, but it could be done. So, in regard to that, this question is somewhat off-topic for this thread but it does have a place:

If I "re-encode" an existing MP3 of maybe 160kb with a program such as Magix Audio Cleaning Lab with better EQ settings for example, is the MP3 file re-compressed and thereby possibly losing quite a bit of quality, or is the "response curve"(?) simply changed within the MP3 so it plays back with the new EQ effects?

Thanks again for RockBox. After using it on both my old JBR (boy I miss that thing) and now my H10, I don't think I would have kept either without it.

Dave

[MarcGuay]

If I "re-encode" an existing MP3 of maybe 160kb with a program such as Magix Audio Cleaning Lab with better EQ settings for example, is the MP3 file re-compressed and thereby possibly losing quite a bit of quality, or is the "response curve"(?) simply changed within the MP3 so it plays back with the new EQ effects?

From what I understand, it's impossible to modify and save an MP3 file without re-encoding it, which necessarily means a loss of quality.  The best thing to do is to work with the originally recorded WAV files, which I assume you kept around for a time like this...?  If not, in the future, you might want to consider using a lossless compressor like FLAC so you don't run into these kinds of problems.  Salut...

[Chronon]

Hello All,

I am glad to see all of the response to my original post about this subject, though I didn't mean to stir everything up.

However, I do agree essentially that EQs and their settings are "supposed" to be used for hardware deficiencies, but that is sort of what caused some, well allot, of my variances in some of my music as far as the way I transferred it to MP3 format.

Coming off of Cd's, or digital sources were no problem. But coming from LPs and tape, I tried to "rejuvenate" some of them and through using a set of headphones that didn't really have good enough response for that purpose, some of my recordings are notably weighted in certain areas.

The topic was mentioned about redoing them, which would be allot of work, but it could be done. So, in regard to that, this question is somewhat off-topic for this thread but it does have a place:

If I "re-encode" an existing MP3 of maybe 160kb with a program such as Magix Audio Cleaning Lab with better EQ settings for example, is the MP3 file re-compressed and thereby possibly losing quite a bit of quality, or is the "response curve"(?) simply changed within the MP3 so it plays back with the new EQ effects?

Thanks again for RockBox. After using it on both my old JBR (boy I miss that thing) and now my H10, I don't think I would have kept either without it.

Dave

Sometimes it's good to stir things up a bit.   

If you re-encode then you'll definitely lose quality.  As MarcGuay suggested you would be better off doing a fresh rip or transcoding from lossless again with better settings. 

As for correcting such shortcomings on the fly, we'll see if this topic gels into a well defined feature request.  So far it appears that there might be some support for the idea of implementing a new kind of relative (and cumulative) EQ setting which should allow you to apply EQ settings to a particular set of files.  I like Llorean's recent suggestion of summing up all of the cumulative EQ modifiers found in the path to a given file. 

[KindOfBlues71]

This offers less flexibility (files are harder to move, for one thing, you can't rearrange your collection) than simply putting them beside the original file.

There's no reason to add limitations to the system just because iTunes has a silly structure. We can't, or shouldn't, spend forever working around it, and sometimes it's better just to say "they'll have to figure out which files are their songs for this one." If there was a graphical UI for creating these individual EQ addition files, it could put them in the right place anyway.

I personally hate the iPod directory structure and how it re-names files.  As I understand it, one reason for it is to cut down on seek-times/hard drive usage.  That's one reason why I've stuck w/ the iPod file structure.

Back to the topic, what isn't flexible about giving users the option to either store the eq/config file in a single folder or in the directory containing the files to apply the settings?  Those are the options Rockbox users have for Playlists, albeit from two different menus - Playlist Catalog and Playlist.  If the developers were concerned about users restructuring their music directories they wouldn't have made the default location for a dynamic playlist the root, nor the default location for saved playlists the Playlist folder.

I would take Chronon's idea one step further, personally, and apply all EQ modifiers in the path.

If you had /music/rock/ACDC/Back in Black/song.mp3 and there was a rock.eq in the rock folder, and a Back in Black EQ in the album folder, they should both be "added" to the static EQ, because the Rock one is there for your preferences for the genre, and the album one is there to correct for oddities in mastering the Album. The offsets should be cumulative.

I disagree with applying EQ cumulatively.  If you take the time to EQ a specific album to your liking, why would you want anything to alter it?  If you're listening to an artist where the Rock EQ is applied and then play a track from the Back In Black album, all EQ settings should be completely changed to the Back In Black EQ only, not a cumulative EQ of both.

[Chronon]

I personally hate the iPod directory structure and how it re-names files.  As I understand it, one reason for it is to cut down on seek-times/hard drive usage.  That's one reason why I've stuck w/ the iPod file structure.

Back to the topic, what isn't flexible about giving users the option to either store the eq/config file in a single folder or in the directory containing the files to apply the settings?  Those are the options Rockbox users have for Playlists, albeit from two different menus - Playlist Catalog and Playlist.  If the developers were concerned about users restructuring their music directories they wouldn't have made the default location for a dynamic playlist the root, nor the default location for saved playlists the Playlist folder.

Well, this whole idea still needs to be fleshed out properly.  I can see that it would be fairly straightforward to simply follow the path of each file in a playlist to look for EQ modifiers.  However, sensibly accommodating Apple's file storage scheme adds some difficulty.  You are advocating for storing the EQs all in one place.  But this doesn't give Rockbox an easy way to tell if an EQ needs to be applied to a given file or not without reading the file lists in each EQ that you have stored there.  It sounds messier and slower than a method that simply groups the EQ settings with the files that you want altered.

I disagree with applying EQ cumulatively.  If you take the time to EQ a specific album to your liking, why would you want anything to alter it?  If you're listening to an artist where the Rock EQ is applied and then play a track from the Back In Black album, all EQ settings should be completely changed to the Back In Black EQ only, not a cumulative EQ of both.

I can see why you might prefer that.

From the other side of the issue, relative settings (modifiers) would allow a presumptive .eq file to act as a filter.  So I would imagine storing a .cfg with absolute settings that correspond to my best assessment of a flat response for the given headphones that I'm using.  This is how I store my EQ settings currently -- I have a .cfg for each pair of headphones that I use with my DAP on a regular basis.  This serves as a starting point and you could further apply filters to this to suit individual files without losing your reference point for the given headphones you're using.

Anyway, it seems like soap has some input regarding this whole issue so we should probably let him weigh in on all of this before we get too ahead of ourselves.

[soap]

I'm actually forgetful about what my main "beef" was with Llorean and Chronon - so why don't I just issue an apology?

If we're bar-talking this subject, though, here are my thoughts:

• Equalizer settings stored in a .cfg file should be absolute.
• Equalizer settings stored in a .eq (for sake of conversation) file can be relative.
  
  • Perhaps a menu toggle to make .eq files relative (cumulative) or absolute.
  • I am attracted to .eq files being searched for in exactly the same hierarchy album art is searched for.
    
    • This solves the iPod database issue.
    • This maintains only one "pattern of association" to learn.
    • This allows you to tweak EQ settings at exactly the granularity you desire.
      
      • You can have your "rock" .eq file associated with an album (copy rock.eq to AlbumName.eq) while still being able to have a track-specific override if you really want to peak to 'White Rabbit'.
  • I do not see the need for having a .cfg EQ setting (hardware), an album-art-style .eq EQ setting ("Back In Black"), AND another .eq EQ setting. (Rock)
    
    • Perhaps a second example of where this would be useful could sway me (as if I have any say), but I agree with KindOfBlues71 on this, and do not see how the system I outlined above conflicts with the usage pattern Chronon describes in his response to KindOfBlues71.

EDIT:  OH - THAT'S RIGHT!  I was arguing semantics - a stupid thing to do!

[Chronon]

Nicely organized.  I think this satisfies all of the concerns so far.  No arguments here.

[Llorean]

If you have a .cfg EQ, an album .eq, and a genre .eq, it means that if you switch headphones to ones that are say have much lower bass, you don't have to reconfigure every single individual EQ to take this into account, you just tweak the hardware EQ and you're done.

Also, many people have genre EQs (why do so many players come with them naturally?) yet Febs suggested there may be badly mastered albums or tracks that need correcting. You could always, I suppose, have the individual album one override the Genre one, but it seems to me ideal to be able to preserve a "flat" version of the album by having an individual album .eq, then if your "rock" tastes change you don't have to redo the overall "Rock" EQ, plus every individual album EQ you've already created for problematic ones.

It just seems having them cumulative the whole way down means "if the situation changes, you need to make the fewest possible changes to your .eq files" so I don't see what the negative of it is in response to that. All I really got was an "I wouldn't use it, so it's not necessary." Personally *I* wouldn't use it, but I still think it's better than having them replace each other, so only one is active. Why even do cumulative to the hardware EQ if you're going to do it like this, why not just say all .eq files must be explicit?

The only apparent disadvantage of cumulative EQ is that it might confuse some people at first. But there's nothing you can accomplish with explicit EQ that you can't with cumulative, while cumulative would allow you much more freedom for adjusting the sound of your whole collection as tastes change or other needs change, while having to alter an absolute minimum of files.