Advanced VMD - Axis / Mobotix Tested

Author: Ethan Ace, Published on Aug 26, 2013

Surveillance professionals prefer motion based recording over continuous but the main weakness of motion detection is its inaccuracy. Sure, it saves a lot of storage over recording continuously but it still can be very inefficient, triggering against leaves, trees, shadows, low light noise and more, requiring time consuming optimization that rarely works completely.

Axis and Mobotix Advanced Offerings

Recently, both Axis and Mobitix have released advanced video motion detection offerings. Axis has VMD 2.1 that it bills as having 'easy setup with few false alarms' while Mobotix promotes MxActivitySensor as offering 'unique motion detection in the camera'.

The IPVM Test

We tested both of these offerings to understand how much of an improvement they made over 'traditional' video motion detection plus how accurate they were.

To do so, we bought 2 Axis cameras (same model) and loaded one with traditional and one with VMD 2.1, running them simultaneously.

We started with a simple conference room and performed tests with moving subjects as well as running the cameras 24/7 to measure false positives:

We then moved on to a more complex, larger area to see how the offerings handled this:

Finally, we went outdoors to a very large area with vegetation to see how they would fare in an adverse environment:

Finally, we ran the Mobotix camera through similar conditions to gauge its performance.

Questions Answered

Our testing answers the following key questions:

  • How much of an advantage did advanced VMD provide over traditional?
  • Where was advanced VMD better than traditional and where was it worse?
  • Is it worth using advanced VMD? If so, where?
  • What barrier or issues are there in using 'advanced' VMD?

************ ******************* ****** ***** ********* **** **************** **** ******** ** ****** ********* ** *** **********. ****, it ***** * *** ** ******* **** ********* ************ *** ** still *** ** **** ***********, ********** ******* ******, *****, *******, low ***** ***** *** ****, ************* ********* ************ **** ****** ***** **********.

Axis *** ******* ******** *********

********, **** **** *** ******* **** ******** ******** ***** ****** detection *********. **** ****** *.***** ** ***** ** ****** '**** ***** **** *** ***** alarms' ***** ******* ************************** ******** '****** ****** ********* ** *********'.

The **** ****

** ****** **** ** ***** ********* ** ********** *** **** of ** *********** **** **** **** '***********' ***** ****** ********* plus *** ******** **** ****.

** ** **, ** ****** * **** ******* (**** *****) and ****** *** **** *********** *** *** **** *** *.*, running **** **************.

** ******* **** * ****** ********** **** *** ********* ***** with ****** ******** ** **** ** ******* *** ******* **/* to ******* ***** *********:

** **** ***** ** ** * **** *******, ****** **** to *** *** *** ********* ******* ****:

*******, ** **** ******** ** * **** ***** **** **** vegetation ** *** *** **** ***** **** ** ** ******* environment:

*******, ** *** *** ******* ****** ******* ******* ********** ** gauge *** ***********.

Questions ********

*** ******* ******* *** ********* *** *********:

  • *** **** ** ** ********* *** ******** *** ******* **** traditional?
  • ***** *** ******** *** ****** **** *********** *** ***** *** it *****?
  • ** ** ***** ***** ******** ***? ** **, *****?
  • **** ******* ** ****** *** ***** ** ***** '********' ***?

[***************]

Answer *******

  • *********** ********** ***** ***** ****** **** *** ******* ******** *** over ***********. ****** ***** ****** **** ****** **********.
  • *** **** ********** ** ******** *** *** ** *** **** where **** ***** **** *******. *******, *********** *** *** *** more ****** ** ******** **** ******* ** ***** ****** ** the ****.
  • ***** ******** *** ** ***** ** ** ****** **** *** application. ** **** *****, *********** *** ** ****** ******* ****** so ** ***** ** ****** ** *** ************ ** ********* was ********.
  • *** ******* ** *** **** *****. ***** **** **** *** Mobotix ********* *** **** *** *** ** ****** ***** ** most ** ***** *******, *** ******* ** * *******. ******* only ******** ******* (** ********). **** *** *.* ** ************ only ********* ** ***** *** *** *** *** *******.

Key **** ********

***** *** ******* *** ******** **** *** *****:

  • **** ***** ****** ********* *.* ******** ****** ** **** ********** than ***** ******* ****** *********, **** ****** ********* ***** *** ranges, *** **** *** ***** ******. 
  • *******, ** *** ***** ****** (~* *** *** ***** ** our *****) *** *.* ****** **** *********** **** *******. *** false ***** **** ******** **** ***.
  • ****' ******* ****** ********* *** ****** ********* ** ******* ** light, **** ** *******, ****** ************, *******, ***., ******** ***** ******* ** ***** ****** *** to ***** *******. 
  • **** ** *** ******** ****** **** *** ****** ***********, ******** VMD ****** ***** ****** ******* ***** *** *.* *** ***.
  • *** *********** *** ********** ***** ******* * *** * ****** cameras, ****** * ****** ******* ********* ** ***** ***** ******, where * ****** ******* ******* *********** ** *****.
  • *** *.* ******** ****** ** *****, **** * ****** ********* included *** ******** ******** **** **** ** *********** *************.
  • ******** *** ****** **** ******* (**** ******** *** *******), ****** zones *** ********** ** ****** *********** ******, ******* ** *********, making ******** ** ***-****** ***** **** *********. ***** **** ********* object **** *** **** *********** ** **** *** ****.
  • ******* **************** (********* ** * ***) *** * **** **** detection **** ** ** ***** **' ** **** *****, ********** direction ** ******* *** * ***** *******, ** ****. *******, the ******'* **** *** ***** *********** ******** ** ****** ***% missed ********** ** * *** ** *****. 
  • ****************'* ******* ***** *** **** ****** **** *** ****** *** moving ****** *** ***, **** ** ****** * ******* ** over ***', **. ~**' **** ****** ******/**** **** *** ******.
  • **************** ** **** ****** ****** ** ****** *** ********** ****** caused ** ********** *******, **** ** **** ******** *** ******. Activation ** ***** ** *********** ****** ****.

**** ***** ****** *.* ** ********* ** ***** ******** ******** **** ****. **************** ** ******** ******** ** *** ******* ******* ******* firmware ******* *.*.* *** *****.

***************

*** ************* ******* ******, ***** *** **** *** *.* ****** easier ***** *** ****** *********** **** ****' ******* ****** *********, at ** ********** ****. *** ***** ****** ****** *********** ** VMD *.* ****** ****** ********** ***** ******** ****** ********* ********* a *** ** ***** ******. ********** *********(****** + **********) **** ******** ****** ********* ** ****** ** ideal ********.

*** ***** ***** ****' ******** ****** *********, ********** ****** *********** so **** **********, *******, ** *********** ******* *** *** ** the *** ** *** ****** ***** ***********. ***** ****** ** night *** ** ***** ****** *** ****** ********* *** ********* and ******* ******** ** ****** *********. *******, ****** ** *** 2.1 ***** ******* **** ******* *** ************.

**************** ****** **** ***** *** ***** ***********, ********* *********** **********, with **** ***** ***** ********. ***** ****** ******** ****** ******** and ******* ********* ** ******* **** ********* **, *******, ** its ******** ******** ** **** ******* **** ******* ****** ****** *** FOV.

******** **** *** *.* *** **************** **** **** ******** ********** over **** ************'* ******** ****** *********, *** ******* ******* * key ********.

Axis ************* **********

**** ********** ******* ***** ********* *** **** *********** *** *** VMD *.*, ** **** ** *** *********** ** *************, ******* the ******* ******* ** *********** ** ********:

Indoor ********** ****

** *** ****** ********** ****, ** **** *****, ********* ** a ***** ******* *** ******** ** *** ******, **** **' the ******** ****** *****. **** *** *.* *** ******** *** detected *** ******* ***% ** *** ****.

*** *.* *** **** ** ******* **** ***** *********, *******, such ** *** ****** ***** *** ** *** ****** ****** of **** *****. ******** ***, **** ** **** *********** *** 1% ****** ****, ***** ***.


** *** *****, **** *** *.* *** ******** **** ******** in **** *****, ** ****. *** *.* ****** *** ********** out ** ******** *****, ***** ******** *** *** **** ***.

*******, ** *** *****, ******** *** *** ****** ***** ** false ****** **** ******* ** *****, **** ** *******, ****** being ******** ** ** ***, ******** *******, *** ******* ******* as ****** ***** **** *** *******. *** *.* *** ********** immune ** ***** ** **** *****.

Indoor **** ******

****** *** ******* ** ** **** ****** *****, ***** **' long, **** ***** *********** *** ******** *** **** ****, ** to ***** **'.

****** ****, *** *.* ****** **** *** **********, ***** ******** VMD ****** ****** ***.

** *** *****, *** ******* ****** **** *** ***** ********** both, ****** ******** *** *** **** ********, ******* ** **********. VMD *.* ****** ***** *** ** ** ** **** **** (~83% ********).

*******, ****** ****** *** *** **** **** ** ****, *** 2.1 ***** *********. ******** *** *** ****** ***% ******** ** to ~**', *** ****** *** ********** ****** ****.

** *** *** **** ** *** ********** ****, ******** *** once ***** *** * **** ***** ***** ****, ****** ** changes ** *****.

*******

********, ** ********, ****** ****** (* ****** ***** ** ******** alarms ** *** *** *********) ********* ******** ****** ****** ** both ******** *** *** *.*. ** ******** ***** ***** **** both, ** ** **** ****** ** **** ********* *********.

**** ***** ** ***** **' **** **** ******** *******, *** ***** ***' *** ****** ****. **** * ******* walking ******* *** ***, **** ******* *********. *******, *** *.* picked ** *** ******* ** ~**-**' **** **** *** ******, while ******** *** ********* **** ****** **'.

****** ******** ****** ** ****** ******** ****** ****** ** **** from *** ****** **** ~**', **** ******* **** **** ** detect *** ******* ** ** ****** ****** *** ***** ** view ** ************* **'.

*** *.* ******** **** ** *** ****** **** ****** *** day. ******** *** ******** **** ***** *% ** *** ******** on *** **** ****.

*** *.* *** **** **** ** ****** ****** ** *** distant *******, **** ***' **** *** ******.

** *** *****, ******** ***'* *********** *** ******* ** *******, though***** *** ******** *******, ***** **'. *** *.* *** ****** to ****** *** ******* ** **** ***** ** ***, ******* all ***********.

** *****, *** *.* ******** ***** **% ** ********, ***** standard ******** **** ***** **%.

*************, *** *.* *** ****** **** ** ****** ******** ** the ******* ******* ** *****, ********** *** **** ** ***** trucks, *** **** ** ****.

** *** *** ****, ***** ******* ***'* ************* ** ****. VMD *.* ********* ******, ********* ***** **% ** ******, ********* even * ********* ****** ********** *** *****.

Sensitivity ***********

** ******** ****' ******* *** *********** ******** ** *** ** there *** * "***** ****" ** ***** ***** ****** **** reduced, ******* ********* ********* *****. ** ***** **** *** ******* setting, **%, ********* **** ** *** ****** ** ******. ******** sensitivity ** **% ******** ** ******* ********* ***********, ******* *** decrease ** ***** ******. ***** ****, ****** ********** ****** **** common, ****** ******* ** ***** ***** ****** ***** ******. **********, increasing *********** **** **% **** ********** ******** *********, *** ***** motion *** ** ******* ***** *** ********* ****** ********. 

M ****** ***********

** ***** ** **** *********** ** ***** **** ****** ******, we ********* *** *.* ** *** * ****** **** *********, the ***** *** *****. ** ***** **** *********** ** **** light *** ********** ** *** *****, *** ********* ** *** light *** ********, ***** *** * ****** ******** ****** ** image ** *** ***** ******.

Mobotix *************

******* ************* ** ********** *******, ****** ******** *** ** *******'* standards:

Mobotix ******** ******

** ****** *****, **************** ******** *** ******* ** **' ***** 100% ** *** **** ** **** *****. **** *** ******** of ****** ********* (*** ****) *** ******** ***** (****** ****).

******** *** ********* ********** ** ******* *** ******* ***** ****** the *** ** ****** *** ******. ** **** **** ********* began ** ***** **', **** *** ******* *********** *** ******.

** *** *****, *******, *** ** *** ***'* **** ********* performance, ** *** ****** ** ******* *** ******* ** ***. It *** ********** ****** ** ***** ****** **** *******, *******, and ***** ******* ** *****, *******.

******* ***********

********, **************** ********* ********* ** *******, **** ******** ********* *** to ***** **' **** *** ******* ****** ****** *** ******. Also, *******, *** ***** ********** **** ** *** ***** ** the *** ***** *** ****** ** *** ******, *** *** excluded **** *** ****** ****, *** ** ***** ****** **** triggered.

****** ****** *** ***, *******, ***** *** **** *******. *** camera *** **** ** ****** *** ******* ** **** ***', as ***** ****:

Comments (31)

It's promising to see improvements on video motion detection. It's clearly one of the most fundamentally important aspects of recording (as well as search to have accurate tagging of real motion events).

The downside remains the lack of VMS support. Mobotix is understandable, given their business model but Axis really needs to promote this and work with their VMS partners to get 3rd party support.

Finally, this is the start of an ongoing series of tests we will be doing on motion detection and analytics from 'regular' cameras to 'smart' cameras.

Has VMS integration improved since 9 months ago?

When you say the VMS doesn't support it, does that mean you can't configure it from the VMS, or that it won't help the VMS detect motion better at all?

Hey James, good question. I'll check with Axis.

In the VMSs we used, "not supported" meant that the VMS simply did integrate motion events from VMD 2.1. They use a different API. It's actually surprising that more haven't integrated it.

Apparently Exacq can be made to support VMD 2.1.

The integration is somewhat less than user friendly.

(my submission for understatement of the year).

I see that Axis has now come out with VMD 3.0

What improvements did you see with that?

Rob, do you have a link to Axis VMD3.0? I do not recall any formal announcement of it.

http://www.axis.com/products/video_motion_detection/ - at the bottom:

=====================

Earlier releases

For those products that don’t support VMD 3 AXIS Video Motion Detection 2.1 is still available. AXIS Video Motion Detection 2.1 is based on the same reliable motion detection technology, but does not include False Alarm Filtering and visual configuration support.

===================

John, it is all about signal to noise ratio in those low light scenes. Until you have cameras that have a better SNR in low illumination the software will have trouble with identification. Keep up the good work you are doing your test pictures are worth a thousand words.

Mark, the main camera we tested was an Axis P 'Lightfinder' series, which is one of the best low light, non IR cameras in the market, so that should underscore the difficulties.

We've been considering another small scale implementation of less than 10 cameras. We think VMD is an important component.

We believe that VMD benefits to total ownership costs are much more substantial than the one-time system procurement cost benefit it delivers through reduced storage sizing. When reliable VMD presents non-real-time reviewers with a high percentage of "valid" motion events, this personnel cost driver which extends over the life of the system can be greatly mitigated.

Thanks for another informative review that is relevant to our immediate needs.

This is a nice start, but rather than seeing how much better Axis' advanced VMD is than their basic VMD, I'd rather see how Axis' advanced VMD does say against Sony's DEPA, or Samsung's VMD, or some VMS's advanced VMD looking at the same scene.

And how VideoIQ does against say AgentVI or IOImage looking at the same scene.

Luis, what cameras are claiming advanced VMD? We can easily test against others but are there specific models that are known (via test/field performance) to be highly robust against false positive motion events?

From the sound of it, though I may be wrong, the Axis VMD sounds like something that would be comparable to Sony's DEPA, which I consider to be a step up from basic activity detection, but below Video Analytics (which actually tries to identify objects and object intentions), which I don't have the impression Axis VMD is on the same level. Same for some Samsung cameras. Anything that has vectored, sizable boxes with some options like trip line and directional alerting I consider a step up from activity detection.

Axis VMD2.1 does not support any lines or rules. It's meant to be better motion detection, period.

Sony and Samung are marketed differently, with rules/events. However, the key question is how well the underlying detection works. I have not heard anything positive about that, even from those manufacturers.

Axis has cross line detection which is likely more comparable to Sony DEPA. We are testing Axis cross line and can try it against DEPA (Ethan/Derek)

Does Sony still include DEPA in Gen 6? I thought it was now "VMF" or something like that. It looks like it includes cross line as well as simply "object detected" analytics, according to the manual (page 30).

I admit I'm not sure myself on that one. (DEPA in Gen 6)

Ethan - Sony Gen5-V, Gen6-V and Gen6-E series all have both VMD and VMF. (All other cameras have VMD, but not the VMF menu.) Basically, you can place a Video Motion Filter (VMF) inside any VMD area. This is the DEPA Advanced solution you setup in the camera menu. Many recording platforms can use VMD and/or VMF as event triggers.

Another reason to use lighting to improve analytics in low light levels. There will never be a replacement for good lighting!

I am looking at proposing Mobotix just for this new and improved analytic. What I wish these vendors could provide is some kind of guideline which correlates PPF and analytic performance?!

Yes there are many variables which will affect actual results - however a decent starting reference point assuming 'ideal' conditions would be nice.

Mark, above comment, mentioned SNR in relation to performance. I'd like to understand that more. What exact signal, and what exact noise?

RJ I have used the new MxActivity sensor down to .3 lux in color according the sensor on the camera with no problem using the new M15 series Mobotix cameras. Detection is very acurate. In fact we installed these for the local water management district to self monitor and the detection and notifications have been so good they have turned off thier false alarm prone security systems with perimeter beam detectors. They tell me this will save them thousands in false alarm fines. The test above was taken using the Q24 in 360 mode instead of on the wall in 180 mode which is a better application for the analytics. Also if lighting is a concern the new M15 has much better low light performance than the Q24.

360 is the most common mode for fisheye / panoramic cameras, ergo our test setup.

That said, it's pretty clear that both analytics worked impressively so long as the camera has sufficient lighting. Whether it's a different model that's better in low light or using IR, that should go along way.

Mike,

That is promising! I have scoped the M15 5MP cameras. 1 color lens, 1 b&W lens.

Can you explain the 'fines' they receive for false alarms? So they had a traditional perimeter beam detection system? I looked at that for this application initially

Great article, I would also like to see a VMD 2.1 test on the Axis Q16xx series, considering their WDR plus low light capabilities

Harold, the P3354s we used are indeed Lightfinder. They're not WDR like the Q1604, but they perform similarly in low light. I'll add this to the post, thanks for pointing it out.

We have been using DEPA already since the begingin with cs-50,and then ch-140 and now the ch-180 with milestone to trigger start/stop recordings in an outdoors environment. For us its a huge money and time saver, since we pay per used GB in storage and it gives us the ability to store high quality pictures (16 fps, 2mb/sec, mpeg-4). Timesaver when we go through recordings since i only stores when something moves in a predefined area. It comes with the camera so you dont need to pay any extras, its already there. Not using it (where its aplplicable) seems a waste to me. You just need to set it up and it saves you time and money.

Granted its not "perfect" and we set the parameters so we rather get a little bit more recordings, but still..

It works great for us.

//Snowman

RJ,

They had many false alarms according to the police which resulted in fines year after year. The result of foliage, animals and what I think was a faulty installtion, but they did not ask me to fix it so I offered the video solution, which quite frankly is much better anyway. I have filtered out all external causes, sun blow out, leaves, trees, rain, shadows etc. The only activations they have had have been real movement and they get notified by email within 20 seconds of the event for varification. They already caught one intruder who was a homeless person. If you wish to get further info or setup details you can email me at mike@dfs-group.net.

John,

Thanks for your input, but in our case most of these cameras are setup as 180 on wall set. Most of our customers prefer the 180 view over the 360 view, the ActivitySensor excells in the wide angle applications. We do however use the 360 in retail for the heat map function. I do have to correct a couple if innacuracies in the setup video. 1. In event settings the event dead time is not a how long it takes for the event to stop, but rather how long it will take for an event to retrigger. 2. The activity sensor threshold of 1 thru 99 is not how many seconds of motion before activation, but rather how much activity in the scene it takes to trigger an event.

I wish to ask, since VMD is pretty much on the camera side. Could we balance out the use of VMD versus allowing the VMS to do the recording on motion only. For example, where the network has sufficient bandwidth to handle continuous resolution video, just stream the images to the recorder permanently and let the VMS use pre and post roll setting to record what is important. Then if you have areas where network bandwidth is week, use the a suitable camera with VMD to only push video when it is required.

You can use VMS server side motion detection - a lot of people do. See our camera vs server side motion detection for pros and cons.

Has there been any improvementsuggestions in advanced VMD since this article?

Login to read this IPVM report.
Why do I need to log in?
IPVM conducts unique testing and research funded by member's payments enabling us to offer the most independent, accurate and in-depth information.

Related Reports

Axis Vs. Hikvision IR PTZ Shootout on Sep 20, 2018
Hikvision has their high-end dual-sensor DarkfighterX. Axis has their high-end concealed IR Q6125-LE. Which is better? We bought both and tested...
Avigilon Announces AI-Powered H5 Camera Development on Sep 19, 2018
Avigilon will be showcasing "next-generation AI" at next week's ASIS GSX. In an atypical move, the company is not actually releasing these...
October 2018 Camera Course on Sep 13, 2018
Today is the last day to save $50 on the October 2018 Camera Course, register now. This is the only independent surveillance camera course,...
VMS Export Shootout - Avigilon, Dahua, Exacq, Genetec, Hikvision, Milestone on Sep 13, 2018
When crimes, accidents or problems occur, exporting video from one's video surveillance system is critical to proving incidents. But who does it...
US DARPA Investing $2 Billion In AI on Sep 11, 2018
The US Defense Advanced Research Projects Agency (DARPA) is granting more than $2 Billion to companies developing new AI technologies. The money...
Dahua Low-Cost 4MP Camera Tested (N44CL52) on Sep 10, 2018
4MP use continues to increase, especially in low-cost models, according to integrators in our 2018 Resolution Usage Statistics. We bought Dahua's...
Directory Of 110+ Video Management Software (VMS) Suppliers on Aug 30, 2018
This directory provides a list of Video Management Software providers to help you see and research what options are available. Listing...
Inputs/Outputs For Video Surveillance Guide on Aug 24, 2018
While many cameras have Input/Output (I/O) ports, few are actually used and most designers do not even consider them. However, a good understanding...
Luxriot VMS Profile on Aug 23, 2018
Luxriot is more popular than Hikvision and Milestone products according to ASMAG which was probably even surprising to Luxriot. The company has...
Synology Surveillance Station VMS Tested on Aug 22, 2018
With so many low-cost NVRs and enterprise VMSes, is there any place in the market for NAS-based VMSes? Recently, IPVM bought a Synology NAS for...

Most Recent Industry Reports

Alexa Guard Expands Amazon's Security Offerings, Boosts ADT's Stock on Sep 21, 2018
Amazon is expanding their security offerings yet again, this time with Alexa Guard that delivers security audio analytics and a virtual "Fake...
UTC, Owner of Lenel, Acquires S2 on Sep 20, 2018
UTC now owns two of the biggest access control providers, one of integrator's most hated access control platforms, Lenel, and one of their...
BluePoint Aims To Bring Life-Safety Mind-Set To Police Pull Stations on Sep 20, 2018
Fire alarm pull stations are commonplace but police ones are not. A self-funded startup, BluePoint Alert Solutions is aiming to make police pull...
SIA Plays Dumb On OEMs And Hikua Ban on Sep 20, 2018
OEMs widely pretend to be 'manufacturers', deceiving their customers and putting them at risk for cybersecurity attacks and, soon, violation of US...
Axis Vs. Hikvision IR PTZ Shootout on Sep 20, 2018
Hikvision has their high-end dual-sensor DarkfighterX. Axis has their high-end concealed IR Q6125-LE. Which is better? We bought both and tested...
Avigilon Announces AI-Powered H5 Camera Development on Sep 19, 2018
Avigilon will be showcasing "next-generation AI" at next week's ASIS GSX. In an atypical move, the company is not actually releasing these...
Favorite Request-to-Exit (RTE) Manufacturers 2018 on Sep 19, 2018
Request To Exit devices like motion sensors and lock releasing push-buttons are a part of almost every access install, but who makes the equipment...
25% China Tariffs Finalized For 2019, 10% Start Now, Includes Select Video Surveillance on Sep 18, 2018
A surprise move: In July, when the most recent tariff round was first announced, the tariffs were only scheduled for 10%. However, now, the US...
Central Stations Face Off Against NFPA On Fire Monitoring on Sep 18, 2018
Central stations are facing off against the NFPA over what they call anti-competitive language in NFPA 72, the standard that covers fire alarms....
Hikvision USA Starts Layoffs on Sep 18, 2018
Hikvision USA has started layoffs, just weeks after the US government ban was passed into law. Inside this note, we examine: The important...

The world's leading video surveillance information source, IPVM provides the best reporting, testing and training for 10,000+ members globally. Dedicated to independent and objective information, we uniquely refuse any and all advertisements, sponsorship and consulting from manufacturers.

About | FAQ | Contact