|
Mike, there are quite a number of additional considerations, but a lot of it is really down to 'gut feeling' which can only really be learnt through experience and experimentation.
Your target audience is home broadband, but broadband covers quite a range of bandwidths, over here in the UK it ranges from around 250kbps to 8Mbps with varying contention ratios, so you may want to consider encoding multiple streams to allow the lower bandwidth users to get a smaller/lower frame rate stream whilst also giving the higher bandwidth users a better experience. If you are unicasting though, then depending on the number of simultaneous users connecting, then you may find that your 3Mbps pipe soon reaches its limit.
If it's a live webcast, then you will find increased limitations as you can't use multi-pass encoding, or change your subject set-up to allow for optimal encoding, and it's a more a case of 'driving by the seat of your pants'.
The subject content is also a consideration, as it depends on whether the emphasis is on the quality perceived by the end user is biased to audio or video, and if it's video, you additionally have to decide over smooth motion (for motion) or sharpness (for stills).
As you have probably already experienced, it's all a number of interconnected trade-offs, whereby improving one aspect usually results in a degrading of another. Also, each end user will perceive their experience differently, both due to technical constraints and to expectations.
I've always tended to ignore pre-supplied profiles and created them from the ground up for each event, as it helps one to think through what it is that you are doing and relate that to your audience. I tend to use the Windows Media 9 CODEC's.
The proof of the pudding however is in the eating, so the sooner you start experimenting and evaluating the end result, the sooner you will start to get a feel for what will give you obtainable results.
Dave
|