Definition: "The JPEG-2000 image compression system has a rate-distortion advantage over the original JPEG. It allows extraction, different resolutions, pixel fidelities, regions of interest, components, and more, all from a single compressed bitstream. J 2000 also has options for lossy and lossless compression."

Well,… it has been a frustrating few days, to say the least, trying to get this JPEG-2000 sequence created. These complicated files are testing the limits of my computer and my patience. I’m fine if something just involves research and work and then even more research. However, when it moves into the realm of computer processing power and involves questions like, ‘Can my hard drive do the job based on the CPU levels required?’, and ‘Are there unresolved bugs in this software?’, then I start to get a bit… troubled.

Tried to create a J2C sequence using a bitrate of 125mb/s simply because that was the rate that had been suggested in the tutorial I was following. This somehow gave my computer or the OpenDCP program a world of headaches. It would stop transcoding as if it were finished. Thinking the process was complete, I’d check the specified folder and see that it was fine down to 8000 images, then it had made a fantastical leap to 100112. Only about 92000 images missing. So, I selectively tried to fill the gap of missing files. Oh, I wish it was that simple. It would only transcode another hundred, somehow randomly chosen by the program, even though I had clearly specified which files I wanted to transcode, then call it quits. Big breath required.

Anyway, after two days of continuing struggles and tests, (even went so far as to contact a lab that specializes in DCP creation), I learned that the issue may be the bitrate I was transcoding at - 125mb/s. Some research revealed that the developer of the program “OpenDCP”, Terrence Meiczimger, had actually gone as low as 20mb/s without image degradation. Other professionals had tested low bitrates and found that, for 2K, (JFG is 2K), 55mb/s was perfectly fine and would result in no quality loss. So, to be on the safe side, I decided to test it out at 75 mb/s. It was a 9 hour nail-biting process… but IT WORKED! Every single .tif file was transcoded to J2C. This was an accomplishment worthy of celebration. Instantly, I ran a test to compare quality to the original .tif file and was thrilled to see that there was no image loss whatsoever. Other than the color shift, which is apparantly perfectly normal.

Before I move to the next stage of this process I thought I’d write this blog, because sometimes the whole process is so complicated that my brain runs out of memory space and I forget how the hell I did something. Not that I hope to transcode again anytime soon, but it may be nice to relive this success.

The image below is the test comparison. The top frame is from the .tif file, Bottom frame is the J2C file.

DCP JOURNAL - TRANSCODING to JPEG2000