This post describes my ideas for what (I call) DSP-Per-Track. I believe this would fairly easy to implement, and very easy, predictable and logical to use.
1. The purpose of DSP-per-track is to customize certain tracks which do not otherwise conform with your listening environment. For example, insert some B52 songs in with some Daft Punk, and you hear a huge difference in studio and mixing style. Some might say Daft Punk is too bassy others might say that the B52's music is very bright, hard, and on some systems almost irritating. But they actually both sound pretty decent compared to other songs in each's genre. So just as one example, DSP-Per-Track to the rescue.
2. DSP-Per-Track is completely separate from the regular DSP options which are unique to each play zone.
3. DSP-Per-Track only uses DSP options that are not global (intended for system use, like Headphones, Room Correction or Convolution). Non global effects which logically CAN be used per track include Equalizer, Parametric Equalizer, Effects, Tempo & Pitch, and Parametric Equalizer 2. Any user VST plugin loaded as a DSP may also be used (as they can also for play-zone global settings, but are still separate in function between global and per-track.
4. DSP-Per-Track is Library based, not Playlist based, so you are essentially tweaking your music library by song, regardless of where or how it is played. That provides the most consistent behavior as well as ease of programming or adjustment.
4a. The Library based notion of DSP-Per-Track may be arguable by some, but having it by playlist means that the same song played in a different list would/could have different characteristics. I suppose in select conditions that might be useful, but ultimately would be a whole bunch of additional busywork by the user, not to mention more memory required to house song parameters for duplicates in playlists. (but see 5a for checkbox options. One could be to make the same for all playlists). I don't know how problematic this would be to program, though.
5. DSP-Per-Track would turn itself off after each track. The idea is that this isn't a global repair of every tune in your library, but occasional corrections on an as needed basic. Whole library changes should be done globally if deemed necessary. Additionally, not turning it off on a per song basic would lead to unexpected results depending on sort orders.
6. I suppose you could add a parameter to the name assignment of the DSP-Per-Track name, which indicated to NOT disable it and the end of the track. But I'm not sure that the current naming of files for each DSP assignment is even necessary under this scheme. Without separate filenames, It would be a simple right-click option from a library file entry, and that would show you the current DSP assignments for that particular song. Additionally, you could leave that window open (much like playlists or tags) so it would show you immediately when a title is selected, just what its settings are (if any). Extra unique options might be included in checkboxes for that filename. Some sort of a visible tag (color?) could display in the file list to indicate any song that has DSP-Per-Track enabled.
6a. You could select multiple tracks and set the DSP-Per-Track assignments for all of them at once. You could also copy/paste from one title to another to duplicate DSP.
USES / EXAMPLES
Simply and easily 'normalize' various sonic characters of titles that just don't fit with each other when played together, whether it be eq, polarity, even mild compression when needed. It would obviously be most usable when the sources are high quality, but could also be used to help blend mp3 or other lossy formats in with higher quality source.
Concise visualization of what tracks are or are not set for DSP, and quick changes can be easily made.
You could make up some nice playlists with tracks that fit nicely together, to save as a playlist, burn to CD, etc.
One of my uses would be when streaming live tunes to an Internet Radio Streaming station, especially tracks 1970 or before to get them more uniform in tonality and level (L128 doesn't work well for this) so that later processing will actually work. This type of thing operates in real time and the music might span 4 decades or more, needing lots of help.
One professional use that would be helpful to me is for track mixing post production of a song, to modify it in various ways, and be able to compare them against each other or inserted in a playlist of other music to see how they stand up sonically.
There's lot's of applications for this sort of thing, but only if it is easy to work with, simple to operate, and completely predictable -- not subject to arcane limitations and behaviors that you're always trying to work around.
Anyway, those are my thoughts on the whole issue. Comments? (thought out, please)
--Bill
