<br><br>
<div><span class="gmail_quote">On Apr 11, 2005 2:24 PM, <b class="gmail_sendername">Matt</b> <<a href="mailto:skd5aner@gmail.com">skd5aner@gmail.com</a>> wrote:</span>
<blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">Hello All, =D<br><br>Sorry this is long: For the short version, just read the first two paragraphs.<br><br> I had an idea a few months ago, and just thought I'd throw it out<br>there. I think that mythcommflag is getting better and better.<br>However, sometimes there are times it'll goof up, and some shows are<br>notroriously harder than others to flag correctly by their unique<br>filmographic nature(Lost, Law & Order, CSI, etc...).<br><br>How about a method where users can validate mythcommflag's results and<br>make that info available to the myth community to strengthen the<br>accuracy of commercial flagging by using verified data.<br><br>For example. You record an epsiode of "Lost". You let mythcommflag<br>run, and it detects commercials 90% correct. Then you go in and edit<br>the recording and import the commercial flags by pressing "Z"... so<br>far, everything like normal. Once the commercial markings are in<br>there, you edit them so they are correct (cut off the begining<br>commercials before the start of the show, make sure commercials are<br>detected correctly, and make sure that there aren't any improperly<br>detected cuts in the middle of the show, etc).<br><br>Then, once you have explicitly defined the commercials, some process<br>could identify these as " Human Verified to be Correct". Then, the<br>next time a user runs mythcommflag on that episode, the service can<br>verify it's results based on those verified results and adjust<br>accordingly.<br><br>This is where it gets difficult and several different solutions might<br>be available:<br><br>1) Digital fingerprints - Not sure if you anyone is familiar with<br>MusicBrainz (<a href="http://www.musicbrainz.org/">http://www.musicbrainz.org/</a>) or not. What it does is<br>take an mp3 file, analysis it and make a digital finger print of it.<br>Then, it sends that fingerprint to a database to compare with other<br>fingerprints that people have submitted. It can then make a<br>comparison and correctly identify the file and populate the ID3 tags<br>of an otherwise un-identified file. The more people that ID a file,<br>the "smarter" it becomes and the better it can do at identifying<br>files. Once the program has been correctly marked, and some kind of<br>"fingerprint" is made from the show, then it could be uploaded to a<br>central database where other's can look for that show and episode to<br>compare mythcommflag results so that the mythcommflag can do a better<br>job based on the verified results sent by users who manually made sure<br>the commercial detection was correct.<br><br>2) Time based - Possibly more difficult, however you have to assume<br>that the time of the show between commercial breaks is going to be<br>(nearly) identical. So, in almost the same way, if mythcommflag has a<br>false positive for a commercial, it can review what others have<br>verified as correct and say "hey, there shouldn't be a commercial here<br>because 40 other users say there is a time block of 15 minutes, not<br>7.5 + commericial + 7.5 more, etc.<br><br>3) Identify individual commercials (not the entire commercial break) -<br>Along the same lines as #1, identify specific commercials. So,<br>anytime an acme company commercial is recorded, mythcommflag<br>could/would be able to identify it as a commercial soley because it's<br>"learned" that it is. This could possibly have more overhead and<br>possibly labor intensive. With the constant turnover of advertising,<br>as well as locally broadcast commercials, it might not be worth it and<br>it'd be a ton of information. However, it could see where this might<br>be useful for short teaser clips you see at the begining or the end of<br>certain shows or that certain networks frequently use.<br><br>4) combination of the above - I think if you went with the<br>fingerprint, a time based solution would almost be included in that.<br><br>5) Other - I'm sure others would have other suggestions that might<br>work for this.<br><br>I know that this is a centralized approach, but seeings how it would<br>be what I call a "bonus" feature, people that don't want to<br>participate wouldn't have to sort of thing. Perhaps their's a<br>decentralized approach as well. Mythcommflag could still work WITHOUT<br>this feature, it's just something that can reinforce the accuracy of<br>commercial flagging in myth. I'm also not sure about the legalities,<br>PC, etc that a methodology like this would have or Issac's stance on<br>the issue as I don't know if it's ever been thought of up to the<br>point.<br><br>Anyway, I'm not a developer, nor am I pleading or begging anyone to<br>develop this... I just think it sounds cool and if anyone has the<br>know-how or if a few developers think it'd be fun to work on it, I'd<br>love to help out any way I can.<br><br>So, what do all of you think? I know it could be a large project...<br><br>Thanks!<br>Matt<br>_______________________________________________<br>mythtv-users mailing list<br><a href="mailto:mythtv-users@mythtv.org">mythtv-users@mythtv.org</a><br><a href="http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users">http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users</a><br></blockquote></div>
<div><br>I've also been thinking about this sort of feature. It seems a little sad, that hundreds of CPU's start comm. flagging the same show minutes after it's ended. What about a task sharing system to aid commercial flagging?</div>
<div> </div>
<div>Let me describe:</div>
<div>My machine and your machine have recorded "Lost" and "Alias". We both subscribe to the same cable service in the same zip code area (basically, we have the same Zap2It settings).</div>
<div>My machine will flag "Lost". Your's will flag "Alias". Then we'll exchange cut points. Initially, the exchanged cut-points could be verified by local commercial flagging. Over-time, groups of machines would form that "trusted" each others' comm. flagging accuracy based on previously verified (and accurate) results.</div>
<div> </div>
<div>Notice that no copyrighted content is exchanged.</div>
<div> </div>
<div>Obviously, more elaborate schemes could be devised so that the load would be fairly distributed and "leaches" would be kept to a minimum.</div>
<div> </div>
<div>Just an idea.</div>
<div> </div>
<div>-- </div>
<div>Christopher David Petersen</div>
<div>(not of nuvexport fame, just a disciple of Jarod's guide)</div>
<div> </div>