Controlware Communications Systems

You are here:Knowledge Base  » Security articles » H.264 compression benefits

We take a look at the benefits of H.264 compression and what it can do for you

Controlware Country Manager Mark Harraway was recently interviewed by International Security Buyers Guide (ISBG) about this exciting technology.

ISBG: Is there a 'horses for courses' argument in terms of the different algorithms given that they have been tested with varying aims and tailored to distinct applications? Is H.264 really so versatile as you believe?

In a word “Yes”. The official (snappy) titles for H.264 are  “ISO 14496-10:2009” or “MPEG4 AVC (Motion Picture Experts Group Phase 4  Advanced Video Coding)” or even - since it is now developed and maintained by the Joint Video Team of the ITU-T and MPEG -  “ITU-T H.264”.  Within the standard there are a number of different profile requirements on how the codec function should work depending on the nature of usage. This can be from very low-end applications where reduction of bandwidth demand is key (such as video to mobile phones) right through to high-definition ultra high-quality broadcasting.  Now, here’s the rub: each of these profiles will affect required processing power and therefore component and build costs. As a result, most manufacturers have only adopted certain profiles in their products with obviously divergent costs and functionality / quality and bandwidth ramifications. It is therefore difficult to be sure that you are comparing apples with apples as everything can still be said to be H.264-compliant but to which profile?

 

ISBG: As an H.264 adherent could you please fight its corner in terms of image degradation.

Again, this is all about how far you go with the compression or profile used.  Many vendors offer functionality within their products to support additional compression or bandwidth limitation. In this way they improve the tailoring of their product to specific projects.  For example, if you are using remote cameras over an ADSL line with a fixed maximum available bandwidth, then it could be acceptable to either drop frame rates or decrease image quality in order not to exceed the available bandwidth and avoid pixilation.

The H.264 standard sets the criteria for multi-picture inter-picture prediction which is an improvement over the previous standards of just examining the differences between one or two frames by sampling many more frames. If it is unlikely that certain images areas will change (fixed structures in the background for example), then bandwidth is not wasted by repeated transmission of redundant data that is in the frame already. This does mean that bandwidth can be very dynamic in locations with frequent motion or sudden scene changes as the algorithm has to deal with such variation. But theoretically, depending on the profile used, there should be no difference in image quality. And if the hardware encoder then makes a choice to drop frames or over-compress the data to meet a user defined requirement, the H.264 standard should not take the blame.

 

ISBG: Given the current state of video management systems, will  the marketplace  allow H.264 to be implemented widely?

Again, the answer is “Yes” if only because it is the stated mission of ONVIF to promote a single standard within the CCTV arena.  Many leading vendors have H.264 in their product sets and many VMS developers already support it.  The Internet protocol CCTV community is well aware of the drive towards standardisation and the need for it.  The key for all of us is the path towards  integration: If you are installing a complete new system into a green field site then I would unhesitatingly recommend opting for H.264 right now.  But if you have chosen a migration approach, then look to encoders to see you through until you replace  your cameras since it should be possible to support both H.264 and non-H.264 on a good VMS system.  But do remember that H.264 in the CCTV world will be driven as a hardware platform.  You will need either an H.264-enabled encoder / codec or camera to create the stream and then a VMS that is capable of recording H.264.

 

ISBG: Has the popularity of M-JPEG caused the CCTV sector to accept low frame rates as the norm? Is there any truth in the adage: “You don't miss what you’ve never had.”?

I wouldn’t say that M-JPEG is more popular than MPEG4 or H.264. Rather, it’s just that it is an older and easier format and therefore more convenient to adopt.  But then I think the key to this is that when MPEG-4 was introduced, the classic scenario occurred in that those who didn’t offer it tried to frighten the industry by saying that as the B or P frames in MPEG weren’t “complete” frames, this would render the footage inadmissible in court on the grounds that it had been “tampered” with. It’s a similar scenario to so many vendors at this year’s IFSEC saying “Go hybrid!” My contention is that they are only taking this position as they don’t have a comprehensive IP solution and are fudging the issue. 

What this has meant is that since M-JPEG makes big demands in both storage and bandwidth, frame rates have been sacrificed meaning that you could miss the crucial shot of a human face. By contrast, with H.264 you can now offer both high frame rate and high resolution very easily over the same bandwidth.  The issue of judicial admissibility of MPEG4 or H.264 has long been solved so I do see H.264 as the future of CCTV in general just as few would argue now that HD is the future of commercial television and movies.

 

ISBG: Sceptics allege that H.264 necessitates higher-powered processing and there are higher lag times. But elsewhere industry pundits argue that the exact opposite is the case. What is your position on this?

It is true that as you do more work at the encoding end to compress the stream. More processing power is required and this is why you are seeing manufacturers release new hardware platforms for H.264. But as with all hardware, the processing power of the chipsets is increasing exponentially so I don’t see why this should cause more lag on the network since you are reducing the transmitted bandwidth.  What may be happening is that certain manufacturers are using poor quality chipsets or are trying to cut costs generally. Similarly, such scare-mongering could be analogue adherents trying to discourage end-users from deploying H.264 yet again.

 

ISBG: H.264 has been described very patronisingly as 'promising', presumably by those in the analogue camp or supporters of other standards who don’t want this new offering  to prosper.  Are there parties in the industry who want to impede the progress of H.264 by ‘damning it with faint praise’?

Again, this is the “We can’t offer or work with it so let’s rubbish it or at the very least pour cold water over it” tactic. IP and H.264 are the future of CCTV and just as magazines are no longer produced with typewriters and the old ‘hot metal’ printing technology, end-users shouldn’t listen to entrenched analogue adherents when making decisions on the future of their systems.  I would be more than willing to place an H.264 offering up against a like-for-like analogue system in an independent test to prove its worth.

 

ISBG: What is more important for end-users, reductions in RAID demands or usable images with reduced bandwidth? How does H.264 fare on both, and are there variations within industry sectors?

Only an end-user can answer this since factors include a mix of budget, space, and quality but it is important to note that H.264 and IP give you the flexibility to tailor the system to meet on-the-ground needs. By contrast, in an analogue or DVR-based system, it is the hardware that dictates matters through supported frame rates or recording space in the box. The beauty of an IP / H.264 system is that you can be dynamic in the offering. After all, why record at a high frame rate or resolution when nothing is happening? The way forward must be to move to event or alarm-driven recording.  And if you really do need 24/7 recording then use the power of H.264 to reduce your bandwidth and thus your storage requirements.

 

ISBG: There is a rumour that H.264 may prove unsatisfactory in situations where low latency is important. Is there any truth in this?

Again, I would have to say it depends on which manufacturer you are using.  Latency is a function of the network – available bandwidth vs amount of traffic generated.  As we have already said, using H.264 should improve network latency as you are reducing the bandwidth on the network by pushing the work to the edge device.   If the concern is image delay (perhaps moving one’s hand in front of a camera and then seeing this on the screen), the issue comes down again to the truism that you get what you pay for. Higher build standard encoders or cameras will give quicker response times as they have faster processors and the manufacturers are likely to have implemented the profile better.  Again this can all be adjusted in the individual codec settings.