Reply to thread

Message

<blockquote data-quote="rsbrux" data-source="post: 850801" data-attributes="member: 49539"><a href="https://forum.team-mediaportal.com/members/hwahrmann.10391/" target="_blank">hwahrmann</a>,Many thanks for the new version.&nbsp; It's amazing how much functionality you have packed into it!In reviewing the settings, I noticed some minor spelling errors in the English version:&quot;Ripping&quot; is spelled correctly in the heading &quot;Ripping/Conversion&quot;, but is spelled &quot;Riping&quot; 3x at the bottom of the &quot;General&quot; section.I was scouring the settings in hope of finding a way to increase the size of the music sample used to &quot;Identify File&quot;s as I have been getting quite a few &quot;false positive&quot; IDs.&nbsp; Attached is a particularly humorous example.&nbsp; The track in question is actually &quot;It's a Man's World&quot; performed by James Brown.&nbsp; It is particularly funny that the albums offered are mostly plausible, despite the incorrect identification of the track.&nbsp; What would a track by &quot;Los Bravos&quot; be doing on &quot;The Very Best of James Brown&quot;?&nbsp; Would it be possible to implement some kind of &quot;plausibility check&quot; in MpTagThat to prevent such self-contradictory results, or are you depending strictly on the information provided by the MusicBrainz API?&nbsp; I can imagine that such a plausibility check, even if possible, might be non-trivial, but would it at least be possible to have the option of using a larger sample to generate the hash?&nbsp; I would be willing to wait longer for the results, if it reduces the number of erroneous matches.</blockquote>

[QUOTE="rsbrux, post: 850801, member: 49539"] [URL='https://forum.team-mediaportal.com/members/hwahrmann.10391/']hwahrmann[/URL], Many thanks for the new version. It's amazing how much functionality you have packed into it! In reviewing the settings, I noticed some minor spelling errors in the English version: [INDENT=1]"Ripping" is spelled correctly in the heading "Ripping/Conversion", but is spelled "Riping" 3x at the bottom of the "General" section.[/INDENT] I was scouring the settings in hope of finding a way to increase the size of the music sample used to "Identify File"s as I have been getting quite a few "false positive" IDs. Attached is a particularly humorous example. The track in question is actually "It's a Man's World" performed by James Brown. It is particularly funny that the albums offered are mostly plausible, despite the incorrect identification of the track. What would a track by "Los Bravos" be doing on "The Very Best of James Brown"? Would it be possible to implement some kind of "plausibility check" in MpTagThat to prevent such self-contradictory results, or are you depending strictly on the information provided by the MusicBrainz API? I can imagine that such a plausibility check, even if possible, might be non-trivial, but would it at least be possible to have the option of using a larger sample to generate the hash? I would be willing to wait longer for the results, if it reduces the number of erroneous matches. [/QUOTE]

Verification