FiiO X5 3rd Gen (FiiO X5iii): Stream of Consciousness Impressions & Review

Discussion in 'IEMs and Portable Gear' started by Torq, Mar 21, 2017.

  1. econaut

    econaut Almost "Made"

    Contributor
    Joined:
    Feb 5, 2017
    Likes Received:
    266
    Trophy Points:
    63
    Location:
    EU
    Unfortunately, I didn't. I had these at the back of my mind, but didn't stumble across them in the settings, just found the Viper stuff. If there's a next time, I will definitely check this out.


    Don't think so. Actually I have tinnitus (those damn metal concerts), so I am quite used to it ;) I listened to some minimalistic vocal-centric stuff and the hiss spoiled it for me on the Fiio.
     
  2. Johan

    Johan New

    Banned
    Joined:
    Feb 23, 2017
    Likes Received:
    12
    Trophy Points:
    8
    Location:
    Resident troll and general Porker
    Are guys using se or balanced with andro? Balanced seems better as the oi is nearly 3 ohms so it would give the andro a more desirable fr then SE. Likely less hiss too
     
  3. gixxerwimp

    gixxerwimp Professional tricycle rider

    Pyrate
    Joined:
    Sep 30, 2015
    Likes Received:
    5,773
    Trophy Points:
    113
    Location:
    small island claimed by China
    It was a bit noisy in the store and I didn't notice any hiss, though I wasn't listening for it in particular. The 2 clicks when switching tracks was very obvious though.

    Fiio X3ii spec says:
    Output Impedance <0.2 Ω

    So if that's your reference for Andromeda bass, then the Zout of the X5iii's SE output would likely give the same bottom end response.
    Pretty much all my listening was with the Slow Roll-off filter, and I thought the tonality of cymbals on Rush's YYZ sounded a bit more realistic from the X5iii than my micro iDSD (which was recently described by @k4rstar star as totally off-timbre and fucked in the treble). I very quickly flicked through the other filter settings at the end of my session and couldn't hear any significant differences. Guess my ears suck. :oops:
    I think everyone who's tried the Andros with the X5iii has used the SE output. I'm waiting for the store to get a balanced cable in so I can try it out for exactly this reason.
     
    Last edited: Apr 8, 2017
  4. Johan

    Johan New

    Banned
    Joined:
    Feb 23, 2017
    Likes Received:
    12
    Trophy Points:
    8
    Location:
    Resident troll and general Porker
    Dont quote me as i javnt tried but going by the specs on fiio website im gonna assume the balanced out will be much better with the andros,
     
  5. Bina

    Bina MOT - Shanling

    Pyrate
    Joined:
    Sep 29, 2015
    Likes Received:
    536
    Trophy Points:
    93
    On unit I had on loan, balanced output was significantly more hissing and actually worse sounding than the SE. Tried with same identical cable on Hum Pristine.
     
  6. Johan

    Johan New

    Banned
    Joined:
    Feb 23, 2017
    Likes Received:
    12
    Trophy Points:
    8
    Location:
    Resident troll and general Porker
    from the andro?

    hmmm well that blows

    its odd, on the fiio site the balanced out actually has LESS power then the SE and WORSE S/N ratio only better spec is crosstalk (cause it is balanced obviously)
     
  7. Grahad2

    Grahad2 Red eyes from too much anime

    Pyrate
    Joined:
    Jan 26, 2017
    Likes Received:
    1,162
    Trophy Points:
    113
    Location:
    Singapore
    Most amps have worse S/N ratio on balanced out since that's 4 channels of amplification vs 2.
     
  8. Johan

    Johan New

    Banned
    Joined:
    Feb 23, 2017
    Likes Received:
    12
    Trophy Points:
    8
    Location:
    Resident troll and general Porker
    depends on the topology, bridged then yes

    still less power is really odd
     
  9. rawrster

    rawrster New

    Joined:
    Mar 19, 2017
    Likes Received:
    19
    Trophy Points:
    8
    Location:
    New York City
    That is definitely disappointing. I may think twice about a balanced cable for the Andromeda now.

    I did find the balanced out having less power a bit odd but it isn't unheard of. I had a Taurus a whole back and the balanced out had less power until a certain ohm limit. That was a headphone amp and not a dap though.
     
  10. Johan

    Johan New

    Banned
    Joined:
    Feb 23, 2017
    Likes Received:
    12
    Trophy Points:
    8
    Location:
    Resident troll and general Porker
    Damn and i was hoping fiio eould turn it around with this dap. Every other product ive ever tried has sounded like total garbage save the original x3 with the wolfson, im still cutious about this though. Maybe ill grab one off of amazon so i can rrturn it if its garbage without issue.

    Im currently using an lgv20 as my portable faily driver and I have to say im rathrr impressed by it. This thing can actually drive my hd800S to beyond comfortable levels and is dead silent eith iems. Thinking i may just stick with it tbh.
     
  11. Grahad2

    Grahad2 Red eyes from too much anime

    Pyrate
    Joined:
    Jan 26, 2017
    Likes Received:
    1,162
    Trophy Points:
    113
    Location:
    Singapore
    The original X3 was rubbish though, but I found the X5iii good, but didn't hear it with the Andromedas.
     
  12. Johan

    Johan New

    Banned
    Joined:
    Feb 23, 2017
    Likes Received:
    12
    Trophy Points:
    8
    Location:
    Resident troll and general Porker
    I liked the original x3. Grn 2 was garbage imo
     
  13. Torq

    Torq MOT: Headphone.com

    Pyrate
    Joined:
    Jan 15, 2016
    Likes Received:
    8,193
    Trophy Points:
    113
    Location:
    Seattle, WA
    On the little tool ... I'm down to just packaging (and testing those packages). We'll see how well that goes ... bleedin' edge versions of Python tend to break packaging tools. If it's really problematic I may move it back to 3.5.x. (or make it harder for everyone and just post the source and let you get on with it).

    Anyway, a little teaser ... which I'll move to the tool-thread when I post the usable version:
    Code:
    Usage: mlcp [OPTIONS] [LIBRARY_PATH] [BACKUP_PATH]
    
      Purge or backup crud files from a specified music library.
    
      "Crud" files are any file that isn't one of the types designated to keep.
      By default this will delete any non-music file (see --listtypes),
      including folder-level album art.
    
      Unless the --purge option is specified, no actual changes to the library
      will occur.  This allows you to simulate the process to see what will
      actually occur before actioning them.
    
    Options:
      -p, --purge                   Executes the actual purge operation.
      -a, --art [keep|remove]       How to handle folder-level art.
      -k, --keep [music|audio|all]  File classes to keep (see --listtypes).
      -b, --backup                  Backup crud files instead of deleting.
      -v, --verbosity [0|1|2|3]     Status/progress reporting verbosity.
      -l, --listtypes               List file types by class.
      --version                     Show the version and exit.
      --help                        Show this message and exit.
    
    In it's simplest form, the command:
    Code:
    mlcp --purge
    Will run through the library starting in the current working directory, deleting anything it considers to be a crud file. This mode is the most aggressive, and you can determine what that means by running:
    Code:
    mlcp --listtypes
    That'll give you a list of two classes of files ... "Music" files, which are any file that would normally contain MUSIC. And then a second set that usually contains AUDIO, but that typically isn't a "music" file/song/track:
    Code:
    Classes of files for -keep option (default is just to retain "Music" files):
    
      Music File Types:
    
        aac, aiff, ape, dff, dsd, dsf, dxd, flac, m4a, m4p, mp3, oga, ogg, wav, wma, wmv
    
      Audio File Types:
    
        3gp, aa, aax, act, amr, au, awb, dct, dss, dvf, gsm, iklax, ivs, m4b, mmf, mpc, msv, mogg, opus, ra, rm, raw, sln, tta, vox, wmv, wv, webm
    In essence, the simple operation DELETES ALL files in your library that are NOT one of the types lists as "Music File Types".

    The "--keep" option defaults to "music", but you can also specify "all" or just "audio" and it'll keep either all "Music" and "Audio" files, or just "Audio" files (which would then result in it deleting the "Music" files).

    For album art, if you keep that at the folder level, you can have it leave those files alone like so:
    Code:
    mlcp --purge --art keep
    A file is considered to be "folder-level album art" if it is a .JPG, .JPEG or .PNG file named as "album", "cover", "folder", "thumb" or "albumartsmall". Anything not matching those predicates will get zapped.

    If you want the tool to backup the files it would otherwise delete, specific the --backup option:
    Code:
    mlcp --purge --backup . /backup_folder 
    And that will result in the "crud" files being copied to an identical hierarchy under the specified "/backup_folder". That allows you to put them back in place simply by copying the entire hierarchy to the original root library folder.

    And, finally, if you don't want the tool to DO anything to your files at all and just want to see what would happen if you ran it for real, then omit the "--purge" option and it'll do a dry-run without moving or deleting anything. Specify "--verbosity 3" as well and it'll spit out a file-by-file list of what will be moved or purged (at "--verbosity 2" it'll show a progress bar and estimate time remaining as it goes).

    Otherwise, a sample run will yield something like this:
    Code:
    Scanning library ...
    
    Total File/Folder Count: 15,295
    
    Music Files Found:
    
      dff: 5
      dsf: 83
     flac: 8,899
      m4a: 1,857
      m4p: 13
      mp3: 1,272
    
    Total Library Size: 12,129 music files, occupying 343.1 GB.
    
    Will Purge: 1,938 crud files, saving 0.46 GB (0.13%).
    The above numbers are generated from a test-subset of my travel library for my WM1A. It runs for me in about 10 seconds or so, and another 10-ish if I use the "--backup" option. Performance is roughly linear with increasing library size.

    --

    The mismatch in total file/folder count with the combined music and crud files is down to me deliberately counting folders as well so that the numbers agree with a) what your OS reports if you do a Get Info/Properties lookup on the library and what your DAP should report with the actual files.
     
    Last edited: Apr 8, 2017
  14. Torq

    Torq MOT: Headphone.com

    Pyrate
    Joined:
    Jan 15, 2016
    Likes Received:
    8,193
    Trophy Points:
    113
    Location:
    Seattle, WA
    As I figured, packaging with 3.6.1 is being a pain in the arse. I've spent more time twatting around with that than I did writing the code in the first place, so I'm done with that for now. I'll update it once the relevant tools are updated. That means this first version (0.0.0.1) will require you to install Python 3.6.1. It also means the binary is bigger than it would otherwise be. Proper version numbers will follow with a proper package and more testing.

    And, finally, it'll likely also require at least a properly POSIX compliant OS, and will possibly only work correctly/or at all on macOS.

    It isn't doing anything macOS/OS X/POSIX specific, in fact the only platform specific code in the entire thing is the stuff to extract volume geometry information under Windows, but my plan to thoroughly test it on macOS, Linux and Windows fell by the wayside with the time spent on the packaging nonsense.

    Do NOT run this on your live library without first having a verified backup - at least until you're sure about what you are (and it is) doing*.

    And if you do run it, I'd recommend making the first pass something like:

    Code:
    mlcp --verbosity 3 /library_path
    From there, you can experiment with other options (see previous post) until it shows as doing the things you want it to do BEFORE you add the --purge switch.

    Note that, at least on macOS, any files this tool purges should wind up in your trash can and can be recovered from there.

    Anyway, to run it you'll need to download and unzip the file, put the resultant "mlcp" file wherever you want it to live (somewhere on your path if you're not just going to run it directly in the library folder) and you'll probably need to mark it as executable**.

    If you encounter issues, PM them to me. Once I have the proper thread for this (i.e. when it's more broadly useable), THEN issues can be posted in that thread - but let's not clutter this one up any further (if anything serious crops up I'll pull the file). Don't PM me without details though ... "it didn't work" isn't useful.

    ...

    Things I plan to add, beyond fixing the packaging:
    • Right now folder-level album-art is an "all or nothing" thing. It either all gets zapped, or it all gets kept. I will add an option to allow you to purge all-but one folder-level album art file, to designate which one is kept (e.g. cover.jpg vs. folder.jpg) and, possibly, to rename the files to a consistent form (.e.g to make a variety of different files all read as "folder.jpg" ... with the necessary conversion if they're in a different format).
    • If purging results in empty directories, which it can (if you happen to have a folder with JUST crud in it for some reason) I'll make it so that those are removed from the library. Currently, they'll get left in place.
    ...

    Removing this version of the file, posting a new version (below), see that post for details.

    --

    *If you do, and something goes wrong, I don't want to hear about it - you have been warned.
    **Not giving explicit instructions for these steps right now in the interests of discouraging anyone that doesn't know what they're doing.
     
    Last edited: Apr 11, 2017
  15. Archaleas

    Archaleas New

    Joined:
    Apr 11, 2017
    Likes Received:
    0
    Trophy Points:
    1
    Location:
    Australia
    Hey Guys,

    So have been using the Fiio X5iii for a few days now, and seem to be encountering a few bugs,

    The biggest and most frustrating one at the moment is that it is not playing certain songs. I am only 4 levels deep in a folder structure but when I tap the song, turns orange, nothing loads.

    Frustrating me no end, anyone else encounter this?

    EDIT: Updated firmware and seemed to fix the problem! Hopefully this will help anyone who has similar issues
     
    Last edited: Apr 11, 2017
  16. Torq

    Torq MOT: Headphone.com

    Pyrate
    Joined:
    Jan 15, 2016
    Likes Received:
    8,193
    Trophy Points:
    113
    Location:
    Seattle, WA
    There are lots of bugs with it. Most are minor. Some not. I've not seen any issues with deeply nested folder structures, but there are circumstances in which some songs won't play. And in all the ones I've encountered, that's been down to either a) the format/bit-rate of the track not being supported or b) there being something wrong with the file.

    So, what firmware version are you running?

    Is the song on internal storage or a microSD card? If it's a card, what format is it in.

    And, most importantly what format and bit-rate is/are the track(s) that are causing you problems?

    You're probably better off posting issues on the Head-Fi thread for this as the FiiO rep(s) actually read that and tend to be reasonably responsive.
     
  17. Torq

    Torq MOT: Headphone.com

    Pyrate
    Joined:
    Jan 15, 2016
    Likes Received:
    8,193
    Trophy Points:
    113
    Location:
    Seattle, WA
    I found a minor issue (not destructive) with the way resource-forks were being handled for hidden/O.S. maintenance files on card storage. In essence, the resource-forks show up as separate files during the scan - when the file they are associated with is deleted the O.S. removes the resource-fork as well ... causing a file-not-found error when the tool goes to delete the fork itself.

    This update eliminates those forks from the purge processing and thereby completely avoids the issue. This will have the effect of yielding correct-but-lower crud-file counts when compared to the previous version.

    Also, when operating directly on SD cards, it's worth noting that depending on your card reader, the speed of your card, and how your drive caching is setup, the tool can indicate that it's finished processing before disk activity stops. Don't pull the card until your OS lets you safely unmount it! With a big FAT32 format card and thousands of files, combined with a slow card/reader it can take several minutes for the cached OS operations to catch up with reality. In most cases you won't see this issue, but I found it when I happened to stick a card in an ancient card reader and figured I should mention it. In a future version I'll see about linking completion with the OS file activity catching up.
     

    Attached Files:

    • mlcp.zip
      File size:
      836.2 KB
      Views:
      226
  18. landroni

    landroni Friend

    Pyrate
    Joined:
    Aug 12, 2016
    Likes Received:
    1,164
    Trophy Points:
    93
    Would it help throwing in a `sync` call there to ensure that file activity has completed?
     
  19. Torq

    Torq MOT: Headphone.com

    Pyrate
    Joined:
    Jan 15, 2016
    Likes Received:
    8,193
    Trophy Points:
    113
    Location:
    Seattle, WA
    If os.sync() worked 100% of the time, it would - but it doesn't ... this being one of the cases where you see it not do what it is supposed to (since it's being used here already). If you run exFat or NTFS file-systems on your storage cards (FAT32 is glacial on large volumes, and is the biggest cause of the issue), use a non-ancient reader (in this case one that is comparatively slow, but too smart for it's own good), or disable write-caching for card storage, you'll almost certainly never run into the issue. And it's a complete non-issue on fixed storage.
     
  20. Bobcat

    Bobcat Friend

    Pyrate
    Joined:
    Oct 3, 2015
    Likes Received:
    400
    Trophy Points:
    63
    Location:
    Texas
    Home Page:
    Man that takes me back to the old days.
    sync
    sync
    sync
    shutdown -h now

    After all, if one sync is good; three damn well ought to be better :)

    Rob
     

Share This Page