Bandwidth vs Low Light Shootout - Avigilon, Axis, Bosch, Dahua, Geovision, Hanwha, Hikvision, Uniview, Vivotek

By IPVM Team, Published Feb 08, 2019, 11:21am EST

Nighttime bandwidth spikes are a major concern in video surveillance, but do all manufacturers' cameras perform the same? Are some more consistent from day to night than others? And do manufacturer models all similar?

To find out we shot out and compiled results from over 50 cameras from 9 manufacturers - Avigilon, Axis, Bosch, Dahua, Geovision, Hanwha, Hikvision, Uniview, Vivotek - to see how each did, what was average, and who was worst and best and how much performance has improved in the past several years.

Manufacturer ***********

**** ** ************ *********** compared ** *** *****:

  • **** *** ***** ****** spikes/most **********:**** *** ***** ********** IR ****** **** ************ produced *** ******** ******* spikes, **** **** ********* close ** **%, ***** 1/3 ** *** ******* average. **** *** *** true ** ***** ***-** models, ***** ****** *********** like ***** ****** (*** below).
  • ******/***** **********, *** ******:******'* ****** ******** ***** 125% ********, ***** **** average *** ****** **** Axis/Bosch. ****** ****** **** generally *** ***** ** average, **** *** ***-***** an *******, ********** **** ~52%. ***** ****** ******** only **** ******** ***** the ****** ******** (***% vs. ***%), **** **** one ***** ******* **** 200%.
  • ********* **** ***********:***** *********'* ******** ** average **** ***** ***%, their ****** ******** *** two ******* ****** ** IR ******* **** ** our *****, **** ***%. There *** ** ***** trend ** *** ********* models *********, **** **** newer ******* ******* ****** than *****, *** **** versa. ***%
  • ******** ****** **** *******:******** ******* ****** **** over *** ******* *******, ~225%, **** *** ***** increasing ***% ** *****.

General ********

** ******** ** *** manufacturer ******** *********** *****, there *** ******* ******* takeaways:

  • ~***% ******* ******** ** Integrated ** *******:****** **+ ******* ******* from ***** ** **, bit **** ********* ** average ** ~***% **** day ** *****. ********* varied **** * *** of ~**% ** **** 400%.
  • ***-** ******* **** ******:***-** ******* ****** ********* in *** ***** **** day ** *****, **** a *** ** ~***% to * **** ** over *,***% (*** *********). However, ***** ****** *** far **** ****** ** surveillance ** ****, **** the **** ******** ** models ********* ********** **, while **** *** *** in *** ******** **** test.
  • ******* ******* ****** ***** past *****:*** ***% ******* **** in **** **** ** roughly **% ***** **** the ***% ******* ** our **** ***** *** to *** ********** ***** of ********** ** ** 2019.
  • *** (*** **** * cap) ***********:** ****** ********* ******, we ********* ***** ******* bit **** ****, ** this ****** *** ***** to **** ** *** as ******** ****** **** complex *****, *** **** bandwidth ** * ********* limit ** ***** ****** network ******* *** *******, with ******* ****** ** video *******.

**** **** ***** **** no ********** ****** ** how ****** ****** ***** on **********, **** **** models ********* *** ********* and ****** ******* ** each ********** ******.

Results ********: ***% ******** *******

**** *** *** ******* sorted ** ************. ****** 30+ ******* ******, *** average (****) ********* ******** from *** ** ***** was ~***%. *** ****** was ***%, ****** ****** increases ****** ***********, **** outliers ** *** ** 17% *** **** **** 400%.

2019 Nighttime Bandwidth Spikes By Manufacturer

Most ********** *************: **** *** *****

** *** * ************* tested, **** *** ***** integrated ** ****** **** consistently ******* ******* ******, with **** ********* ***** to **%, ***** */* of *** ******* *******. Note **** ***** ****** fewer ** ****** **** Axis, **** **** ** their **** *** ****** lacking *** ******* (*.*. the********* **** *********), ***** **** ****** integrated ** ** *, P, *** * ****** domes *** *******.

Others **** ******

***** ************* ****** **** more ******. *** *******, some *** ********* ****** increased ** ***** ***** to **** *** *****, below ***%, ***** ****** produced *** ******* ****** of ** ******* ******. Additionally, ******* ******* ******** the **** ****** ********** increase (**%) ** ***** 4MP *****, *** *** of *** ******* ** the ** ******.

Much ******* ****** **** **** *****

** ******** ** *** 145% ******** **** ** current *****, ******* ** our **** **** ******** 498% ****** ********* ** night, **** *** ****** spiking ****** *,***% ******.

2014 Nighttime Bandwidth Spikes By Resolution

Integrated ** ******* ****** / ***-** ******* ***** *********

*** **** ********* ****** in ****** ********* ****** is ******* *** ****** includes ********** ** ** not. *** **** ********* single ****** ** ****** performance ******* *** ******** test ** **** *** models ****** ** **** is *** ********* ** integrated ** ** ****** every ****** ******. ** 2019, **** ************* ******* integrated ** ** **** models, **** **** *** cameras *** * ******** of **** ******* ******* it. *******, ***** *** some, **** ** *****, whose ******* ********** ****** (such ** ************ **** ***** *****) ** *** ******* IR, ****** ***** *** the ********.

***** ***-** ****** ***** much **** ************* **** non-IR ******, **** **, 10x, ***, ** ****** increases ********.

2019 Non-IR Cameras Spike Drastically

***** ****** *********** *** improved, **** **** ********* encoders, ****** *** **** sensitive *******, *** ****** processing, *** ******** ** infrared ***** ** ********* surveillance ******* *** *** biggest ******** ****** ** bitrate.

Fisheye *********** ****** ***********

*** ******** ** ***** the ******** ***** ***** down ** ******* *******. In *** *****, ******* models ****** **** *********** than ******, **** **** the ******* ****** ******************** *****. **** ** generally *** ** **** low ***** ***********, **** night **** ****** **** darker, *** **** **** detailed **** *******, *** this ** *** *********** true, ** *** **** M3058-PLVE *** *** ** the **** *** ***** performers ** ************ ******* ****** ********.

2019 Fisheye Cameras Vary Drastically

Root ***** ** ********* ******: ******** *****

*** ****** ** ***** spikes ** *** ******* moving *****, *** ****, static, *******, ** **** that ** ********* *********** **** *******.

****** *** *** ***** to **** * ***** or ***** **** ***** was *** ****** ****** is ******** **** **** noise, ***** ***** **** this:

** ** *******, **** noise, **** *** **** contrast (***** *** ***** near **** *****) *** rapid ********* (******* ******), make ** ****** ** encode. ******,********* ******* *** ****** / ***** **** *** similar *** **********. ***** undermines **** ** *****.

Recommendation: *** *** **** * ***

*** *** ** ***** massive ********* ****** *** be ******* ** ******* *** **** * cap, **** ***** ** MBR(******* *** ****). **** method ****** ******** ** fluctuate ** ********* ** fit *** *****, *** only ** ** * specified *****, ** ***** point *** ****** ********* compression ** ****** ** drop ****** ** **** within *** *** (********* the ****** ** *********).

** *** *****, **** when **** ******************** ********* **** ** average ** ~** ** 35, ***** ******* *********** were *******, **** ****** bandwidth ***** ** **** 80%

bandwidth cap minimal impact on details

Constant *** **** *** ***********

******** *** **** *** also ** **** ** prevent ******, ***** ** locks *** ***** ** one ********* ***** ********** of ******** ** *** scene. *******, ****** *** day, **** ****** *** less *******, *** *** waste *********/******* ** *** camera ***** **** *********** encode *** ***** **** set ** ***. ******* of ****, *** ** not *********** ****** ** the **** ****** ***** does *** ******* ***.

Digital ***** ********* ******

******* ****** ********* ***** is **** ** ****** spikes *** ** ********* noise ** ******* ***** reduction (***), ***** ******** to ******* ***** ****** by ********* ** ****. There *** *** *** methods, ******* (***** ********** in * ****** *****) and ******** (***** ********* by *********** ******** ******), which *** ********* ******** in ******. ***** ***, spikes *** ** *********** cut, **** **-**% ********** in *** **** ******** *** *****.

*******, *** *** **** introduce ******** *******, ** well, **** **** ****** of ******** ***** ********* causing **** ** ***** (shown *****), ********* ******* to *** ******* ** slow *******. ******* ** this, ***** ****** ** careful ** *** ******** DNR ** ******* ******.

*** **** ******* *** testing, *** ********* *** (******* ***** Reduction) ******** ****.

Test *******

** ****** **** ** cameras **** **** ************* day (~*** ***) *** night (~*.* ***) ******** motion ****** ** ** FPS.

*** ******* *********** ****** were************ ***-** ****** ******* smart ****** **. ***** codecs **** ******* ** their ********* ****** ***** on *** *****. ***** codec ******** ****:

  • ****:****************, ****, ******* *** 300, ******* *** ***
  • *****:***** *.***+ **
  • ******:***************, ******* *** ***, Dynamic *** ***
  • *********:*.***+ **
  • *******:*-**** **, *****
  • *******:*********** **/***, ******* *****-***** period *******

*.*** *** **** *** consistency, ** **** ******* did *** ******* *.*** streaming.

Comments (24)

Undisclosed Manufacturer #1

02/09/19 04:18pm

I might have missed this, but what compression did you test this on, H.264 or H.265? I'd be interested to know if H.265 on the same camera would lower the percentage increase. The overall bitrate would most likely be lower. What would intelligent codecs do, Zipstream, Wisestream, H.265+, etc.

Agree
Disagree
Informative: 4
Unhelpful: 1
Funny
Avatar

Ethan Ace

IPVM | In reply to Undisclosed Manufacturer #1 | 02/11/19 02:26am

You didn't miss it, I forgot to put it in! These measurements are all H.264, with smart codecs on. I'll double check some things in setup to add detail to the report on exact settings and post back tomorrow.

Agree
Disagree
Informative: 3
Unhelpful: 1
Funny
Avatar

Guilherme Barandas

In reply to Ethan Ace | 02/14/19 11:16am

Ethan, so you did the tests on H265, could you show us.

Agree
Disagree
Informative
Unhelpful
Funny
Avatar

Ariel Levy

In reply to Undisclosed Manufacturer #1 | 02/12/19 02:38pm

H.264 is dead.

Agree: 1
Disagree
Informative
Unhelpful: 3
Funny

Undisclosed #2

IPVMU Certified | In reply to Ariel Levy | 02/12/19 03:25pm

Long live H.264+!

Agree
Disagree
Informative
Unhelpful
Funny
Avatar

Jörg Steuerwald

02/11/19 01:00pm

Can you describe more the settings of the cameras , like Frame-rate, and image quality settings. In some cases the 4MP produce less bandwidth as a 1080 resolution from the same vendor.  

Agree
Disagree
Informative
Unhelpful
Funny

Jason Crist

02/11/19 01:05pm

Very interesting that Sony wasn't tested, the VB770 and newer gen low light cameras are the hottest thing in the industry.

Agree: 2
Disagree: 2
Informative
Unhelpful
Funny

John Honovich

IPVM | In reply to Jason Crist | 02/11/19 01:13pm

No, they are not the hottest thing in the industry. Indeed, in our favorability survey of Sony just released only 1 out of 200+ integrators even mentioned Sony's low light offerings.

That said, we have tested the VB770 here.

Agree
Disagree
Informative
Unhelpful
Funny

Jason Crist

In reply to John Honovich | 02/11/19 03:09pm

I'm new here so I'm learning the ropes.  The surveys of integrators clearly are the least useful of any of the resources here, HIK gets the most good and the most bad (paraphrasing but it's a clear phenomena).  I expected a consumer reports style engineering based review system, which certainly many features on here are but the surveys are clearly not.  And I didn't bring it up, you referenced it in this context.

Sony hasn't done a great job of keeping their products in front through marketing recently, they aren't going to win any popularity contests, but the VB770 has garnered some accolades in it's short time on the market.  But low light has been the center of what they have focused on in the past couple of years and it's disappointing not to see them included.  Not your fault they've been quiet for a while, but the superior performance of their products from video quality and reliability has rarely been factually contended.

Agree: 2
Disagree
Informative
Unhelpful
Funny

John Honovich

IPVM | In reply to Jason Crist | 02/11/19 03:26pm

I expected a consumer reports style engineering based review system

Which we do - Sony 35mm Super Low Light Camera Tested and we also did Axis vs Sony Super Size Shootout (Q1659 vs SNC-VB770)

The surveys complement are engineering / technology tests to show what a wide variety of integrators are using.

Sony's low light performance is very good - but it's very expensive. It's hard to sell many $5,000 cameras when pretty good low light cameras are widely available under $1,000. There are applications where the VB770's price premium is justified but, because of the significant price delta, it's a niche.

Agree: 3
Disagree
Informative
Unhelpful
Funny

Undisclosed #2

IPVMU Certified | In reply to Jason Crist | 02/11/19 05:01pm

...the VB770 and newer gen low light cameras are the hottest thing in the industry.

Sony hasn't done a great job of keeping their products in front through marketing recently, they aren't going to win any popularity contests...

...but the superior performance of their products from video quality and reliability has rarely been factually contended.

”They’re performing like hot cakes!”

Agree: 1
Disagree
Informative
Unhelpful
Funny: 2

Jason Crist

02/11/19 05:12pm

Again, Sony's fault for not getting them out there but the newest releases feature much expanded low light capabilities like the VM772R that are a fraction of the cost of the VB770.

Agree: 1
Disagree
Informative
Unhelpful
Funny

Undisclosed #2

IPVMU Certified | In reply to Jason Crist | 02/11/19 05:35pm

...like the VM772R that are a fraction of the cost of the VB770.

I would hope it’s a very small fraction, since it’s 1/16th as low-light sensitive as the VB700.

 

Agree
Disagree
Informative
Unhelpful
Funny: 1
Avatar

Ethan Ace

IPVM | In reply to Jason Crist | 02/11/19 06:13pm

To be clear, the VM772R is decent for a 4K integrated IR camera. We tested it. But it uses a much smaller sensor than the VB770 and its performance is nowhere near that camera. From our tests:

Also, last I checked that was a $2,000+ camera (street price), which makes it a clear outlier when most cameras are under $1,000, typically far below $1,000. So even if it's a fraction of the cost of the VB770, it is still priced well above average for an IP camera in 2019.

Additionally, it is not a fair comparison to include either the VM772R or VB770 in this test, as they do not include a true VBR implementation (they do not fix compression and vary bitrate as is typical). We've mentioned it in our tests of those specific cameras, as well as this report. So testing them against true VBR cameras which vary bitrate more widely (both higher and lower) would be misleading. If they've fixed that in firmware, we'd be interested to take a look, but last we checked, it was still true.

Agree
Disagree
Informative: 1
Unhelpful
Funny

Jason Crist

02/11/19 05:54pm

would've been nice to have tested it against the others ;)

Agree: 1
Disagree
Informative
Unhelpful
Funny

Undisclosed #3

02/12/19 04:54am

What type of light meter do you use? Can you add a column in your reports to show 1 or 2 different lux meters used in the actual tests? I trust what you say but it would be nice to see the numbers from a few measurement devices for a baseline. Thanks in advance.

Agree
Disagree
Informative
Unhelpful
Funny
Avatar

Edouard Sonnenschein

In reply to Undisclosed #3 | 03/13/19 02:14pm

I would also be interested in the light meter you are using.

Agree
Disagree
Informative
Unhelpful
Funny
Avatar

Ethan Ace

IPVM | In reply to Edouard Sonnenschein | 03/13/19 03:08pm

We use an Extech LT300. It's one of the few we've found that measures down to 0.01 lux. Most meters are not made for low light measurements. Many people can probably get by with 0.1 accuracy, but for our testing we are routinely below that level.

Agree
Disagree
Informative: 1
Unhelpful
Funny
Avatar

Edouard Sonnenschein

In reply to Ethan Ace | 03/13/19 07:47pm

I use the Extech HD400.  Good to 0.01Fc/0.1Lux.  The Extech LT300 specs show 0.01Fc/Lux. So I was not sure it it actually went down to .01 lux.

Thank you for the information.

Agree
Disagree
Informative
Unhelpful
Funny

Alfredo Santiago

ALFREDO'S TELEPHONE SERVICE
02/19/19 08:46am

Good info but at that low level useless if you need to see who or what. You need light.

Agree
Disagree
Informative
Unhelpful
Funny
Avatar

Ethan Ace

IPVM | In reply to Alfredo Santiago | 02/19/19 12:38pm

I would agree with your statement in general, just want to clarify that light doesn't have to mean visible light. Most of the cameras in this test were integrated IR, so there was non-visible light in the scene. 

Agree: 1
Disagree
Informative: 1
Unhelpful
Funny
Avatar

Juan Banda

IPVMU Certified | In reply to Ethan Ace | 08/01/19 03:54am

Thank you for providing the light perspective follow-up. #OnPoint

Agree
Disagree
Informative
Unhelpful
Funny

daniel tobias

09/05/19 09:47pm

Editors,

Thank you for this informative article.

Can you please clarify the unit of measurement for the day and night bandwidth values?

Are those mbps, Mb/s, or an arbitrary value?

Thank you.

Agree
Disagree
Informative
Unhelpful
Funny
Avatar

Ethan Ace

IPVM | In reply to daniel tobias | 09/06/19 04:15am

Figures in the charts are in megabits per second (Mb/s, Mbps, however you prefer to notate it). I'm going to send these charts to our graphic designer to add that notation and will report back when it's done.

If you have any other questions, let us know!

Agree
Disagree
Informative
Unhelpful
Funny
Read this IPVM report for free.

This article is part of IPVM's 6,943 reports, 926 tests and is only available to members. To get a one-time preview of our work, enter your work email to access the full article.

Already a member? Login here | Join now
Loading Related Reports