Advanced search

Message boards : News : different video segment lengths

1 · 2 · 3 · Next
Author Message
Travis Desell
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 16 Jan 12
Posts: 1813
Combined Credit: 23,514,257
DNA@Home: 293,563
SubsetSum@Home: 349,212
Wildlife@Home: 22,871,482
Wildlife@Home Watched: 212,926s
Wildlife@Home Events: 51
Climate Tweets: 21
Images Observed: 774

              
Message 1508 - Posted: 4 Sep 2013, 0:09:35 UTC

Susan and I had a discussion today and we're interested in testing out different video segment lengths. I'd like to try 5 minute, 10 minute and 20 minute segments.

The idea would be to use these different segment lengths and then compare if there is any effect on accuracy of your observations. Does anyone have any issue with longer video lengths? With this change you'd get a mix of 3, 5, 10 and 20 minute videos (as we wouldn't be changing the ones that have already been converted into 3 minute segments).

Profile Lemon
Send message
Joined: 10 May 13
Posts: 229
Combined Credit: 476,659
DNA@Home: 190,781
SubsetSum@Home: 225,957
Wildlife@Home: 59,921
Wildlife@Home Watched: 11,190,214s
Wildlife@Home Events: 0
Climate Tweets: 0
Images Observed: 0

        
Message 1509 - Posted: 4 Sep 2013, 0:19:38 UTC - in response to Message 1508.

Would this change affect all viewing areas? If so, keep in mind you already have over 1000 hours of unviewed video for both Belden and Blaisdell. It could take weeks before anyone even gets a longer video for those areas.

Travis Desell
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 16 Jan 12
Posts: 1813
Combined Credit: 23,514,257
DNA@Home: 293,563
SubsetSum@Home: 349,212
Wildlife@Home: 22,871,482
Wildlife@Home Watched: 212,926s
Wildlife@Home Events: 51
Climate Tweets: 21
Images Observed: 774

              
Message 1510 - Posted: 4 Sep 2013, 0:37:26 UTC - in response to Message 1509.

Would this change affect all viewing areas? If so, keep in mind you already have over 1000 hours of unviewed video for both Belden and Blaisdell. It could take weeks before anyone even gets a longer video for those areas.


Yes it would. But videos are shown somewhat randomly, so videos will randomly be longer (but not very frequently until most of the 3 minute videos are watched).

Profile DoctorNow
Avatar
Send message
Joined: 31 Jan 12
Posts: 328
Combined Credit: 2,333,912
DNA@Home: 94,617
SubsetSum@Home: 162,940
Wildlife@Home: 2,076,355
Wildlife@Home Watched: 1,977,485s
Wildlife@Home Events: 589
Climate Tweets: 128
Images Observed: 2,678

              
Message 1513 - Posted: 4 Sep 2013, 5:44:30 UTC - in response to Message 1508.

I'd like to try 5 minute, 10 minute and 20 minute segments.

The idea would be to use these different segment lengths and then compare if there is any effect on accuracy of your observations. Does anyone have any issue with longer video lengths? With this change you'd get a mix of 3, 5, 10 and 20 minute videos (as we wouldn't be changing the ones that have already been converted into 3 minute segments).

Well, as I already mentioned somewhere I personally would have no problem with 5 minutes - maybe even 10, but that would be the absolute maximum.
But I doubt that it does effect the accuracy much. Can't explain it in detail, it's only a feeling... ;-)
____________
Wildlife & SubsetSum Badges-creator ;-)
Mod/Guru @ the german team BOINC@Heidelberg
My BOINC-Stats / My Badges

Travis Desell
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 16 Jan 12
Posts: 1813
Combined Credit: 23,514,257
DNA@Home: 293,563
SubsetSum@Home: 349,212
Wildlife@Home: 22,871,482
Wildlife@Home Watched: 212,926s
Wildlife@Home Events: 51
Climate Tweets: 21
Images Observed: 774

              
Message 1517 - Posted: 4 Sep 2013, 17:56:57 UTC - in response to Message 1513.

I'd like to try 5 minute, 10 minute and 20 minute segments.

The idea would be to use these different segment lengths and then compare if there is any effect on accuracy of your observations. Does anyone have any issue with longer video lengths? With this change you'd get a mix of 3, 5, 10 and 20 minute videos (as we wouldn't be changing the ones that have already been converted into 3 minute segments).

Well, as I already mentioned somewhere I personally would have no problem with 5 minutes - maybe even 10, but that would be the absolute maximum.
But I doubt that it does effect the accuracy much. Can't explain it in detail, it's only a feeling... ;-)


Any particular reason 20 minutes would be bad, other than bandwidth? (However at 5 or 10 minutes it would take the same bandwidth just in more clicks).

Profile DoctorNow
Avatar
Send message
Joined: 31 Jan 12
Posts: 328
Combined Credit: 2,333,912
DNA@Home: 94,617
SubsetSum@Home: 162,940
Wildlife@Home: 2,076,355
Wildlife@Home Watched: 1,977,485s
Wildlife@Home Events: 589
Climate Tweets: 128
Images Observed: 2,678

              
Message 1522 - Posted: 4 Sep 2013, 18:31:11 UTC - in response to Message 1517.
Last modified: 4 Sep 2013, 18:34:01 UTC

Any particular reason 20 minutes would be bad

Because it's already exceeding the limit I can divert from my work breaks etc... ;-P

Edit:
But since no one else posted his opinion I guess you are going for it anyway...
____________
Wildlife & SubsetSum Badges-creator ;-)
Mod/Guru @ the german team BOINC@Heidelberg
My BOINC-Stats / My Badges

Profile Lemon
Send message
Joined: 10 May 13
Posts: 229
Combined Credit: 476,659
DNA@Home: 190,781
SubsetSum@Home: 225,957
Wildlife@Home: 59,921
Wildlife@Home Watched: 11,190,214s
Wildlife@Home Events: 0
Climate Tweets: 0
Images Observed: 0

        
Message 1527 - Posted: 4 Sep 2013, 22:52:19 UTC - in response to Message 1522.

How about limiting the varying lengths to just 2 of the viewing areas, such as Lostwood and Terns (or Plovers)? That would still allow for accuracy testing while providing a "safe haven" for those who don't want longer video lengths.

Senilix
Send message
Joined: 2 Feb 12
Posts: 96
Combined Credit: 8,073,607
DNA@Home: 106,112
SubsetSum@Home: 114,729
Wildlife@Home: 7,852,766
Wildlife@Home Watched: 7,118,967s
Wildlife@Home Events: 1,563
Climate Tweets: 5
Images Observed: 26

            
Message 1528 - Posted: 4 Sep 2013, 22:53:40 UTC - in response to Message 1508.

I'd say let's give it a try. IMO I'd not expect any big improvement to the accuracy numbers.

I guess Travis can always add some checkboxes to the settings page to let users opt in/out longer video segments if they want to.

GLeeM
Send message
Joined: 1 Jul 13
Posts: 118
Combined Credit: 47,541,025
DNA@Home: 28,994
SubsetSum@Home: 231,079
Wildlife@Home: 47,280,952
Wildlife@Home Watched: 3,888,714s
Wildlife@Home Events: 628
Climate Tweets: 0
Images Observed: 0

          
Message 1529 - Posted: 5 Sep 2013, 0:38:10 UTC

The original idea was to try one or two or more nests separately and in chronological order.

The advantage would be:
1. The watcher could get to know one nest really well.
2. The watcher would know when the bird leaves and only have to watch for when the bird returns.

Imagine not having to watch a three minute video multiple times for a bird movement when that bird is not there!!!!!! This would not improve accuracy or speed up watching?? To me this would be HUGE, I wonder what I am doing wrong?

Profile DoctorNow
Avatar
Send message
Joined: 31 Jan 12
Posts: 328
Combined Credit: 2,333,912
DNA@Home: 94,617
SubsetSum@Home: 162,940
Wildlife@Home: 2,076,355
Wildlife@Home Watched: 1,977,485s
Wildlife@Home Events: 589
Climate Tweets: 128
Images Observed: 2,678

              
Message 1532 - Posted: 5 Sep 2013, 6:03:30 UTC - in response to Message 1528.

I guess Travis can always add some checkboxes to the settings page to let users opt in/out longer video segments if they want to.

That would be a good idea, I can live with longer videos if that is implemented. ;-)
Otherwise I simply skip them.
____________
Wildlife & SubsetSum Badges-creator ;-)
Mod/Guru @ the german team BOINC@Heidelberg
My BOINC-Stats / My Badges

GLeeM
Send message
Joined: 1 Jul 13
Posts: 118
Combined Credit: 47,541,025
DNA@Home: 28,994
SubsetSum@Home: 231,079
Wildlife@Home: 47,280,952
Wildlife@Home Watched: 3,888,714s
Wildlife@Home Events: 628
Climate Tweets: 0
Images Observed: 0

          
Message 1533 - Posted: 5 Sep 2013, 10:52:33 UTC - in response to Message 1531.

I scratch my head in amazement at what this person saw that I didn't that would indicate a Bird is Present. I can watch that same Video numerous times & never once see anything that would lead me to believe a Bird is there.

It is not what the person saw, it is what the person KNOWS.

Example:
If you SEE a bird leave at time 1:00 and you don't see it return (with lots of movement) during time 1:03 then the bird is NOT there.

If you SEE a bird return at time 1:06 and you don't see it leave during time 1:09 then the bird is still present even if you see no movement.

GLeeM
Send message
Joined: 1 Jul 13
Posts: 118
Combined Credit: 47,541,025
DNA@Home: 28,994
SubsetSum@Home: 231,079
Wildlife@Home: 47,280,952
Wildlife@Home Watched: 3,888,714s
Wildlife@Home Events: 628
Climate Tweets: 0
Images Observed: 0

          
Message 1535 - Posted: 5 Sep 2013, 14:33:53 UTC - in response to Message 1534.
Last modified: 5 Sep 2013, 14:38:58 UTC

skip the rest.

After the validation process gets fixed it will again validate Video Issue and Unsure will not be Invalid. Until then please do not skip the rest as new people will get frustrated and quit. Or they will take your example and only turn in the absolutely clear videos ... pretty soon there will only be the skipped ones left :-(

Also, if you skip the rest your Accuracy Rating will have no meaning!

Edit: On the dark videos mark Unsure Parent Absent/Present and No for everything else.

GLeeM
Send message
Joined: 1 Jul 13
Posts: 118
Combined Credit: 47,541,025
DNA@Home: 28,994
SubsetSum@Home: 231,079
Wildlife@Home: 47,280,952
Wildlife@Home Watched: 3,888,714s
Wildlife@Home Events: 628
Climate Tweets: 0
Images Observed: 0

          
Message 1543 - Posted: 6 Sep 2013, 11:07:59 UTC - in response to Message 1539.

Well my feelings are if your just Marking Unsure a lot your Accuracy Rating has no meaning either, your just passing the Buck on down to someone else to make the Decision for you & insuring yourself a Valid Video when you Mark Unsure.


At first my feelings were the same as yours until I thought about it a little further.

They started to show accuracy rating because it is important to them. This says to me that I should not guess at an answer.
If I am writing a software program to examine video for whether a bird is present and I use a segment of video in which you say a bird is present and my program doesn't see it, then I will make a change to the algorithm to get the program to see the bird.
If you guessed wrong that the bird is there then I am making changes for nothing. I would have examined the video myself if you had marked it Unsure. This is why there is an Unsure button. It is not passing the Buck, it is doing what they want.

I think my logic is correct, if not, someone please speak up!

Travis Desell
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 16 Jan 12
Posts: 1813
Combined Credit: 23,514,257
DNA@Home: 293,563
SubsetSum@Home: 349,212
Wildlife@Home: 22,871,482
Wildlife@Home Watched: 212,926s
Wildlife@Home Events: 51
Climate Tweets: 21
Images Observed: 774

              
Message 1581 - Posted: 10 Sep 2013, 3:36:49 UTC - in response to Message 1534.

I scratch my head in amazement at what this person saw that I didn't that would indicate a Bird is Present. I can watch that same Video numerous times & never once see anything that would lead me to believe a Bird is there.

It is not what the person saw, it is what the person KNOWS.

Example:
If you SEE a bird leave at time 1:00 and you don't see it return (with lots of movement) during time 1:03 then the bird is NOT there.

If you SEE a bird return at time 1:06 and you don't see it leave during time 1:09 then the bird is still present even if you see no movement.


I'm not talking about something that is as Obvious as seeing the Bird leave & not returning or returning during the same Video, I'm talking about seeing nothing happening in a very Dark Video, no movement, no Bird leaving or returning, nada, nothing but some waving Grass with a dark back ground.


In these, I'll use the scroll bar to quickly scan through. If it really is too dark to tell if anything is happening then you can click the video issue button (which with new updates to the validator will be the same as if you had just marked everything unsure).


On another note I see some of the newer people are starting to mark those type of Video's that have a lot of waving grass & darkness as Video Issue's, I've been marked Invalid on 3 of them because I just marked Unsure for Bird Present or Absent and 2 others Marked Video Issue when it was just to Grassy & Dark to tell anything. I reported them to see what Susan or whom ever looks at them has to say about it.


If it's too dark to see anything, then video issue is the way to go. If it's not so dark you could tell if the parent left or came back, then do partial unsures and no for the ones you could have seen.

Until I get an answer from whom ever Reviews the Reported Video's I guess with the Validation system in place now the best thing to do is only turn in Videos that are absolutely clear that the Bird is Present or Not & skip the rest. Marking unsure now for Bird Present or Not with some of the Darker Video's will get you Invalids with the present Validation System.


I'll get this updated. Might need tom implement some kind of partial accuracy along with the partial credit. :P If you get 6/8ths the markings right, then I guess that shouldn't be just a wrong in terms of accuracy which right now is just calculated as (right / (right + wrong)).

So instead of doing accuracy on a per video basis, I could do it on a per observation basis.

1 · 2 · 3 · Next
Post to thread

Message boards : News : different video segment lengths