Comparing the IL to the S3

[tigerwillow1]

This post compares a 6mm S3 against a 6mm IL. It is not a review. Most of you focus on image quality, while my priorities are reliable animal detection and reduced false triggers. Within reason, image quality is priority #3.

The subject cameras, in 5442-S2 mounts because I don't like the S3 and IL fixed focal mounts. When the wind is blowing, the dancing tall grass results in frequent false triggers from both of them.




My overall conclusion is that in most (not all) cases the IL image is better, but remember my priorities and judge for yourself.

• The IL is obviously better when faced with a larger range of brightnesses in the same image.
• No info here to judge the built-in LEDs. They are turned off and I use external illuminators.
• I think that the daytime images with the IL are overall a bit better, but the S3 is good enough.
• The IL is better with false triggers from the blowing grass.
• The IL is so far pretty poor at detecting animals, but I want to hold judgement until having more time to experiment with settings.

The PRO I have is reasonably good at detecting animals, and they use the same firmware, so something doesn't make sense.

Reasons to use the IL over the PRO:

• If you want 6mm fixed focal, IL is the only choice.
• It's significantly smaller for fixed-focal. The IL is the same body size as the S3.

Reasons to use the PRO:

• Doesn't use the silly (to me) newer mounting base.
• More LEDs that I assume is better.
• A few more features that might be important to some.

The firmware in the IL is 3.146.000000.4.R, 2025-11-11

Most images are pretty small because the subject is pretty far from the camera. The edge of the house is 75 feet away.

The only daytime shots I have to offer right now:


Shots with Christmas lights trying to ruin the image:







More pictures in the next post.

 

[tigerwillow1]

I didn't catch the exact frame on this one. Shots within one second of each other.

Cat in motion, S3 did better.

 

[Molbo]

So, it looks like, on average, IL static handles static scenes better, but S3 works better for moving objects.
Can you post settings for both of the cameras?
Assuming they have the same sensor(IL and S3), and IL having symptoms similar to PRO, it means that there is some hope for PROs in future firmware updates

 

[tigerwillow1]

Can you post settings for both of the cameras?

Image: Defaults, except for night on the S3 with brightness=60 and gamma=71, for a better overall match to the IL's night image.
Exposure: Daytime auto, night 0-4ms, except the IL likes to keep changing it by itself to auto, so I really don't know what it was for the previous pictures.
FPS: 10
Bitrate: 8192, H264
The IL also likes to turn its LEDs back on by itself.

I caught a couple new daytime shots today. They illustrate the poor animal detection of both cams. The IL never recognized the deer, and the S3 didn't pick it up until it was getting close to exiting the FOV a few seconds later. What I'm seeing is the the AI is very slow to detect an animal, worse on the IL than the S3, and the animal has sometimes come and gone before the AI detects it.

Neither camera was close to detecting the animal at this point:


The S3 detected the animal a few seconds later, the IL never detected it, but it does do better on the darker area of the picture:

 

[natedawg]

Where did you get your IL if you don’t mind me asking? I’m getting cameras soon and honestly the image quality here looks pretty good!

 

[Molbo]

Image: Defaults, except for night on the S3 with brightness=60 and gamma=71, for a better overall match to the IL's night image.
Exposure: Daytime auto, night 0-4ms, except the IL likes to keep changing it by itself to auto, so I really don't know what it was for the previous pictures.
FPS: 10
Bitrate: 8192, H264

Is it possible to disable AI-ISP on IL?

 

[tigerwillow1]

Where did you get your IL if you don’t mind me asking? I’m getting cameras soon and honestly the image quality here looks pretty good!

Got it from Andy. Special order took about a month.

Is it possible to disable AI-ISP on IL?

I've been looking at the IL, PRO and S3 today. I thought I saw an on-off switch for AI-ISP on either the IL or PRO but can't find it now. On the S3 there's a switch for AI-SSA and I might have confused that. Where would I look for an AI-ISP switch?

 

[Molbo]

I've been looking at the IL, PRO and S3 today. I thought I saw an on-off switch for AI-ISP on either the IL or PRO but can't find it now. On the S3 there's a switch for AI-SSA and I might have confused that. Where would I look for an AI-ISP switch?

This is from PRO, there is no AI-ISP in S3.

 

[bigredfish]

It’s only in the very latest December FW

It made litter no difference in the 54PRO or 3449 TIOC PRO. I would expect same.

Unless it’s a different sensor, which I’m told it’s not, there is not going to be much difference.
Again unless a different sensor, the image quality with respect to moving targets can’t be fixed with FW.

 

[tigerwillow1]

PRO with 2025-08-07 firmware ----------------------------------------------IL with 2025-11-11 firmware No AI-ISP switch on either one.



Regarding the AI-ISP switch, if it does indeed do nothing, there's precedent. In the 5442/541 series, the object filter on/off switch never did anything.

 

[tigerwillow1]

Frame-by-frame shots of a cat running full blast through the S3 and IL FOV, plus bonus from the T54PRO that picks it up right after. At 10 FPS, the whole sequence lasts less than a second. The cat was 100% undetected by all 3 cameras. They cannot pick up fast moving animals. I knew about it only because I saw it with my eyes. The PRO shots aren't a good comparison because the cat was closer to it, and at a different angle. If there's an overall quality difference here between the S3 and the IL, I'm not seeing it.

 

[Holbs]

Do you live next to a zoo?

 

[tigerwillow1]

Do you live next to a zoo?

No, but sometimes it feels like it. Deer, coyotes, cats, bobcats, dogs, snakes, possums, skunks, foxes, squirrels, badgers, rats, porcupines, just off the top of my head. A few years ago there was an escaped cow at our back door. No bears, though, and none of the snakes are poisonous.

 

[tigerwillow1]

By now I have a pretty good feel for image quality between the S3 and IL, so probably my last mass-posting of side-by-side images. Conclusions are:

• In situations with even daytime lighting, it's an overall tie.
• In situations with large difference bright and dim areas, IL wins.
• At night with IR lighting, IL wins.

I've realized there is a bigger elephant in the room, in that so far the IL's object detection kind of sucks. Details in the next post.

Images where the S3 looks better (IMO of course):





Images where the IL looks better:

 

[bigredfish]

That would be a major deviation from the 54PRO and 3449 TIOC Pro if so.

It would be nice to get hands on one to be able to test real motion video with each camera tuned to its best ability

My S3's dont produce that noise. Weird

 

[tigerwillow1]

The IL has so far turned out to be horrible at object detection, to the point of being close to useless if I don't figure out how to fix it. My PRO does not have this non-detection problem, interesting because they use the same firmware. The IL does have a newer firmware version than the PRO. These are shoot-from-the-hip numbers, not precise measurements.

Of all the animals detected by either camera:

• The S3 detects 90%
• The IL detects 50%

(In a few cases, the IL detected and the S3 didn't)

Of the statistically small number of human encounters:

• The S3 detects 100%
• The IL detects 50%

The only thing the IL does better is far fewer false triggers.

I will try to work through this. Right now both cameras use tripwires exclusively. The S3 rule sensitivity is 4 and global sensitivity is 6. Both the rule and global sensitivities on the IL are set to 8.

Things I will try over time:

• Turn off large scale models.
• Intrusion box instead of tripwire.
• Different firmware version.

Anything else to try? Am I missing something obvious?

 

[natedawg]

How are stills with movement at night compared to the S3? Either animals or humans moving or cars driving by?

All of the threads on the 54Pro and TIOC Pro have all had pretty bad picture quality with moving objects at night.

 

[bigredfish]

The S3 has so far turned out to be horrible at object detection, to the point of being close to useless if I don't figure out how to fix it. My PRO does not have this non-detection problem, interesting because they use the same firmware. The IL does have a newer firmware version than the PRO. These are shoot-from-the-hip numbers, not precise measurements.

Of all the animals detected by either camera:

• The S3 detects 90%
• The IL detects 50%

(In a few cases, the IL detected and the S3 didn't)

Of the statistically small number of human encounters:

• The S3 detects 100%
• The IL detects 50%

The only thing the IL does better is far fewer false triggers.

I will try to work through this. Right now both cameras use tripwires exclusively. The S3 rule sensitivity is 4 and global sensitivity is 6. Both the rule and global sensitivities on the IL are set to 8.

Things I will try over time:

• Turn off large scale models.
• Intrusion box instead of tripwire.
• Different firmware version.

Anything else to try? Am I missing something obvious?

Did you type that wrong?

You start saying the S3 is bad then show numbers saying it detects Better?

 

[bigredfish]

And the more I look at your pics, though they're not good full res, I would swear you have the labels/names backwards

The 5442 S3 (Aka 54IR) does not look anywhere near that bad unless you have some really bad settings?

 

[wittaj]

Yeah not being full resolution distorts the imperfections.

I mean this is from an older 5442 and the S3 would be even better