To H.265 or not to H.265

Disclaimer: I am not a technical expert on the finer details of video codecs, nor I have studied H.265 in depth (yet), so I may be off on a few things here. Feel free to comment and correct me if I am wrong.

A tender was released sometime back for a city surveillance project. The project required a few 1000's of cameras integrated into a state of the art command and control center, mobile surveillance vans, etc. This was probably the first of such a tender I had seen with the following requirement under the CCTV technical specification "Video Compression Standard required - H.265". Though, at first I did think it was a typo (did he mean H.264) - obviously majority of the reputed manufacturers would not meet this requirement today. I mean there were one or two OEMs who had products out but it was too early to just put H.265, it wasn't mainstream yet. I was pretty sure that it would be changed to read "H.264 or better" in the first corrigendum.

As expected almost every OEM argued that since H.265 is relatively very new and not yet fully implemented industry wide and requested for the following change - "Video Compression Standard required - H.264 or H.265". This would allow fair competition amongst reputed manufacturers eventually benefiting the project.

The consultant/customer were convinced that the system they designed was 'better' and "using the latest technology" and the manufactures who objected (complained) were behind the curve. One even said this was "forward thinking, so the state will save a lot of money in the future". Surprisingly the corrigendum's released thereafter had no changes regarding this.

Realizing this could be a missed opportunity, sentiments then changed to "Oh they are just favoring a particular brand, everything is pre-decided" or slave driving the engineering team to get a working demo out. Fair enough - it could take a year or more to close the tender so they have time to have a proper release. I didn't bother. Other projects such as this took 3-4 RFP's. Governments have changed from the time of the first RFP to the last one :) and I was pretty sure this would be a waste of time now.

Update - Only two bidders eventually qualified technically (They had cameras using the H.265 codec). Bid evaluation is in progress and the consultant/tendering authority are confident they will award it to one of them. Sure, lets see.

Over the last few months, I have started hearing this from multiple customers, consultants, etc. - "We are planning to put H.265 in the technical specifications of our next project" or "When are you planning to implement H.265 in your product range". But when I heard it from my most favorite customer, I had to study it a little better.

Frankly there was not a lot of information available to assist an end user. I mean it was either too technical (the actual spec, comparative matrices) or plain marketing rubbish (Guaranteed 50% savings!) but nothing that would help an end-user to answer the question 'do we need it now or later'. Here's what I think.

Firstly, lets understand this,

1. H.264 and H.265 are examples of Video Coding Formats or Video Compression Formats. These are documented by a detailed technical specification document known as a Video Coding Specification. Some such specifications are written and approved by standardization organizations thus becoming the Video Coding Standard.
2. A specific software or hardware implementation capable of video compression and/or decompression to/from a specific video coding format is called the Video Codec.
3. Now, here is where it gets interesting "The Video Coding Format will only define or support what type of compression and only define its parameters for standardization. It does NOT define the actual method of compression (which algorithm to use for the actual compressing of the data). That is left to the implementor of the codec. So as a manufacturer - I will use H.264 as my guideline standard - but I can choose any number of available compression techniques, or make my own.

Confused ? Ok try this "Imagine there are two manufacturers A and B. Both will mention "H.264" on their data sheet as the compression standard. Considering all other factors identical, record 10 minutes of video at 1080p @ 25fps at Variable Bit Rate. When you compare the results (e.g bandwidth required by camera A vs. camera B, storage required, image quality, average fps, latency, etc.) the results will be different.

Why ? Because it all really depends on how the software engineers have optimized their compression algorithms. Benefits of such standards will be different depending on how they have been implemented. (I can go on and on about how this relates to the actual selling price of a camera but lets leave that for another time)

Anyway - you can think of a standard as a recipe. The recipe for butter chicken is the same, yet it would taste different when cooked by different cooks. This is because, while the recipe (standard) lists the same identical basic ingredients and how to cook it - the eventual end result (taste) will depend on the individual cooks caliber, style and experience in cooking. If you order butter chicken from the menu - don't expect it to taste the same everywhere.

From what I have understood so far, by design itself, I agree H.265 is significantly better than H.264 in terms of bandwidth consumption, storage required (promises a 50% reduction for the same image quality). How I utilize it to its best potential - that is entirely up to me.

Note to the consultant above - I apologize, theoretically you are correct but practically, we will just have to see.

Here's why I would wait for sometime before I put it in a specification today.

Lets start with the cameras - The two CCTV manufacturers who did end up getting technically qualified (i.e. had H.265) are what I call 'box-pushers' (and usually are from our north eastern friendly neighbor). They see a camera as a commodity (nothing wrong in that - there is a market for it). They have a relatively fast turn-around time, keep up with the latest (even upcoming) technical trends and are usually first to market with some of their products. Their high volume production allows them to play well in markets focused on price (L1). Product quality is great for generic, low-end applications - not for critical security applications for sure.

Legacy companies who are experts in video surveillance (the big boys, the real industry drivers) - usually never jump in early without spending considerable time/money in R&D and their overall strategy and future plan. It requires significant planning, engineering, development and testing. They usually do no push out a product without all of the above the above to be 'first to market'. They focus on their product quality, have a much smaller range than the box-pushers, do a few things - but do them well. If you are designing a tender for a critical security application (City Surveillance, Airport, Refineries, etc.) use them. You may pay a little more at the start, but during the long run - they have a better ROI.

Software for Video Management - For a video codec to be implemented both the encoder (CCTV) and decoder (VMS) will have to be use the same codec. So a camera using a H.265 encoder will have to use a VMS to correctly decode the stream.

Video Management Software used in such projects are highly specialized, scalable and focused on the critical security market. Box-pushers usually ship with what I would describe as a low-end, personal use VMS, It can never compete with the software from the experts and definitely never scale up to meet the software requirements of a city surveillance tender. The integrator will have to have use a multi-vendor approach right from the cameras to the VMS and integrate it into the PSIM at the control room. So unless the standard is widely accepted, tested and used by majority of the VMS experts - it wouldn't put it in a specification. (Shameless Plug No. 1 - Always bet on a single vendor end-to-end solution. We take the responsibility from the camera to the software and do not play the blame game).

(Note: Except one or two - I am not sure if any of the popular, mainstream VMS's or PSIMS are shipping a stable release of H.265 implementation today I may be wrong so let me know if you are aware of any).

Cost implication of moving from H.264 to H.265 ?

Another misleading statement I have read and heard is "Because H.265 will need half the bandwidth and about half the storage for the same quality of video - It will save money?" - No.

Think about it, when H.264 started replacing MPEG4 as the main compression standard in CCTV systems, it was predominantly also due to the rising popularity, demand and greater adaptation of HD video (720p/1080p). Along with this, there was an entire shift in the entire ecosystem - you needed powerful graphics cards, faster processors, better monitors. You would just not swap your system from what it is just to use H.264, you would upgrade it entirely to view or record HD video.

Compare this to the cable TV/Set top box industry. We are all seeing ads constantly about Cable TV providers (Tata, Reliance, etc.) providing HD Set Top boxes - more so today than three years back. No point getting one if your television is still old and does not support HD. Makes sense if you have a HD TV already, makes sense if your existing old television set has lived its life and needs replacement - then buy a HD TV because the industry has standardized it and is making content for it.

If you really want to effectively utilize the benefits of H.265, you should plan for an entire upgrade which will mean upgrading your hardware (cameras, workstation clients, servers, monitors, etc.) and your software. By swapping one or two pieces - you will never be utilizing it to its fullest extent. Sometimes it doesn't make sense to throw out something that works and is serving your requirements at the moment. Ride the technology curve with the industry leaders, they know what they are doing.

I would wait till 4K is more mainstream, common and I can choose products from multiple vendors. If one or two people are selling a next generation technology today, personally I would wait to see how it pans out before betting on it.

To conclude

I think the tipping point of cross-over between H.264 and H.265 as a globally accepted (implemented) standard in mainstream products is still is sometime away.

2. Thinking that using H.265 vs. H.264 will guarantee the benefits as described (or even selling it) is not correct. It all depends on how the manufacturer has implemented the standard and where it fits in the scheme of things. Products which are usually 'first to market' do end up with a lot more glitches than products that come in later once a technology is more standardized and stable.

Shameless Plug No. 2 - If you agree to my critique above but would also like to get the benefits of H.265 today (view and record a great quality image, consume lower bandwidth, reduce the amount of storage required and get a better ROI as compared to any other product or solution in the market) then I suggest you look at the end-to-end solution offered by IndigoVision. With features such as ACF on the edge, lowest latency, guaranteed no frame drops, an open system with no-single point of failure - you can future-proof your investment starting now.