360 Video Test Using Ambisonics

This test was aimed at developing some ideas for using Ambisonics along with 360 video. That is,if you were really standing beneath Edinburgh Castle, you would hear sound based on the direction you were looking.

So, try this:

1. Put headphones on.
2. Watch the video, orienting yourself to look toward the castle.
3. The first bus that passes should sound as if it moves from your right to your left.
4. Now, replay the video facing AWAY from the castle.
5. This time, the passing bus should sound as if it is behind you.
6. This means you will hear it first in your left ear, moving to your right ear.
7. So, instead of stereo, this is decoding the sound based on where you are looking -- in theory, that's more immersive and similar to our real-world experience.

You might also try this in a headset https://youtu.be/NwcYpf6kQWA - but be sure to use headphones.

360 Camera Virtual Reality Test Shots

VID_20171212_120910_010_1 from Ted Fisher on Vimeo.

Here are three views of a location, shot in 360-degree virtual reality. You should be able to see this in a VR headset or just by holding your phone at arm's length. If you are seeing this in a browser you should be able to click and drag around, although that's not as much fun.

UPDATES: 1. This works in Firefox. On mobil phones, it doesn't seem to work in Safari. It works in the Vimeo app. It may be best to go straight to Vimeo: https://vimeo.com/246982292

The reason for shooting in this particular spot is that it is a real resolution challenge -- tons of detailed plants, detailed ground, etc. So far, I've seen the detail vary greatly in different platforms. I've tried it in a Google Cardboard with an iPhone in it, and that worked well. A bit immersive. And I've tried it with an iPhone 6 Plus held at arms length. It looked good, but that's a less immersive method. On a laptop, it works (as on this page viewed in Firefox) but if the goal is immersion, then the headset is the real target.

More experiments ahead. 

Timelapse Experiment

Cloudlapse from Ted Fisher on Vimeo.

I wanted to create a timelapse that also slowly zoomed out.

So, here's how I did it:

1. I shot stills at an interval of 2 seconds apart.  I shot for about 15 minutes.
2. These were then dropped into a 1080/30p sequence in Adobe Premiere. Each still became one frame in the video.
3. Since each frame was 2 seconds apart, one minute of shooting creates 30 frames or about 1 second of video. So - 1 minute of filming equals one second of timelapse video.
4. This sequence was exported as a QuickTime movie at 1920 by 1080, 30 frames per second.
5. The exported clip was imported back into Premiere and dropped into a 720p/30 sequence.
6. By using keyframes to control "scale" I was able to start with the clip at 100% scale -- larger than the 1280 by 720 frame of the video, and therefore "zoomed in" -- and then set a keyframe at 67% scale ("zoomed out"). This keyframed change in scale results in the feeling of zooming out.

Microphone Comparison

Microphone Comparison Test from Ted Fisher on Vimeo.

Above: a comparison of three microphones:

• an Audio-Technica ATR-3350 Lavalier
• a Rode Lavalier Microphone
• and an Azden Barrell Shotgun Mic With XLR Outputs

All three are presented as recorded with no changes in level or equalization. The ATR-3350 does very well (and sells for a very reasonable sub-$25 price) but the Rode delivers what I think is a much more pleasant tone for a documentary interview. The shotgun mic does well, and with careful microphone placement and adjustment of recording levels can do even better.

GH1 1080p Camera Test (After Firmware Hack)

Panasonic GH1 1080p Camera Test from Ted Fisher on Vimeo.

Previously, I posted a one-minute video showing a "Stress Test" of the Panasonic GH1 in video mode.

For that, I shot 720p / 60 footage in Central Park -- and found generally great results, but that the camera clearly had its limits. A sharply-focused shot of highly-detailed trees combined with camera motion could hit the limit of the codec -- resulting in "mud" in the shot. (That is, poorly-resolved digital smearing. Think of a low-resolution JPEG.)

On the plus side: most footage turned out great, and if you planned your shooting you could get excellent HD video without some of the issues (moire, for example, and false detail) that the Canon HDSLR cameras were reported to be experiencing. (And the Panasonic allowed twice as much footage per gigabyte of memory, had no need for a specialized add-on viewfinder, could autofocus during video, presented a live histogram -- and so on.)

Since moving to Los Angeles I've applied the supersecret firmware hack to the camera. (Google it. I'll post more on the details soon.) So I decided it was time to test out the 1080p / 24 mode as well.

For this shoot -- purposefully handheld, shaky, and shooting subjects that tended to bring out "mud" in the stock GH1 -- I used lpowell's "40Mbps AVCHD High Reliability Patch" settings, recording 1080p/24 AVCHD clips. I then processed the MTS files using Voltaic and edited those transcoded clips with Final Cut Pro. I exported a 1080p QuickTime file, then used compressor to create the 1280 by 720 file I uploaded to Vimeo.

My thoughts: the "40 reliability" settings are excellent in quality and stable enough for documentary shooting. As well, all clips play back in camera. Follow the link to Vimeo if you want to download the 720p version -- but trust me that the full 1080p file is even better.

Camera Test A

Panasonic GH1 "Stress Test" A from Ted Fisher on Vimeo.

My first session with the Panasonic GH1. The goal of the shoot was to find the camera's limits. Everything was shot handheld, seeking out conditions which might prove difficult on this type of hybrid system.

So I shot quite differently than one normally would. Rather than panning smoothly, avoiding shots with extreme or complex detail or aiming for the often-recommended shallow-depth-of-field look, I aimed to make the camera struggle.

I think it performed very well. The next test will include quite varied conditions and a different style of shooting. I'm hoping to do a low-light test as well.