Testing Axis Zipstream

By Ethan Ace, Published on Apr 13, 2015

Has Axis found its breakthrough differentiator? Can they demand a premium for this?

Axis claims Zipstream can "lower bandwidth and storage requirements by an average 50% or more."

We upgraded the firmware on the Axis Q1615 we first tested last September to test Zipstream.

Here is a 1 minute preview of the Zipstream test:

Undoubtedly, Zipstream can drop bandwidth consumption dramatically but here are the big questions:

What type of image quality degradation exists and when did it occur?
How much bandwidth consumption savings exists?
What mostly drives the bandwidth consumption savings - the dynamic RoI compression or the dynamic GOP variation?
How does Axis Zipstream compare to the most efficient bandwidth cameras on the market and to other bandwidth savings modes (like Arecont and Vivotek)?
How might it compare to H.265?
What type of issues occurred with VMSes?

Key ********

**** *** *** *** findings **** *** ****:

******* *** ************** **** ** *** bandwidth ******* (**% ********* in ********* ****** **** this *****) *** ******* to ** *** *** innovation ****. **** ** the ******* ** ****** the * ***** ******** on ******, **** **** there ** ******* ******** (saves *********) *** ****** an * ***** ******** on ****** ****, *.*., something ***** ******* ** the ***** (******* ***** quality **** *** ******. In *** *****, **** Dynamic *** *******, ** saw * ***** ********* commonly ***** **** * to ** *******.

***** ***** ******* *********** **** ** ********** *********** on ***** ** ******* plates *** **** **** clear ***********, ********** **** dynamic *** *** ******* on ******* ** ***** (examples ******** ******).

******* *** **** ********** than ***** ********* ****** features** **** ****** ****** which ****** ************* ******* quality (*.*., *******) ** did *** **** **** much ** * ********** / **** *** * very **** ****** (*.*., Vivotek).

****** *** ***** ******* against *** ********* ********** cameras. ** ****** ****** from *****, ********* *** Samsung. **** ********* ***, Axis *** *** **** competitive. **** ********* '*****' out, **** *** ***** was *****. ** *******, especially **** ******* *** disabled, ******* **** *****.

**** ********* **** *.*** with ********* ******* ***: ***** *.*** *** reduce ********* ** ** 50% ******** ** *.***, Dynamic *** **** ****** bandwidth ** **%+ ** many / **** ************ scenarios **** *** ***** benefit **** ** ******** no ******* ** *** VMS **** (***** *.*** does) *** *** ********* to **** **** *.*** for **** *******.

** *** *********** ************ *** ********* ******** with ********, *****, ******* and *********.

Market ******

********* ** * ********** improvement *** *** **** a *********** ****** ******. However, ******** ********* ** limited ** ******-* ******* which **** * *** existing **** ******* *******. Most *** ******* ***** released ***** ******* **** support ******-* *** ********* Zipstream.

*** *** ******** ** whether ***** ************* *** enable / ******* / source ******* *** ** match ****.

********* ** ********* * 'Smart *.***' ***** **** claim ** ******* ** Zipstream *** **** *************, including *********, *** ********** H.265. *** ** **** will **** ********* * much **** ******** ************** for *** *********** ******.

Zipstream *************

********* ******* ******* *** a *** *** ** the ***** ****** ******** page ** *** ******'* web *********, **** *****.

**** *** ******** *** options:

"*.*** ******* *********" *********** adjusts *********** ****** ** different ***** ** *** scene, **** ****** ******* less **********, *** ****** background ******* **** **********. Settings ***** **** *** to ****, **** **** different ****** ** ******* reduction *** ******* ******.
"******* ***" ****** *** camera ** ****** *** I-frame ********, ** * range **** ******** ** set ** *** *.*** tab, ** * *******, to * ******* ** 1200. *-***** ******** ** decreased **** ****** ******* are *******, *** ********* when *** ***** ** static ** ****** *********.

** **** ***** ** review *** ************* ******* for *********, ******* ******* quality *********** (** **** thereof) ** ***, ****, and ******* *** ******.

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

*********, ************ *** ******** as * ****** **, with * *********, ***********, ********** ** ********* being **, ***, ***, medium, ** ****. ***********, **** can *** ** *******. With ****, *******, *********** parts ** *** ***** are ***** ** ** more ****** **********. ***********, it ******* **** **** is **** ********* *** default *********** ***** ***.

*** ***** ***** ***** the ********* ******** ** two *****, *** **** Zipstream ****** ***, ** the ****, *** ******* with ** *** ** High, **** ******* *** on, ** *** *****. Both **** * ****** quantization ***** *** **** deviation **** ***.

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

*** ***** ***** ***** * comparison ** ******** ***** GOP (** ** **** case) ****** *** ******* GOP **** ** ********* (set ** **** ***), as ******** ** *********.

***** *-****** ** *** left *** ******, ***** 10 ******, ******** **** on *** ***** ******, from **, ** **, to **, ** *** camera ******* ******** ********* to **** ** ** the *****.

Bitrate ******

********* ******* ******* ** all ****** ******, **** best ******* ** *** motion ****** *** ********* scenes.

** ****** *********'* ******* on * ****** ***** and *** **** * moving ******* ** **** light ** **** *****. Bandwidth ********* **** *.* Mb/s **** *** ***** was ****** ** ***** 80 **/* (*.** **/*) when ***** ********* **** plus ******* ***.

*** ** **** ***** example ** ****** ********* reductions ** *** **** scene ** ***** *.* lux, **** **** ******* effects, ******** ******* **** nearly ** **/* ** ~2.7 **/*, ** ~**% reduction.

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

** ***** ******** ******** in ** ****** *** motion *****. ** **** light, ~*****, ********* ******* Dynamic *** *** ********** effective, ******** ******** ** 14% ** *** *** 25% ** **** ******** to ************ **** ********* off. *******, ******* ******* GOP ** ** **** still ***** ******* ******** by **** **% ** just **** *** **/*.

** * ***, *** effects ** *********'* ******* compression **** **** **********, reducing ******* ** **-**% Turning ******* *** ** in **** ***** ***** resulted ** ** **** 90% ********, ****** *** as ***** * **** as ** **** *****.

** ******* ******* ******, Zipstream *** ********* ********* to *** ****** ******* scene, **** ** ** 26% ********* ***** ******* ROI (******* ******* ***). Turning ******* *** ** increased ********** ** ** and **%, ********* ** the *****. ***** ********** are ***********, *** ******* compared ** ****** ***** due ** ****** ******* of ****** ** *** scene, **** ******* ********, foliage, *** ***** *********.

High ***** **** **. ***

********** ** ********* *** to ********* **** **** similar **** ***** **** high ***** ***** *** low. ** *** ****** still *****, ******* ********** due ** ********* (******* dynamic ***) **** ******** higher **** ***** ** FPS **** ** ***. However, ********** *** ** dynamic *** **** ******** greater **** ***** ** FPS.

Image ******* ******

** *** *****, ** saw ******* ****** ** image ******* ****** ** Zipstream. ****** ******* *** text ******** ******* *******, without ******, ***** ********** objects (*******, *****, ***** and ***** ********, ***.), less ********* ** *** scene, ********* **** ***********.

** ****** ***** ******* in ******** ****** ****** below.

***** ********/****

*** ***** ***** ******** images ** *** ******* and **** ***** ***** with ********* *** (** the ****) *** * fixed *** ** ** vs. ********* *** ** High, **** ******* *** turned ** (******* ** 1200).

******* ******* ** **** is ********* *******, **** no ******* *********** ********* from *** ******** *********** or ****** *-***** ******** used ** *********. ******* details ** *** ******* are *******, *** **** the **** **** ** the **** ***** ** line * (*****) ** similarly ******* ** ****.

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

******* **** **** ******* when ********* * ****** vehicle, **** ***, ****, and ******* *** ******** all **** ** ******* a ******* ***** ** ~15-20 *** ** * ~20' ***** ** **** without *****.

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

******* ******* *********** **** most ********** ** ******* areas **** ****** *******, such ** *** ***** below. ****, ******* ******** and ****** ***** **** clear **** ********* *** appeared ******* ** ******* when ** *** ****** on, ******* ** *** Zipstream **** ***** *****. When ******* *** *** added, ***** ******* **** most **********, ********* ****, seen ** *** ***** below.

Versus ***** *************

********* ** ****** ********* at ******** ******* ** supported cameras. *******, ***** ************'* cameras ***** ******* ***** bitrate ** **** ******.

*** **** ** *** scenes ** ******, ** did *** **** ***** test *** *********** *** the ***** ***-****, ********* 4132 *** ******* ***-****. Below *** *** *******:

VMS ***********

** *** ** ****** streaming ** ******* **** Zipstream ***** ** ******** Control ******, ***********, ******* Security ******, ** ********* XProtect. ********** ** ********* or ******* *** ********, cameras ********* *** ******** as ******, *** ******** was ******, ******* ***** gaps, **** ******* *** reverse.

**** **** **** ********* or ******* **** ***** in *** ***** ****** or *********, ****** *****, with *** ******** ********** a ***** ****** **** only *-***** ******* ******* until *** **** *-*****. This *** ** *********** when ******* ***-*********** (***, MKV, ***, ***.) ***** in **** ********.

******* *** ********** ********* settings **** ***** ****** Tool, ******** ***** ** set ***** ********** **** that ******, ******* ** using *** *** *********, shown ****:

Test *********

***** ******** ******* *.**.* was **** *** **** test. **** **** **** firmware ** *** *** publicly *********, *** *** supplied ** **** *** testing.

*** ****** ******** **** left ********* ****** *** WDR, ***** *** ****** off ** *** ******, and ********, ************ ** 1/30s.

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

******** ******* ****** *.*.*.**
*********** *.*.*.*****
********* ******** ********* **** 9.0c
******* *.* ***

Comments (12)

Brian Rhodes

IPVMU Certified | 04/13/15 05:26pm

Are changes to Dynamic GOP triggered only by movement? Does changing motion sensitivity affect Dynamic GOP (I-frame sampling) sensitivity?

Also, if I understand it right, are the efficiencies diminished the lower the framerate is set to begin with (ie: greater at 30FPS vs 8 FPS)?

Ethan Ace

IPVM | In reply to Brian Rhodes | 04/13/15 08:13pm

30 FPS vs. 10 FPS results are added above. In short, the differences are not that great, and close enough that I'd be hesitant to declare a winner one way or another.

Undisclosed #1

IPVMU Certified | 04/13/15 05:31pm

Thorough testing!

Note that when streaming or playing back clips in VLC Media Player or Quicktime, errors occur, with the programs displaying a blank screen with only P-frame changes visible until the next I-frame.

Are Zipstreamed streams available as ONVIF streams?

Do you think that the h.264 is possibly non-standard, which might explain the VLC and AVinaptic behavior?

Ethan Ace

IPVM | In reply to Undisclosed #1 | 04/13/15 05:38pm

Yes, they are available via ONVIF.

No, I don't think it's non-standard. It's an I-frame/P-frame issue in those players. We're talking to Axis to get their read on the AVInaptic analysis.

Marc Pichaud

04/14/15 01:05pm

I'm really wondering when you enable dynamic GOV/GOP how it doesn't impact according to your test - the blur in image... when you provide a key frame every ..10 seconds or worse what is your chance to get a plate or a face, or a sudden move..oups... Zip is focussing on moving parts, so like smartstream keep a better quality on moving pixel... but DGOV kills that, by adding more P-frames.

For outdoor non detailed scene (without LPR or Face reco in dark envir. with quick moves) I agree Zip+DGOV is one of the current option to decrease both bandwidth and storage.

To be compared also with adaptive streaming based on 1 key frame and no P-frame when no activity is reported by motion detection, raising to normal fps when motion arises.. In that case, you can keep your key frame for police, it's supposed to be correct

I have experience Smart stream with good quality on moving pixels and/or Roi and big compression on static scene. MAin concerns is to explain to the customer why the static scene pixelise when darkness arises... even with DNR is enable

Undisclosed

04/14/15 04:58pm

I should read the review more carefully. But... grumpy engineer-user response:

so you came up with this overly complicated video compression scam, then you deployed just enough of it to work, and now you're "tuning" it (if you insist calling tweaking a compression algorithm "tuning") and you want me to get excited about it?

Any camera vendor can twist the H.264 voodoo. It's gotta have a billion knobs and dials on the inside no matter how much is exposed through the interface. I don't see this as a differentiator I see it as an enabling technology for more incompatibility.

John Honovich

IPVM | In reply to Undisclosed | 04/14/15 05:22pm

Rodney, you really should review the review more carefully. Why would you preface a sweeping, ungrounded, attack with that fatal admission?

How is 'incompatability' an issue given that we tested it and it worked with 4 of the biggest VMSes in the world without anything done on their part?

Secondly, if 'any camera vendor can twist the H.264 voodoo', why isn't Axis doing the vodoo on all of their cameras instead of just the new ARTPEC-5 ones? Any if this is just knob and dial turning why hasn't anyone else done this?

There is something genuinely new and useful here. Period.

Irina Manea

04/15/15 11:46am

I have understood the priciple behind the idea. They want to mimic Human brain. Only use the most relevand data in the image and the rest compress it to maximum.

After a couple of modification to the logic behind the software i think it could make a big diference in certain aplications.

Undisclosed #1

IPVMU Certified | 04/24/15 04:08am

I'm curious what the net effect of Dynamic GOP running together with VMD framerate speed-up on the VMS would be. It occurs that VMD framerate speedup functions like a poor man's Dynamic GOP. Although not increasing the cardinal interval between I-frames, VMD speedup does increase the realtime interval between I-frames and thereby reduces bitrate.

Of course motion detection has to be setup, and there are only two levels of frame-rates, so cruder for sure. Still I think if someone is using VMD speedup aggressively, they will not see as large a benefit from Dynamic GOP.

btw, Geurtebruck apparently has encoders that use Dynamic GOP...

John Honovich

IPVM | In reply to Undisclosed #1 | 04/24/15 11:22am

Those Geuterbruck encoders are MPEG4.

Undisclosed #1

IPVMU Certified | In reply to John Honovich | 04/26/15 06:30pm

Yes, from back in 2008!

Here's a spec sheet from an updated one:

Ethan Ace

IPVM | 08/11/15 09:33pm

Genetec has integrated Zipstream settings for Axis cameras to their Config Tool, seen here:

These settings directly correlate with Zipstream settings in the web interface, so users may now configure them from here.

Read this IPVM report for free.

This article is part of IPVM's 6,590 reports, 889 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.