I've been using the new iPi Recorder 2.0 beta (2.0.0.10) to convert raw 1.0 video for 2.0 use. First impression...very nice! The interface looks more polished and it's clear what the new tools do and how they work.

One bug found: when using the Converter mode, the Output directory is ignored and the new video is saved to the source directory. A minor bug since the software doesn't overwrite the original files. Details are described here: Conversion Bug

After converting several files, I viewed them in Recorder's new Player mode. The depth video look much nicer--the new colors make the depth range clearer than how the same data was depicted in iPi Studio 1.0. (I haven't seen at this looks in Studio 2.0 yet...will do so soon.) I love the various layout options and switches in Recorder 2.0, and the 'Hide Background' option is very useful for isolating and analyzing the quality of data during a shoot. This should make it easier to determine if a re-shoot is necessary long before getting the data into iPi Studio and tracking it. Kudos, guys!




I'll post more thoughts about iPi DMC 2.0 as I make time to use it this week. This is very exciting! :)

G.

_________________
Greenlaw
Artist/Partner - Little Green Dog | My Demo Reels (2013,) (2015,) (2017,) and (2019)

Watch a one minute excerpt on Vimeo now!

I noticed that the file size of the converted files is less than 1/3 the total size of the original data files--in one case, it was 1/4 the size. Wow! That's excellent! Apparently the savings comes from keying out the background in the depth data directly in the file format.

I haven't tried recording original 2.0 data yet but I'm assuming it uses this same BG keying process to the depth data to reduce the bandwidth while you're recording. Very clever!

G.

_________________
Greenlaw
Artist/Partner - Little Green Dog | My Demo Reels (2013,) (2015,) (2017,) and (2019)

Watch a one minute excerpt on Vimeo now!

This is just my initial impression but being able to analyze the video in the new Recorder 2.0 Player mode seems to make it clearer why the Studio 1.0 tracker had difficulty at certain points in the '3 Happy Cats' video. Of course, Studio 1.0 also had a capable player but with Recorder 2.0's player it's somehow easier to understand exactly what the issues are. (Note: this is just an impression...I haven't directly compared it yet.) Anyway, if this is what we see when recording original data in 2.0, it will definitely help in deciding what motions need to be re-shot during a session.

G.

_________________
Greenlaw
Artist/Partner - Little Green Dog | My Demo Reels (2013,) (2015,) (2017,) and (2019)

Watch a one minute excerpt on Vimeo now!

After my first go with ipi2, I agree this is a great improvement.
The interface, workflow and accuracy are all miles better.

If someone invents a hi res sensor by version 3, ipi will really rival the pro systems.

I use a dual K system, and this is the first time I've got the results I was really after.

The jitter removal seems to work better as well, my results are much smoother.

I like this v2 very much, so far great. The conversion of my .avi have been compressed dramatically eg form version 1 4gig reduced in version 2 to 265 meg, what a reduction, I have not tried using the recorder with my 2 kinects yet. It seems so much better. I can see that the ipi recorder is now only going to be used by ipi studio because of its format.

One last thing, previously in v1 studio the wrist did not turn, but when done in v2 the hands turned, v2 seems more sensitive (like it). Well done the ipisoft team :0)

Did you try the record-mode yet? What happened to me is that whenever I attempt to go to the record -mode, the software starts to "eat" a huge amount of resources (CPU, I/O and memory) and doesn't respond well. It is not completely "dead" but it doesn't do anything within a reasonable delay. I have the old iPi still installed since I need it constantly but somebody said elsewhere that they could co-exist. After some tweaking, I noticed that the freezing might be related to the "cameras in a row" -slider. Whenever I dare touch that, the software goes bonkers. If I leave that as it is, I can get to the next step. Which would be to evaluate... so I guess iPiRecorder v2 is not going to be free to use (like v1 has been).

The conversion, however, goes just fine. For PS3 -videos, I've converted 114 GB of six-camera videos which resulted to 79,2 GB (30% smaller) iPiVideo -format. For some of the early videos we had clear background at the end of the video instead of the default (2 seconds at the beginning). The conversion doesn't seem to allow this, but maybe it just doesn't compress those videos as well. You can see something of value with PS3 cameras and the new iPiRecorder as well by hiding the background. This is an educated guess, though. It would seem now that our open doorways are problematic. It wouldn't have been a big deal to close the doors during the recordings but we have never thought they could pose a problem.

Haven't installed the new iPiStudio yet, since I cannot risk the ongoing tracking process with the old version. We happened to record a lot of stuff last week. I'll probably have to try it at home and see how it works since I just cannot wait! I wish I could have noticed the new beta during the weekend.

Can you give us specs of your computer: CPU, memory, OS version? It may help in investigating performance issues.


The EVALUATE button just starts the background evaluation process (now it is handled by Recorder, not Studio). It is not related to software evaluation, as you could think. iPi Recorder will continue to be totally free for use. Maybe we should replace it with something like ESTIMATE or ANALYZE to avoid such confusion.


Yes, conversion can evaluate background only by the beginning of the video, as this was the most common pattern for making records for iPi Studio.


Open doorways were and will always be the cause for background noise. Regarding tracking in iPi Studio, there will be no more problems with them in v2 than in v1. Regarding size of records, more background noise leads to decreased compression ratio when the compression method called Background subtraction is used.


You have nothing to worry about, iPi Studio v2 installs side by side with v1, as well as iPi Recorder. So you can continue to make a production using v1 and give a try to v2 at the same time.

Or "Evaluate background"? Or if a single word is needed, how about PREPARE.

I noticed now that if I remove 3 cameras from the computer, so that 3 are left, the recorder works quite normally. With 4 cameras attached it is alrealy sluggish and with 5-6 almost unresponsive (but only if I touch that slider in question).

Here are the details of the computer:

Hi, ppe

Do you experience the same performance issues with the last release 2.0.0.12 of iPi Recorder?

Nope. Whatever the problem was, it is now gone.