SMD 3.0 dahua

cd36

Pulling my weight
Joined
Oct 31, 2019
Messages
118
Reaction score
112
Location
Canada
@cd36 - wasn't sure how you determined when to take the pics, but your method sounds good.
Sorry I should have been more clear in my initial post on when I took the screen grabs. But yes, all screen grabs are from the first time the camera picked me up when walking into the scene. It should be essentially what you asked for, just without showing the IVS rule as I used "Visitor Picture" from SmartPSS to save it from the recording, which of course strips out the IVS rule information.
 

looney2ns

IPCT Contributor
Joined
Sep 25, 2016
Messages
15,521
Reaction score
22,657
Location
Evansville, In. USA
@cd36 - any chance you have SD cards in those cameras and could do the same test with IVS rules and then show pics with the IVS rules on showing the first time the box turned red?

Many of us experienced significant lag with the 3.0 firmware, but your setup of two cameras basically side-by-side would be a great test to show those differences.
Lag in what way?
 

Wildcat_1

Known around here
Joined
Dec 9, 2018
Messages
2,046
Reaction score
5,861
Location
US
Also on the AI detection (SMD + IVS issues etc) Just one thing to remember is that 05-06 was primarily focused on SmartIR improvements (unfortunately as I’ve shown yesterday that still has huge issues). The detection issues were carried over from the 04-03 release so believe Dahua will focus on SmartIR issue fix first then AI detection + IVS for next release as they move toward SMD 3.0 stable release etc. Will be interesting to see what the next test release brings and will put it back through testing when available.
 

wittaj

IPCT Contributor
Joined
Apr 28, 2019
Messages
24,429
Reaction score
47,550
Location
USA
Lag in what way?
In my testing, with SMD, the recognition of the subject is delayed compared to previously. I suspect that is in part due to additional analysis being performed. It still recorded on every event, but in many instances, the vehicle would have moved all the way across the screen and be half a car from outside of the field of view all together before the vehicle would get the AI box around it (so actually a half a box). Previously it would draw the box around the whole car when it was in the center of the frame.

I do not know if this update also impacted IVS AI, but it did in mine. Similar things happened as with SMD. Previously, the camera would trigger on the leading edge of the box around a person. Now the person is completely through the IVS before it would trigger. It still got them and triggered, but was delayed.

This impacts if you use it as a spotter cam for a PTZ. The jogger would be almost out of the field of view before the IVS triggered well beyond the jogger crossing the line right in the middle of the field of view. This impacted the spotter cam in that by the time the camera recognized it as a person and sent the command to BI for an ONVIF trigger and then the PTZ would move, the jogger was out of the frame. This did not happen previously.
 

looney2ns

IPCT Contributor
Joined
Sep 25, 2016
Messages
15,521
Reaction score
22,657
Location
Evansville, In. USA
In my testing, with SMD, the recognition of the subject is delayed compared to previously. I suspect that is in part due to additional analysis being performed. It still recorded on every event, but in many instances, the vehicle would have moved all the way across the screen and be half a car from outside of the field of view all together before the vehicle would get the AI box around it (so actually a half a box). Previously it would draw the box around the whole car when it was in the center of the frame.

I do not know if this update also impacted IVS AI, but it did in mine. Similar things happened as with SMD. Previously, the camera would trigger on the leading edge of the box around a person. Now the person is completely through the IVS before it would trigger. It still got them and triggered, but was delayed.

This impacts if you use it as a spotter cam for a PTZ. The jogger would be almost out of the field of view before the IVS triggered well beyond the jogger crossing the line right in the middle of the field of view. This impacted the spotter cam in that by the time the camera recognized it as a person and sent the command to BI for an ONVIF trigger and then the PTZ would move, the jogger was out of the frame. This did not happen previously.
Yeah I'm seeing some of the same with the intrusion and tripwires.
Dahua engineer is logged remotely into my cam now to see what they can find.
 

EMPIRETECANDY

IPCT Vendor
Joined
Nov 8, 2016
Messages
8,259
Reaction score
23,717
Location
HONGKONG
Has some progress under @looney2ns great help. He helped make a remote link for engineer checking.
About @Wildcat_1 2 bugs reported.
Bug 1 issue looks like this 2019 cam can’t do real hard reset/factory default to clear the history setting. they do some clear via a special tool, wait looney to check if the smart Ir can active now.
Bug 2. Will have a new firmware tomorrow, hope it can really work out. Bug 2 if sorted the cam will have a good improvement.

IVS issue will have another fw baseline will be released in June. Will have some improvement then.
 

Wildcat_1

Known around here
Joined
Dec 9, 2018
Messages
2,046
Reaction score
5,861
Location
US
Has some progress under @looney2ns great help. He helped make a remote link for engineer checking.
About @Wildcat_1 2 bugs reported.
Bug 1 issue looks like this 2019 cam can’t do real hard reset/factory default to clear the history setting. they do some clear via a special tool, wait looney to check if the smart Ir can active now.
Bug 2. Will have a new firmware tomorrow, hope it can really work out. Bug 2 if sorted the cam will have a good improvement.

IVS issue will have another fw baseline will be released in June. Will have some improvement then.
Thanks @EMPIRETECANDY. Bug # 2 is a big one and speaks to the way SmartIR was coded therefore fixing this, the priority it takes in assessing the scene (including it's own production of light, IR) + how it interacts with Backlight Programs when used together will make a huge difference, not just to the Volt line but across other cam ranges in the portfolio. Look forward to putting the next beta release through its paces in the same test and will report back.
 

Wildcat_1

Known around here
Joined
Dec 9, 2018
Messages
2,046
Reaction score
5,861
Location
US
All

Happy to update that I’ve been testing the Dahua beta release (21-05-20) focused on the bugs I reported and that in my updated testing the issues have now been fixed for the most part. I’ve included a video showing how SmartIR now works correctly after an upgrade and how it works when utilizing with HLC at various levels. All of this testing was completed using the same configuration as before.

Now onto the details. In looking at how this was fixed it appears that the issues were tackled by hardcoding behaviors in this release. This is certainly one way to approach it and does fix as you will see. However this does mean that lower backlight levels do not adjust as well and that at this is not adjusted algorithmically that other lighting conditions ‘may’ cause SmartIR issues. Either way, this is a great update and improvement over not only 05-06 but previous generations of FW as it finally looks to fix (for Volt chipset cams right now anyway) the SmartIR issue I raised (Bug #2) that is deep rooted in FW and widespread across all SmartIR cams in the portfolio of products.

1 key note is that in the new test version (now 21-05-21) there are also adjustments made to the backlight menu in general, namely that you can no longer enable BLC with SmartIR. This does make sense as the 2 functions really have always worked against each other. BLC looks to amplify light on scene, SmartIR looks to adjust accordingly and there are situations where they just don’t work well leading to either washed out (over exposed) or under exposed targets with properly lit backgrounds. This too could be tackled further in future updates with the algorithm approach I mention above to create a truly ‘Smart’ and adaptive solution, but for now at least this approach removes washout being an issue with these competing functions. So do keep this in mind IF you use BLC and again I hope in future that the SmartIR algorithm approach can be looked at.

Lastly and most importantly, I really do hope that Dahua now uses this fix as the base for ALL SMARTIR FW across all cameras going forward. I cannot stress the importance of this enough. This Bug #2 I found and reported has been a long time issue, is NOT limited to Volt chipsets and rather than reinvent the wheel on future FW OR having this SmartIR issue present on other cams, it would be great to know that this becomes (for now) the gold standard of SmartIR base code in current and future releases for ANY Dahua camera.

HTH, please see the video for performance and as always let me know with any questions

My thanks to Dahua for working with me quickly on these bug fixes and to @EMPIRETECANDY for his continued commitment and helping push this forward.

Video Link - Timestamps Included In Video Description

PLease note for some reason the HLC 5 section did not encode so you can skip that part :)
 

tech101

Known around here
Joined
Mar 30, 2015
Messages
1,472
Reaction score
2,125
Location
SF BayArea, USA
Wow thank you so much @Wildcat_1 for working on the fix with Dahua and Andy. Seems like you are going to make Dahua products so much better and they will work the way it was suppose to just not on paper but in real like in this instance for Smart IR.

I am hoping this firmware will also come to Laser PTZ. I see on the laser IR when a human is at distance smart IR cannot take advantage of IR right now like the Beam Angle specially it can definitely do focus beam but it does not and at night at a distance Laser has to be controlled manually to get any useful footage specially for human facial details.
 

cd36

Pulling my weight
Joined
Oct 31, 2019
Messages
118
Reaction score
112
Location
Canada
It would still be very handy for those of us that use VMS software for motion detection.
You can request it from Andy if you want it. The only advantage of this firmware over the last release version is letting you use a backlight program and smart ir at the same time in your situation.

I just think it will be a mistake to post it as a release firmware as some unsuspecting person will install it and have issues. The main benefit of this stream of firmware is the improved smd 3.0,which is still flawed.
 

Wildcat_1

Known around here
Joined
Dec 9, 2018
Messages
2,046
Reaction score
5,861
Location
US
There are still smd/IVs issues with this firmware, I don't think it should be considered ready for release yet. I have it on a tester but haven't deployed it to any cameras in the field.
Yes as I noted above, BLC & SmartIR cannot be triggered together as of this new release. Also this new release is focused on 1 component, the SmartIR related bugs I reported. AI and IVS issues still need to be fixed BUT at least removing these 2 bugs I found will hopefully now allow Dahua to focus on getting SMD 3.0 out the door
 

sebastiantombs

Known around here
Joined
Dec 28, 2019
Messages
11,511
Reaction score
27,690
Location
New Jersey
Given that Smart IR is a fairly important feature, at least there is progress. We can debate the usefulness of this update and whether it should be posted or not, but the bottom line it that it is a major improvement in the Smart IR department and just about everyone uses IR in one form or another for night vision which is why I suggest it be posted.
 

EMPIRETECANDY

IPCT Vendor
Joined
Nov 8, 2016
Messages
8,259
Reaction score
23,717
Location
HONGKONG
thanks a lot to @looney2ns and @Wildcat_1
Later have a double testing at dahua and release it, right now the version still at Testing one, So Monday i can release the official one.
Will have a newer fw at later June, will has some update on the AI function accuracy.

So later we may call these models as S2 soon. :D
 
Last edited:
Top