screwed up my installation a little

[kolbasz]

so, one thing I did not account for was IR flair (if that's the technical term). Installed during the day and then night came. Things are still decent but curious how much I can do to make it better. Also, because it is my first cameral, maybe I am overreacting.

I tried shifting down and over slightly, but then I lose the end of the driveway and it does not get a ton better. At the end of the day, I assume since I am still capturing the important stuff that maybe I do not need to worry. Important stuff being the walkway to the door.

Just curious for some opinions as I venture into this new space.

 

[sebastiantombs]

That's why you see all the "test, don't guess" comments. The right side is the real problem. Maybe a little jury rigged "baffle" on the IR emitter area might help.

 

[kolbasz]

That's why you see all the "test, don't guess" comments. The right side is the real problem. Maybe a little jury rigged "baffle" on the IR emitter area might help.

Curious, if the wall wasn't there, would it change the overall image? Discussing with the wife, because it is in the doorway she says it is where she wants it. Plus, now there is a hole in the wall, so there is not a ton of moving.

If the image itself of the yard wouldn't change, I won't fret it. If in fact the image would be way different without that, then maybe we explore the idea of the baffle.

 

[samplenhold]

That's a nice pic. But it really tells nothing about the cam's performance. No one really cares about the bushes. You need to see what it does to a person walking up to your door and away from your door. That is the real test.

One comment is that this cam is probably way too high to get a face shot of anyone at the door.

 

[wombsganz]

Since the area of interest is still being lit by IR the issue seems to be that the camera under-exposes the area of interest since the IR flair around the edge is bringing up the average brightness of the scene. I have been meaning to test something in this scenario. If you set the shutter as slow as possible without introducing blur it might help (say 1/120 or 1/60 depending on your sensor size). Of course the best option is to reposition the camera but this might work in a pinch.

On second viewing of your pictures the exposure actually isn't too bad. I would recommend to go outside and walk to the door at night both at normal pace and brisk pace. See if you have any issues with blur. It's easy for any camera to capture a good still picture but it's the moving stuff that counts when you want to identify an intruder.

 

[kolbasz]

That's a nice pic. But it really tells nothing about the cam's performance. No one really cares about the bushes. You need to see what it does to a person walking up to your door and away from your door. That is the real test.

One comment is that this cam is probably way too high to get a face shot of anyone at the door.

I did under 7' as suggested in the wiki. Perhaps it's deceiving as the walkway is a step down so the walkway itself would be just over 7` and the doorway area under.

 

[kolbasz]

Since the area of interest is still being lit by IR the issue seems to be that the camera under-exposes the area of interest since the IR flair around the edge is bringing up the average brightness of the scene. I have been meaning to test something in this scenario. If you set the shutter as slow as possible without introducing blur it might help (say 1/120 or 1/60 depending on your sensor size). Of course the best option is to reposition the camera but this might work in a pinch.

On second viewing of your pictures the exposure actually isn't too bad. I would recommend to go outside and walk to the door at night both at normal pace and brisk pace. See if you have any issues with blur. It's easy for any camera to capture a good still picture but it's the moving stuff that counts when you want to identify an intruder.

Good idea, I'll give this a shot. I guess there's a bit more to the testing than reviewing the image on the computer while holding and testing positions in the middle of the day.

 

[samplenhold]

I guess there's a bit more to the testing than reviewing the image on the computer while holding and testing positions in the middle of the day.

Yes. Since most shit happens late at night, the real test is at night and with a person moving around in the field of view. That is what you are trying to capture, not the bushes.

 

[kolbasz]

Yes. Since most shit happens late at night, the real test is at night and with a person moving around in the field of view. That is what you are trying to capture, not the bushes.

agreed.

So I went out and tweaked it a bit for the better, I think. So before, I had it faced more to the right to capture this lower corner onto the entryway. The thinking was that if an intruder crawled up there I would miss them. Well, what freaking robber is james bonding to the door. So we shifted and now most flair is gone and it raises a new question. How is it best to fix positioning on the door. With the new position you see way more, but still the lower left has a side light, am I capturing enough or should I tilt more down. Surely it is opinion at the end of the day, but still I think there is valued input here that can coach to what is proper door coverage.

 

[wombsganz]

I think this positioning is good all things considered. You'll get enough advanced warning if you set up motion sensing on the outer skirt of your driveway and there's a good chance an intruder will be approaching the camera dead-on which reduces blur and aids identification (capturing people moving laterally is more difficult due to blur). And you'll be able to see them trying to jimmy the door. Or more realistically in this case smashing the glass.

For now I would suggest to tweak the shutter and gain to get the clearest picture for things in motion at night, as well as to test that any alerts (or whatever they are called in your NVR) are delivered to you quickly. For example, walk to the front door at 11pm and see how long it takes your phone to start buzzing if you have it set up that way. That way you can quantify how much time you have to react after you are notified.

Beyond that, start planning out other areas that you might want to cover with cameras or consider ancilliary upgrades - motion sensing floodlights for example. Consider buying a varifocal (zoom) camera and test it in the same spot - see if a more telephoto lens (zoomed in lens) allows you to get a cleaner shot of someone's face. Also consider some basic home security upgrades. If that is a glass door I hope you have a solid core wood door behind it. Have you reinforced your door strikers and door frame? Will it take a beating like this? A reason to add one more camera

I am making some assumptions about why you are running these cameras. If you are only trying to catch package thieves these kind of measures may not be necessary.

 

[samplenhold]

So we shifted and now most flair is gone and it raises a new question. How is it best to fix positioning on the door. With the new position you see way more, but still the lower left has a side light, am I capturing enough or should I tilt more down. Surely it is opinion at the end of the day, but still I think there is valued input here that can coach to what is proper door coverage.

Without a person moving in the view, nothing can be said about the flair, positioning, etc. This is the third time I have said this and I am done giving help here.

 

[alastairstevenson]

How is it best to fix positioning on the door.

I'm guessing from the wide field of view that the camera has a 2.8mm lens.
I believe you could improve the usefulness of the image a lot with a camera having a smaller field of view, or even with a varifocal.
This would also increase the level of detail on any 'objects of interest'.

 

[handinpalm]

I would try rotating the cam 90 degrees and adjust video setting in cam GUI to correct, assuming your camera allows this. That would give you a better view of sidewalk/drive. You still want to see the bushes, because that is where someone's face will be when walking up. You did not state which cam you have?

 

[kolbasz]

Without a person moving in the view, nothing can be said about the flair, positioning, etc. This is the third time I have said this and I am done giving help here.

I went out there and walked around. Watching on my phone I was able to see decent, but my recorder was crashing so I couldn't get saved footage. Separate issue I need to solve.

I heard your suggestion loud and clear which is why I went out there. Once I capture some footage, I will share it with you to get your opinion.

 

[kolbasz]

I would try rotating the cam 90 degrees and adjust video setting in cam GUI to correct, assuming your camera allows this. That would give you a better view of sidewalk/drive. You still want to see the bushes, because that is where someone's face will be when walking up. You did not state which cam you have?

I have this guy, IPC-T5442T-ZE, soI believe it supports flipping the image, as I saw it in the settings.

 

[Fastb]

Don't overlook your other options:
- Turn off internal IR lighting. Mount an external IR illuminator somewhere nearby. Like on the house, shining onto ft walk.

What IR illuminator would be best?

Here is another IR light for $60 (although shipping is quite high) which will do wonder to those Hiks or any cam with crappy night vision http://s.click.aliexpress.com/e/l3bfELW Atom posted sample images comparing with Longse here...

ipcamtalk.com

Choosing an IR illuminator

Is having too much IR illumination as bad as too little? For the purposes of nighttime facial ID would having too much IR give a washed out or overexposed face and therefore no detail? I really don't need more than 50 feet of IR illumination.

ipcamtalk.com

- Set cam to "corridor mode" as handinpalm suggested. The narrowed horizontal FOV will reduce the reflection from the white wall.

- extra light (Ir or otherwise) on front walk and driveway. This should make the cam reduce its IR level, since it doesn't need to lighten up those dark areas.

- lens change

Good luck.

 

[kolbasz]

I'm guessing from the wide field of view that the camera has a 2.8mm lens.
I believe you could improve the usefulness of the image a lot with a camera having a smaller field of view, or even with a varifocal.
This would also increase the level of detail on any 'objects of interest'.

I bought the IPC-T5442T-ZE for its Varifocal capabilities and my venture to get into cameras around the house for surveillance.

 

[alastairstevenson]

Ok, so you should be able to zoom in a bit so the view doesn't include irrelevant areas that are adversely affecting the image.

 

[looney2ns]

Testing before mounting, every location, as outlined in the Cliff Notes.
Means, making a temporary mount, with bucket, 2x4x8', rocks for ballast. Mount cam on 2x4, test location for a minimum of 24hrs, preferably longer while recording.
Worrying about watching the driveway with this cam, falls into the newbie category of trying to do too much with one camera.

 

[looney2ns]

When wives decide the suitable location, it almost always is a detriment to getting an optimal placement for face ID.
Paint the camera to match the house, 98% of visitors will never notice it's there.