They can appear quite cryptic, so hopefully these explanations will make sense and allow you to configure the app for your particular needs. One thing worth pointing out on that subject is that there is no one "perfect configuration"; every single scene is different, so you have to experiment with each setting to find the right balance for you.
All the settings can be found in the app under Settings > Advanced:
Detections For Trigger
This number represents the number of frames in a row that must contain motion in order for a motion trigger to be fired, and therefore perform a capture. So if you want to detect motion that is more pronounced, set this to a higher value so you can be sure some genuine and interesting motion occurred. For example, if you set this to 10 then each of 10 successive frames from the camera input must contain motion in order for a picture (or video) to be taken.
Processing Scale
Each frame from the camera contains lots of pixels to process, so it is wise to scale this down in order to speed up processing. So lowering this value makes the frame to be processed more pixelated (you can see this on the Mini-Preview within the app as the camera is running) and thus means less processing time/power. It also means the frame contains less information so you risk missing the more subtle motions in the scene (although this of course might be a good thing for your needs). So if you want to process faster and become less accurate, set this to a lower value.
Sensitivity
How sensitive to motion the app is. Can't say more than that really!
Frame Delay
The gap between the frames that actually get processed. So rather than processing every available frame, only process the available frame every certain amount of milliseconds. Similar to Processing Scale, this means less processing time/power needed to run the app. Of course it also means that you can catch more prolonged motion to cause a trigger. For example, you could set this to be 1000ms and Detections For Trigger to be 10 which would mean only motion that has occurred for roughly 10 seconds (10 x 1000ms) would take a photo (or video).
Gaussian
This smooths out the current frame to be processed in order to reduce noise in the picture. It means slightly more processing time/power is required though, which is the trade off for using this setting.
So combine these settings to match the needs of your scene. However, there is no combination of settings that I know of that can ignore the motion of trees - the app can't distinguish between what type of object is actually moving, so there is no "Ignore Trees" setting I am afraid... all you can do is try adjusting these settings to try and lessen the amount of times moving trees to cause a trigger. A moving tree is motion after all!
Awesome app, and thx for the "Help" post... I was wondering how these settings work... I'll play around with them. Still hoping to solve the following:
ReplyDelete- Squirrel and butterfly triggered events
- A ~1 second delay from the time a car enters frame to the time the video starts recording... hoping to get this down to 1/2sec.
- Any way to auto-delete old recordings (like 7/14/30 days or something)... I'm afraid my card is going to fill up fast and my phone is mounted outside the house ~ 15' off the ground... hoping to set it and forget it.
hiergehts weiter zu Spycam kaufen
ReplyDeleteWould love to know how to store to sd card automatically rather than after the event. Just like it use to.
ReplyDeleteFirst you need to figure out the proprietary location path that the sd card is mounted. Linux is still in old days where external devices are mounted in some random directory, rather than at a root level as in reality, above the file system-which is on only one of many drives. Then, under advance setting, toward bottom, you manually type in the name of the external storage card to some directory that you create. So, for me it is "storage/sdcard1/spy". Spy is a directory I created. Sdcard1 is my external card, since linux starts counting at 0, unlike real people. And, 0 is the internal user area (partition) that the phone allows.
DeleteI have tried to change it with no success.
DeleteUsing a Note 4 on M (6.0.1) The storage location for me should be
/storage/3093-9544/spy
but when I try to enter it it says
"Cannot create or access directory. Check storage not full or protected!"
I know the path is correct (the /3093-9544 folder is where all my ext sdcard files exist)
Any idea what may be wrong?
What is the exact difference between the video resolutions? Which one produces the clearest and sharpest video? Is the resolutions "high" referring to the maximum resolution of the device? (Everything notwithstanding the file size)
ReplyDeleteGreetings
More pixels mean more resolution.
DeleteAlso, it should be discussed the advantage of black and white v. color. In my first tests, I was saving %7 by doing black and white. For my use as a job site time clock/supervisor/antitheft, uploading via cell, the lowest resolution will save cell data and battery life (using 7800 mah, and jumper box usb). The fisheye lens dropped the typical oiutdoor shot file size from 50k to 32k. The 7% was indoor b&w saving, I haven't tested outside, nor yet time to reproduce these results on any consistent basis.
Outside test just done: 43k v. 33 k Color v. bw.
DeleteSo saving %25 cell upload and some number up to %25 on battery savings
This is a great way to ditch the time punch card on jobs when the boss is away-using an old cell phone.
ReplyDeleteHowever, this may mean the need for every 3 minute captures and only 1 dropbox upload every 20 minutes over cell data (sync rest on wifi). Still looks like I might get 80 megs per 40 hours. Also, there is a huge need for specific dropbox folders per phone. So, phones don't overwrite eachother and get shuffled photos. This, could be averted by opening up a separate Dropbox account, I suppose.
Doesn't look like it wide upload to dropbox when recordong. The ideal is every twenty when recprdong so I can see the job. But can use every minute. The work around until a fix is ftp and dyndns client. But looks like the dyndns clients don't work on cell towers. Hmmm. Really need this function, at least one dropbpx jpg up per twenty while record every one to three.
ReplyDeleteDoesn't look like it wide upload to dropbox when recordong. The ideal is every twenty when recprdong so I can see the job. But can use every minute. The work around until a fix is ftp and dyndns client. But looks like the dyndns clients don't work on cell towers. Hmmm. Really need this function, at least one dropbpx jpg up per twenty while record every one to three.
ReplyDeleteI am taking this back. Looks like it is uploading in real time.
DeleteThe trick to delete multiple dropbox file is shift arrow. If you just delete the folder, then you must unlink and relink the securet app.
Yes, the real time upload every minute is very useful for breakins where they take the phone.
DeleteLooking closer at the scaling via a large full screen preview, it is clear that its sole purpose is to demonstrate motion detection. The 1/8 is clearly blockier than 1/2, might not work for fisheye camera or motion at a distance--though this setting obviously is for cpu/power during motion detection, not for the captured image.
ReplyDeleteI would like Paul, or someone, the meaning on the previews red target and entending matrix. How do I use these cues to see if my settings will trigger motion, via the preview? It seems they pop onscreen even in a still picture.
Also, I have spent much time trying to get widest view. Laying the Samsung phone on side seems to work better to get wider view. I did make the mistake of gluing the securing mounts inside the clear tupperware case, which make the video upsidedown. I will need to tear everything apart and start again, unless there is a software fix somewhere. Tips on lens to maximize width, would be great. I ordered some fisheye and wide clip on lenses on amazon. But have yet to test and mount these, or adjust the settings to compensate (assume I will have to go back to 1/2 scaling).
ReplyDeleteWhy is the email setting under resolution and not in same place as dropbox?
ReplyDeleteAlso, I would like an email on first trigger in xx minutes, not every motion and capture.
Is there any way to rotate the photo? I've just bought the app for a front door camera and it's working well but would be great if it would save the right way up. Am I doing something wrong?
ReplyDeleteThanks, Duncan
The camera can only be rotated 90 degree clockwise. So, if your cord only allows 90 degrees counter clockwise, everything will be upsidedown. You can turn your comupter screen upsidedown to view captures. Or get a batch rotator. Or stand on your head when viewing. Or wear upsidedown glasses. Or put phone/tablet on table, gently rotate screen. Or hope Paul gets around to photo rotation. Not a game stopper for me. I would rather see a webp format support, so my 90k files can be 40 to 50 k with more clarity. Also, cool down periods to make the phone more stable. Also, autostart in photo mode, or app restart, replace instance, every 3 hours.
DeleteThanks for the details on trigger settings, but I still get many false alarms. If you also included an area mask filter, perhaps a 5x5 grid overlay that user could toggle what to include in trigger processing you could avoid the tree, and increase performance since area may only be a couple squares.
ReplyDeleteThis program isn't on android marketplace anymore it seems.
ReplyDeleteWill there ever be 4K support?
ReplyDelete