Shootout: 4K vs 1080p vs 720p Video

Author: Ethan Ace, Published on Jul 17, 2013

4K is the next big thing in video. While 1080p has become the 'standard' for HD, the next step is 4K, quadrupling resolution / pixel count from ~2MP to ~8MP. It is already showing up in TV sets as well as consumer cameras. Indeed, the first 4K surveillance camera was announced this spring, with many more announcements certain to follow this fall.

Can 4K really make a difference for surveillance? Or is Axis right that the 'megapixel race is over'?

We bought a GoPro Hero3 Black camera that delivers 4k resolution (indeed up to 12MP). What is especially interesting is that they use the same chip manufacturer (Ambarella) used by cameras from many well known manufacturers (like Avigilon, Bosch, Pelco, Hikvision, Honeywell, etc.).

We took the Hero 3 Black and put it to the test at FoV widths from 20' to 160' feet to understand 4K performance.

We tested at 12MP, 4K, 1080p and 720p to determine what actually quality difference each resolution / pixel count delivered.

** ** *** **** *** ***** ** *****. ***** ***** has ****** *** '********' *** **, *** **** **** ** 4K, *********** ********** / ***** ***** **** ~*** ** ~***. It ** ******* ******* ** ** ** **** ** **** as ******** *******. ******, ******** ** ************ ****** *** ********* **** ******, **** **** **** ************* ******* ** ****** **** ****.

*** ** ****** **** * ********** *** ************? ** ** Axis ***** **** *** '********* **** ** ****'?

** ****** ****** ***** ***** ********** ******** ** ********** (****** ** ** ****). **** ** especially *********** ** **** **** *** *** **** **** ************ (Ambarella) **** ** ******* **** **** **** ***** ************* (**** Avigilon, *****, *****, *********, *********, ***.).

** **** *** **** * ***** *** *** ** ** the **** ** *** ****** **** **' ** ***' **** to ********** ** ***********.

** ****** ** ****, **, ***** *** **** ** ********* what ******** ******* ********** **** ********** / ***** ***** *********.

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

Key ********

** ******* **** * ******** ********** ** ***** ******* ** each *** ***** ******, ****** **** ********* ********* *******:

  • ** ***' ****, ***** **** ****** **** ** *** ********* an ******* ** ********, ********** ****** **** *** ****** *********** ones **** *****.
  • ** **' ****, ****** ******** ******* ** **** **** ** whereas *** ***** ********** ** *** ***** ***** ******** ****
  • ** **' ****, ****** ******** **** ***** ***** **** ** while **** **** ******** ******** **** ***** *****. ********, ******* characters **** ******* *************** **** ** ****** *** **** *****.
  • ** **' ****, ** ******** ********** ** ****** ******** ******* but *** ******* ***** ** *** *** ***** **** ******* sharper *** ****** ** **** ** ** **** *****.

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

** ***** ***** (** ** *** ****) ** ****** ** see *** **** ********* ******. **** ******** *******, ******* ****, long ********, ***. ******, *****, ****** ***** *** ****** ** get **** ****** *** ******* ***** *******. *******, ** ******* ranges (** *** *****), ****** ***** ***** ******* *** ** interest (**** * ****** ** *******'* **** ** *** ****** number ** * ****), ** ** ****** ** ** ********.

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

****** *** ******** ** ********** ** *****, ******* ** *** jump **** ******** ********** ** *********, ***** ****** ** ***** that *** ***** *********** *** ****** ** ** *******, ********** considering ** ******* **** **** ***** ***** ** ******* **** in *** ***** ** *** **** *-* *****. *** ***** performance ** ** ******* ******* ** ** **** **** ********** models ****** *********.

*** ***** ***** ***** *** ** */*.* ***** ** ***** high *** ************ ****** *********, ****** *** ***** ********** ********. Certainly, ************ ******* **** ******* ***** * ***** *** ** do *** **** *** ***.

12MP **. ** ***********

** *** ** ***** ***********, **** *** ** *** ****** close ** ****** *** **** (**** ********** ****** **. ****). However, ** ****** **** ** **** ********** ***** *** *** reasons:

  • **, *****, *** **** ****** **** ******** **** ******, ** slight ****** *** ******** ********* *** *****. **** ****** **** captured ** **** **** *** ******.
  • *** ***** ******* ***** ** **** *** ********* ******* ***** from *** ******. ****** *** ** ***** ****** ** **** were *** ** "****", ***** *** ****** ***********, * ****** of * *** ****** ** ****** *** **** ** ***** edge ** *** *****.

160' ***** ** ****

*****, ** ***** **** ** ********* **** ***' ********** ***** of ****. 

** **** ***** *** ****** **** * ** ** ** resolution *********. **** *** ** *********** *** **** ** ******* some *******, **** ** ******** *** **** *****, ** **** as ***** *-* ** *** *****. **** * ** *********** readable ** *****, ****** *** **********, *** ******* ** *** subject *** ******* *********** ** **** ********** *** *****.

80' ***** ** ****

****, **  ******* *** ***** ** *** ***** ** **** by ****, ** **'.

** **** *****, *** *********** *** ******* ****/** *** ***** is **** **********. ******* ** *** ******* *** ****** ** see, *** *** ***** *** ** **** ** ** ***** line * ** ** **********, ***** **** ***** *-* *** be **** *** ** *****.

40' ***** ** ****

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


 ** **** *****, *** ******** ** ****/** *** ******** ******** to *****. ******* ** *** ***** *** ** **** **** to **** */** ** *** ** *****, ***** **** ***** 4/5 *** ** **** ** *****. *******, ******* ** *** subject *** ****** ******* **** ** ****.

20' ***** ** **** 

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

** * ***** ** **** **** ******, **** *** **** camera ******** **** ****** ** *** *******, **** ******* ** the **** ***** ******* **** ** **** */*, ****** ****/** likely ******** *** **** **** ******. 

Low ***** ***********

** ****** *** ***** ** * ~* *** ***** ** gauge *** ***** ***********, ** **** *****. *******, ******* ****** not ******** **** * **** **** ** ** ************ *******, as *** **** * **** * ******** ****** *** **.* lens, **** ** ***** ********** ****** *** ***** ***********.

Test ********* 

*** **** ****, ** **** ****** **** * ***** *******. ****** * ******** ****** ******** *** "********* ******" ************, it ********** ****** **** ***** **** ******** ************ *******. ***** on *** ****** *** ** *******:

 

Comments (23)

So 4K is essentially just a fancy way of saying 8.3MP (or 12MP)? Just like 1080P is a fancy way of saying 2.07MP (with progressive scan)?

Don't forget that it's 16:9. Also, who really wants to remember 2.07MP or 8.3MP? Isn't 4K easier to remember and reference?

I'm the kind of person who is apt to say the glass is 2/5 full, or I'm feeling under the weather today at maybe only 70%. I of course am probably a unique development in the human species.

It's not just a fancy way of describing megapixel. 720p, 1080p and 4K all comply with the HD standard as defined by the SMPTE. Whereas MP just tells you how many pixels there are. As the HD standard increases in resolution (8K format already developed), broadcast, professional and perhaps video surveillance will adopt over non standard compliant sensors.

Hey Ethan,

Are 4K lenses available now? I didn't think they were out and if they aren't how big of an impact is this on your test results?

The Sony part number is IMX117. Keep in mind that low light performance is limited by pixel size. Sony has another 4K sensor for Industrial applications (security and machine vision) with larger pixels that would yield better low light performance. A faster lens would also help as was pointed out in the article.

Is Ambarella A7 capable of coding 4K?

Zhang, the A7 is spec'd at doing video encoding up to 10MP@10fps. So it's not 4K/30fps but the GoPro Hero3 spec's 4K/15fps.

Thanks, John.

This article is quite interesting and helpful. However, as some other reader pointed, the Hero3 is a consumer product, not an industrial level surveillance camera. Is it possible to have a close look at the Bosch 4k camera, which was announced in the ISCwest this year? That could give us more specific direction on the future of 4k IPC.

John Yong, if you get us a Bosch 4K camera will test it! ;)

We spoke with Bosch recently. They noted that the scheduled release date for their 4K camera has been pushed back to early 2014. We'll know more this fall both from Bosch and others who try to release 4K cameras.

This test was more to understand the fundamental technology. That said, consumer products, as many argue, tend to be leading indicators of what industrial surveillance gets a year or two later.

Good to know that. It looks like not all technical barriers have been lifted for 4K surveillance camera. 1. 4k lens; 2. 4k sensor with large pixel size; 3. 4k real-time compression with efficient encoding scheme (HEVC?); or instead using 6G-SDI or DP to transmit the video?

Will 4k industrial camera come to the market at the end of this year?

in survillance scenario, it's long way to go for 4K before it could solve the problem of storage, low light performance, low frame rate/long exposure time for rolling shutter sensor.

for narrow FOV, to prove 4K is better, should not you leave far away from the camera and zoom in the image to check if you can see the details, that'll tell the difference between 2M and 8M, am i wrong?

Christian, you say, "to prove 4K is better, should not you leave far away from the camera and zoom in the image to check if you can see the details, that'll tell the difference between 2M and 8M"

I am not sure if I am understanding you correctly. You want us to optically zoom? digitally zoom? Leave what far away from the camera? The test subject?

If you can clarify, we'd be happy to respond.

For John Z, I don't suspect 4K cameras are going to get HEVC/H.265 anytime soon. My gut feel is that they would start with H.264 just like 'regular' HD cameras do today. Of course, going to HEVC/H.265 would be harder for 4K than 1080p given its higher pixel count.

john,

sorry for not clarifying my concern clearly,

i mean let test subject be far away from the camera, and do digitally zoom.

About H.265 for 4K2K, we have no choice but using H.264 technology, because today's H.265 consumes far more hardware cycles than doing H.264, but seems nobody really achieves the expected encoding efficiency,

By using H.264? 4 times storage of using H.264 but most of the cases even worse than 2M performance in security scenario (like low light performance) , i believe there is market share but should not be so big(actually for traffic camera, since long time ago, 5M and 8M is used).

"I mean let test subject be far away from the camera, and do digitally zoom."

That's what we did here. We kept on moving the subject farther away from the camera and then digitally zooming in. For example, for the 160' wide FoV the subject really far from the camera and we zoomed in to see what details could be delivered.

I do not think 4K will be 4x the storage of 1080p. Many pixels will be of the same color / value so they can be compressed more heavily. This is not unique to 4K to 1080p. For instance, 720p is not typically 3x the storage of 480p / SD even though it has 3x the pixels (~300k vs ~900k). Ultimately, this depends on the amount of motion and complexity of the scene. I'd guess on average 2x more in storage for 4K vs 1080p (everything else equal).

Low light performance will likely be a significant issue. It will be interesting to see. If manufacturers are clever, they will immediately launch an integrated IR version to overcome this. But let's see!

In our experiment, HEVC could achieve 20%-30% improvement easily comparing to H.264, if not the 50% as claimed. Specifically, HEVC is suitable for 4K thanks to its bigger CTU size. The computation complexity could be a problem however we know HEVC encoding ASIC chip is under developed and will be delivered soon. I agree H.264 won't be replaced by HEVC in the IPC world in short term, but could it be used firstly in the 4K cameras?

For low light performance, in CCD domain, we have seen sensors with larger pixel size, e.g., sony ICX 814 (1''). There are also some other expensive global shuttle CMOS candidates we could choose. It is pricy though.

John, I do not know if there will be enough processing power to support encoding HEVC on a 4K camera. I assume there are cost as well as heat constraints on doing so.

As for larger imagers, you rarely see production surveillance cameras bigger than 1/2". As you mention, they are pricy, which is a strong barrier against it. I think integrated IR is a much better solution than larger imagers. Do you or anyone else see a barrier on combining 4K with integrated IR?

I heard TI claimed that HEVC has been implemented on their 667x DSP. The HEVC standard, like its predecessor, provides the tradeoff between its complexity and performance. So a balanced encoder could be achieved by a careful designer. Although I dont know the details of their HEVC encoder, I still believe the trend is promising.

The sensor of 1/2 bigger is rarely seen in regular surveillance camera, but not in the so called ITS (Intelligent Transport System) application. We have seen 5M pixel CCD network camera with 2/3'' CCD frequently used in the ITS system. As you said, they are not cheap, one piece is around $1500. But the market in some countries demands this.

I don't see any problem using an Integrated IR with 4K camera. However, IR cannot solve all the low luminance problems. In some surveillance applications, color information could be crucial. In addition, since 4K camera is a "Ultra HD" high-end product, many user could expect much higher image quality delivered, not just the image size, but also every other aspects. This could make the 4K camera design extreme difficult.

Color, night time vision is a niche. Most users are fine with black and white. If users need to pick between a $1000 4K camera with integrated IR and a $2000+ one with a much bigger sensor, low light color optimization, etc., most are going to pick the less expensive, integrated IR one (especially since overall the less expensive one with still deliver more details in b/w at very low lux levels).

I am sure there are chips that can do HEVC. Beyond price, the issue will be what frame rate can it support (full vs 15, 12, 10) and how many streams (just 1 or 2, 3, etc.). We'll likely see a pattern for HEVC similar to earlier H.264 cameras where streams and frame rate is limited.

What distance from the camera is the test subject with eye chart?

Sal, the distance from the camera ranges from 160' to 20', listed in descending order in the report above.

Those look like field of view width markings no? Am I missing something?

Sal, you're right, I am wrong. Yes, those are the FoV.

Ethan, what is the distance(s) from the camera?

Btw, it's a fixed focal lens so there's no adjustments to FoV/distance.

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

Genetec Privacy Protector Tested on Nov 12, 2018
Genetec has built Kiwi Security's Privacy Protector into Security Center, an analytic which anonymizes individuals in cameras' fields of view...
Dahua Dual Imager Dome Camera Tested (HDBW4231FN-E2-M) on Nov 07, 2018
Dahua has introduced a dual-imager dome model, the HDBW4231FN-E2-M, with two independently positionable sensors including integrated IR, not found...
Wyze Explosive Growth Disrupting Consumer IP Camera Market on Oct 30, 2018
Wyze, a company founded only in 2017, is poised to disrupt the consumer IP camera market by combining American marketing and Chinese manufacturing...
Axxon Face Search Tested on Oct 26, 2018
AxxonSoft has brought facial recognition to their Axxon Next VMS for free with the simply named Face Search, claiming to allow users to find...
Hanwha Dual Imager Dome Camera Tested (PNM-7000VD) on Oct 18, 2018
Hanwha has introduced their first dual-imager model, the PNM-7000VD, a twin 1080p model featuring independently positionable sensors and a snap-in...
ADI's Financials Revealed + W-Box Growth Priority on Oct 15, 2018
  ADI is one of the most powerful distributors in the security industry but how big are they? How much profit do they make? How much do they sell...
Dahua Face Recognition Camera Tested on Oct 15, 2018
Dahua has been one of the industry's most vocal proponents of the value that AI creates: As part of this, Dahua has released a facial...
Last Chance - October 2018 Camera Course on Oct 04, 2018
Today is the last day to register for the October 2018 Camera Course, register now. This is the only independent surveillance camera course,...
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...
Amazon Ring Spotlight Cam Tested on Sep 17, 2018
Amazon's Ring has released their latest camera entry, the Spotlight Cam, which we bought and tested in our Consumer IP Camera Analytics...

Most Recent Industry Reports

Ideal SecuriTest IP Vs Unbranded IP Camera Install Tools on Nov 21, 2018
In our recent IP camera installation tool shootout, multiple members questioned the Ideal SecuriTest IP's features compared to low-cost unbranded...
Intel Neural Compute Stick 2 / Movidius AI Test on Nov 21, 2018
AI is a major trend in video surveillance with manufacturers paying significant attention to Intel's Movidius Myriad chips. Indeed, Avigilon has...
Openpath Access Control Tested on Nov 20, 2018
Big investment in access startups is uncommon, but Openpath has recently attracted $20 million doing just that. The company has limited security...
No GDPR Penalties For UK Swann 'Spying Hack' on Nov 20, 2018
The UK’s data protection agency has closed its investigation into Infinova-owned Swann Security UK, the ICO confirmed to IPVM, deciding to take “no...
Milestone Disrupts Milestone With Arcules on Nov 19, 2018
Milestone is now competing against... Milestone's own spinout Arcules. New IPVM testing shows that Arcules has incorporated a substantial amount...
Pressure Mounts Against Dahua and Hikvision Xinjiang Business on Nov 19, 2018
Pressure is mounting against Hikvision, Dahua, and other companies operating in Xinjiang as an international outcry brews against the Chinese...
Arcules Cloud VMS Tested on Nov 19, 2018
Arcules is a big bet, or as they describe themselves a 'bold company', spun out and backed by Milestone and Canon.  But how good is Arcules cloud...
'Sticker' Surveillance Camera Developed (CSEM Witness) on Nov 16, 2018
The Swiss Center for Electronics and Microtechnology (CSEM) has announced what it calls the: world’s first fully autonomous camera that can be...
ISC East 2018 Mini-Show Final Report on Nov 16, 2018
This is our second (updated) and final show report from ISC East. ISC East, by its own admission, is not a national or international show, billed...
Facial Detection Tested on Nov 16, 2018
Facial detection and recognition are increasingly offered by video surveillance manufacturers. Facial detection detects faces in an image/video...

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