Shannon's Journal Discussion Thread, Vol. 5

[Shannon]

Shannon's quote "I have had to do, and then re-do, and sometimes re-do again, every part of the E5 build. It is pushing new boundaries, finding new ground, and forcing me to adapt new ways of doing things to keep up. Some of these steps take a long, long time to do. It feels like I am slogging through a hell of boredom working on this. And because it now must be built as an 80 minute track, it is taking a lot more disk space and time to process. Reading, writing, copying this much audio at a time is slow, even on my high end SSD. I feel like it will take forever. I was even working on and off yesterday on my day off. Can't wait for this to be finished.

At least I know better what the new limits and requirements are after this."


Wow!! So Shannon does this also Mean that these new things you are learrning and developing will go into stage 3 of UMS v2 and possible stage 4 as well? or does it have nothing to do with it what so ever? Because  all of this is so new,in teh build, expansion, etc.... its hard to know what goes where or what will go where.  these are the reasons Im asking.
sounds like some major R & D to Me.   thanks. Keith

It's not going to change anything for the UMS v2 build.  The configuration for that works - barely - in 60 minutes per track.  This program has a very different goal, with a very different set of sub-goals, which require a very different configuration.

As it turns out, some future titles will be do-able as 60 minutes per track, and others will need 80 minutes.  It will depend on the number of goals and sub-goals, and the resulting configuration balance.  E5 has multiple inter-supportive, but very different, sub-goals.  UMS has many more sub-goals, but they're much more similar.  That seems to make a big difference.

UMSv2 settings were set up before I even built Stage 1, and the whole was designed to be uniform from stage to stage to minimize the amount of time, work and space it requires to build, without giving up anything by doing so.  It is a multi-stage program; E5 is a single stage program.  Multi-stagers will be rare, if not extinct, in the future because of how difficult and complex it is to build this level of tech.  Even single stage programs are becoming so complex to build that it is requiring more time to perform QC and correctness checks, which require more time because there are more steps and more opportunities for errors to creep in.  

Fortunately, I've designed the build process such that if there is a failure in Step 1, it will make step 2-3-4 impossible to complete; and so on.  This allows me to catch any errors in the build process and correct them before the program goes out the door.

But for 5.75.7G we are still discovering the range limits based both on time, processing power, disk space, complexity and software limits.