Hikvision H.264+ Tested
** '*******' *.*** **** ** ** a ***** ** *** ****?
**** ****, **** ******* *** '*****' H.264 ***** **** ************** *******. ***** **** ****, ********* ********* ***** own ********, ***** *.***+.
**** ***** '***** ******' ** ** *** ** *** most ********** *** ************ ** *** market *****.
*** **** **** *********'* *.***+ ****? IPVM ******** * ********* ******* (*** DS-2CD4024FWD-A, **-**********-** *** *** **-***********-*) ** H.264+. ** **** ****** **** ******* 2 **** ********* ******* (*********** ********).
** *** *****:
- ****** ***
- ****** ****
- ******* ********
- ******* *******
** **** ******** *********, ***** ******* of ***** * ******* ** **** as ****.*** ********* ****** ******** ** ********* **** impact *********'* *.***+ ***** ****.
*******
** *** *****, ********* *.***+ ************* reduced ********* ** *** ****** (**-**% reductions **** ******), **** *** ****** indoor ** **** ****** ******** ****** of ****, **** ******* ******* ******, in **** ****** ******** **** *********'* low ********.
********, **** *.***+ *** ********* ** the ******** ** *** ********* ****, including *** *** **** ******, ****** well ***** *** ***** ********* **** Zipstream ******, ********* ****** ***** ********** to **** *** *** *** ** the ******.
*.***+ ** ****** ** **** * very ******** ****** ** *********'* *********** position, ******* ** *** **** ** takes *** *********** ** ******* **** their *** ***** ***** *********. **** Axis *** ********* **** ******** ********** bandwidth *******, ** ******* *** ************* need ** ********** ********* ***** *** smart ***** *******.
H.265 ******
**** *********'* *.***+ *** **** ********* delivered *** **** ********* ******* **** H.265, ** ***** ** *** *.*** vs *.*** ********. ** ****, ** are ************ ********* ***** *** ***** term ****** ** *.*** ***** *** bandwidth ******* ** '*****' *.*** ******** with *** ******* ** **** **** existing *****.
Camera *************
*.***+ ** ********* ** ****** ********* cameras ******** ** ******** *.*.*, ********* from ********* ******* [**** ** ****** available]. **** **** ********* ******* ********** support ********* ******** ********, **** ***** America ******* *.*.* ** ******, ***** Europe *** **** ***** *.*.* *** 5.3.6 *** ********. **** ******, ***** should ******* ***** ******* ** **** the ****** ******* *** ***** ******.
*** ******* * *** * ****** Hikvision ******* ******* *.***+. ** ** supported ** *** * ****** *** **** cameras, ***** ******** *** *** ****** *** ************** ** ** ****. ***** * series ****** ** *** ******* **.
**** ***** **** ********* ******* ***** in *.***+ *** ******** ** **** low ****, ~$*** *** *** ******. Compared ** **** *********, ***** ** supported ** ****** **** *** ******, with *** ********* ** *** *** range ********* ******(~$***-***), ***** ****** *** ************* **** *********.
Limited *************
*********'* *.***+ ******** *** ************* ********, with *********** **** * ****** **/*** control *** ********* *********. ** ******* of ******* *-***** ******** ** *********** are ********, ****** **** ********* ***** includes ******** ******** *** **** ** these.
*** **** ************ ********* ********** **** using *.***+ ** "******* ******* *******", which ******** ** **** ******* ***** a *** ***** **** ** ******** period ** ****. *********'* ************* ****** at ***** ***** **** *** ******** for *** ****** ** ******** ****** to **** *******.
** ****** ***** ******* ************* ******* in **** *****:
Stream ********
** *** ***** ***** ** ****** stream ******** ** *.***+ *** **** Zipstream ***** ***** ******* ********* ********.
*******, ******* ************ ***** *.***+ ** Hikvision ****** *** ~** ** ** FPS. **** ******** ***** ***. *******, Hikvision ****** * ****** **** **** widely, **** ~** ** **, ***** Axis ****** **** ~**-**.
**** **** *** ********* ******** *-***** intervals ** ***** ** ******* *** in *** ***** ** ********.
Bitrate ***********
** ******** ******** ** *** ******* with ***** ****** ** *** *** in ******** ******:
****** ******
*****, *******, ***** ******* ** *** framerate, ******* *.***+ ** ******* ******** of *** ********* ******* ** ** average ** ***** **%. *******, **** at ***** ******, ******* ** ***** roughly ******* ** *** **** ***** with ********* ***. ******* ********* ** (****, *** ****), the *****'* ******* ** ***** **% lower **** *** ********* *******.
** ***** ** *** **** *****, H.264+ ******** *** ***** ************* ***** (>80%), *** **** ***** *** ********* **** and **** *** ************* ****** **** the **** *****. *******, *** ** its *****-** **, *** **** *** the ****** ******** ** **** *****, about **% ***** **** *** *****.
********** ********* ** ** ***, *.***+ (and *********) ******** *********. ********* ******** dropped ** ** ******* ** ~**%, as *** ********* ** *** *****. The **** *** *** ****** ** the ********* ******, ******** ***** **** the **** ******.
******* ******
****, ** ****** ********, ***** * treeline **** ****** ******* *** ******* in *** ***. ** **** *****, H.264+ ******* *********, ********* ~**%, ***** Zipstream ********* ******** ** ~**%.
** **** *****, *********'* **** *** 4132 **** **** ******** ***** **** the **** *****, ***** *** **** was ********** ******.
*******, ** ****** ** * **** road *** **** **** ******* ** motion. *.***+ ****** *** ********, ********* just **** **%. *** *****, *** Axis ****** **** ********* *** ******* lower, **** **** *** **** *********** its ******* **** *.***+ **.
Image ******* *******
*.***+ ****** ** ***** ******* *** very ******, **** ******* ** **** fine ******* ** *** *********** ** tested ******, **** ** *** ******* in *** ********* *****. **** *.***+ off *** ******* ** **** *******, with ********** ******** *** ****** *******. Turning *.***+ **, *** ****** ** the ***** ******* ********** ******* *** hazy, **** ******* **** ******* *** brighter **********, ** ****.
*******, **** **** ***** *********** *** minor, *** ** *** ** ********* issues ** *** ***** ** ****** objects **** ** ****** ** ********.
Test **********
*** ******* **** ****** ***** ******* settings ****** ***** ********, ********** ** shown *****.
******** ******** **** *** **** ****:
- **** *****: *.**.*.*
- **** *****: *.**.*.*
- ********* **-**********-*: *.*.*
- ********* **-**********-**: *.*.*
- ********* **-***********-*: *.*.*
Lew, do you recall the model numbers? And that is for H.265 or H.264+ (or as LTS labels its H.264 Zip+)?
Was there any increased CPU usage on the client? Does choosing H.264+ have any effect on the cameras ability to deliver multiple streams or higher frame rates?
No increase in CPU.
Turning H.264+ on in the 4 and 6 series reduces multistreaming to two streams instead of three, and you can no longer get 60 FPS in cameras which support it (e.g., the 6026). But we saw no issues with 1080p30 streaming being full frame rate, nor 3MP/20 FPS in the 4132.
A methodology question:
In the H.265 tests, the h.265 camera were compared to mainstream h.264 cameras from different models and from other brands. The key conclusions of the report came from this comparison, not from the cameras own h.264 performance vs h.265 performance:
Net / net, these two H.265 cameras offered minimal to no bandwidth savings compared to modern / current H.264 cameras.
explaining
We do not emphasize [intra-camera results] because it does not really impact the general question of whether H.265 is 'better' than H.264.
Though in this tests there are no comparison to mainstream h.264. We only have comparisons among h.264+ capable cameras, switching between h.264 and h.264+ of the same camera to draw conclusions.
This would seem to set the bar higher for the h.265 cameras, then h.264+ cameras, no?
"We only have comparisons among h.264+ capable cameras, switching between h.264 and h.264+ of the same camera to draw conclusions."
The smart codec bandwidth results (Hikvision and Axis) are incredibly low compared to 'regular' H.264 cameras.
So for now it seems H265 has a competitor to "enhanced" H264 methods, but wasn't one benefit of H265 supposed to be you only needed one stream from which you could pull different resolutions, such as high-res for recording and low-res for viewing?
No. I've never heard anyone even claim that before. The 'pulling' different resolutions is a feature of scalable video codecs and H.265 being shipped / implemented is not scalable, just like 'regular' H.264 is not.
No. I've never heard anyone even claim that before. The 'pulling' different resolutions is a feature of scalable video codecs and H.265 being shipped / implemented is not scalable, just like 'regular' H.264 is not.
To be clear, a few manufacturers claim to use scalable h.264 today, as well as supporting 'regular' h.265, so there is reason to believe that they will implement it in h.265 at some point.
Also, amberella released a chip a couple years ago that had h.265 and svc on it, don't know if that ever ended up in any products.
To be clear, no one has claimed to implement scalable H.265 for IP cameras or recorders. And, in terms of models shipping, about 0.1% or less claim to be supporting scalable H.264.
Scalable video codecs have theoretical benefits, they just have never gained enough practical implementation to be a factor.
...about 0.1% or less claim to be supporting scalable h.264
Doesn't at least one of the Hikvision cameras from the test have H.264/SVC?
I think the Smart Series has has had it for a while now.
To be clear, no one has claimed to implement scalable H.265 for IP cameras or recorders.
Aventura claims a lot of things for a long time but there is no evidence any of it is for real. +1 for you on the technicality but this does not change the fundamental market situation.
Lew, thanks for the feedback. I am curious what the upgrade delivers. I have not heard of any existing NVR appliances being field upgradeable to H.265. I am sure it's possible but generally the increased processing needed to do so makes it impractical for existing units.
I recently installed 4 of the new 4MP value plus cameras and enabled h.264+. I installed them on a Milestone system and noticed a significant delay in the video streams "starting" in the viewing client versus turning h.264+ off. Anyone else try this out in Milestone and experience anything similar? Running the 2016 Smart client on a Core i5 with HW Acceleration enabled. I ended up disabling h264+ for the time being due to the delay being pretty bad at times up to 30 seconds or so when double clicking on a camera to enlarge then going back to a multi-view. CPU wasn't high just a long delay...
Hi Michael, I didn't see that in my tests, but we haven't updated to 2016 yet, and didn't have the firmware update for the 2 series when we began testing. I'll update both and report back.
Michael,
"Running the 2016 Smart client on a Core i5 with HW Acceleration enabled."
Can you try to see what the difference is between HW acceleration enabled vs disabled? We are discussing this with Milestone and that was one element they suggested we consider.
Finally got around to testing this. The delay is still present with HW acceleration on and off. I've got these h264 plus cams on a few different systems and all of them still have the same results. They are also noticeably slower to appear in milestone mobile.
The cause of this is probably that the client will have to wait for an I-frame before it can start decoding the rest of the frames. Smart codecs do have higher GOPs when there is no motion on the camera, which means a potentially long wait for the first I-frame. If you generate a lot of motion e.g. aggressively wave your hand in front of the camera, is the delay nearly as long?
Thanks for the test and report. H.264+ turns out to be useful and probably more and more low cost H.264+ cameras will come out soon.
It's very funny to compare h.264+ with h.265, since the same technology that h.264+ used can be used for h.265 as well.
Agreed, it can be, and some manufacturers are already using smart CODECs with H.265 (such as Vivotek, see our H.265 IP Cameras Test).
The point in comparing it here is that H.264+ clearly has significant benefits over H.265, without the additional complications of H.265, like limited VMS support, increased processor usage, etc.
When H.265 is mainstream and smart CODECs are using it, yes, things may be lower still, but I personally don't see that widespread acceptance happening in 2016.
...but I personally don't see that widespread acceptance happening in 2016.
If by widespread acceptence you mean the Western VMS manufacturers will have comprehensive connectivity to h.265 cameras, I agree with you, it ain't going to happen this year.
On the other hand, if widespread acceptence is measured by the number of cameras and matched NVRs that will be actually using the technology, I think it will be substantial.
Why? Looking at the new models coming out of both Chinese republics, they seem to be mostly h.265. At this point, it probably doesn't cost them much more to use that chip, so I would expect by year's end that pretty much EVERY device they make to be h.265 capable.
Costco's already selling h.265 NVR's!
When this next wave of h.265 cameras hits big box, usage will depend not on people's acceptence of h.265, but only on the way the NVR and cameras communicate by default.
This, of course, wasn't the way it was planned. H.265 should have been an elite feature on high-end, high-processing cameras connected to high-performance workstations. But??
But regardless of why that didn't happen and regardless of how good it is or not, by the end of the year, I would expect the low-end IP market to be primarily h.265 capable.
IMVHO.
"Looking at the new models coming out of both Chinese republics, they seem to be mostly h.265."
More specifics please. The pattern that I see is that H.265 is or was a hot spec item inside of China but that the export / overseas market is still firmly H.264.
The fact that H.265 is more advanced does not mean it has to be successful in business. The same fact happened to MPEG-2, which had been used in the broadcasting industry for super long time, even if MPEG-4 was proved to be much more efficient in coding than MPEG-2.
If no big players like Google/Apple/Microsoft are embracing H.264/HEVC, and making it to be a successful market standard, then I don't see HEVC will soon be widely accepted, and the licensing issue of HEVC may not be quickly solved either.
The history why H.264 license was made to very low or free, was because Microsoft invented VC-1 to be competitor and made H.264 to be at a difficult position. After H.264 becomes cheap or free, it was soon supported by all popular operating systems, Windows, MacOS/iOS and Android.
Because H.265 has no competitor right now, it's hard to see H.265 license issue to be solved soon. so H.265 may be only limited, and use in some 4K or above 4K video only. also H.265 in surveillance market may take a few years to grow to bigger size, that's why H.264+ can be more useful than we thought.