[Ryan Holmes]

[Walter Soyka] “transcode your source clips to DNxHD with mxf files”

Walter is exactly right. DNxHD wrapped as a MXF file is likely your best bet for an intermediate workflow on Windows.

Ryan Holmes
http://www.ryanholmes.me
@CutColorPost

[Ryan Holmes]

The reason it records that way is a limitation of the file system used on the CF cards (FAT32). So it has to break it up since the file system can’t understand a file being bigger than 2GB in size. Yes, it’s 2013 and we’re still dealing with this problem! Ridiculous IMHO.

Premiere tries to make this easy by reading the card structure and parsing the XML file that’s on the card. That XML tells a given piece of software (Prelude, Premiere, Media Encoder, Canon’s own XF Utility, etc.) where to find the clips, how to link them together, and how long each clip is. If that XML gets ignored, misinterpreted or corrupted then thing like what you see (clip 1 repeated for the duration of your interview) is what happens.

The best fix I’ve found is to use something else to ingest the footage and transcode it (ProRes or DNxHD are the most common mezzanine formats, and would mix seamlessly with your other footage). So you can try Prelude or directly into AME to transcode. If neither of those work, then you’ll need to get your hands on Canon’s XF Utility (available as a download from their site) and from within there you can spit out each MXF file as it’s own standalone file (meaning it doesn’t reference any XML file, it’s just a self-contained MXF that any system could read – Premiere, Avid, Smoke, etc.).

I’ve had this happen with our C300 more than once. Your data should still be in tact….you just need to quite relying on Premiere to parse it correctly. Sometimes Premiere just can’t seem to get this right. At that point transcode to an intermediate format or spit out standalone MXF’s and you’ll be back on the road.

Ryan Holmes
http://www.ryanholmes.me
@CutColorPost

[Ryan Holmes]

I haven’t seen many success stories taking a t3i and recording out via HDMI. Since the HDMI port on these cameras gets cropped significantly to 1620×1080 (3:2). The advantage of going out over the HDMI port compared to recording internally is minimal (unless you’re doing green-screen work then a 4:2:2 signal is better than the internal 4:2:0). Otherwise, you won’t see much difference for all your effort. So just stating that upfront.

A couple things to investigate: the Canon DSLR’s have notorious problems for outputting correctly over HDMI. The t2i, t3i, t4i don’t natively output 1080 as mentioned above over the HDMI port. I’m not familiar enough with ML hack on these cameras to verify that it does maintain a 1080p output. Or is it just stripping off the on-screen information but still outputting at 1620×1080? So are you sure that the signal coming out of the t3i HDMI port is in fact 1080p23.98? The best place to get help for this type of stuff is on Magic Lantern’s forums. https://www.magiclantern.fm/forum/index.php

The HDMI port may be bad but that’s very atypical. So can the Hyperdeck ingest other cameras that you have? If you don’t have another camera can you borrow someone’s camera and try it out?

Ryan Holmes
http://www.ryanholmes.me
@CutColorPost

[Ryan Holmes]

It always depends on budget! 😀 Working at a school myself I understand about requesting money and you often only get one crack at it!

Lens wise I think those are all good suggestions. If you can spring for L-glass I certainly would. I have the 50mm, 24-70mm, 70-200mm all L series and shoot them frequently. The optics are great on them and they perform as expected day in and day out.

I’m torn on follow-focus. It often depends on what type of shooting is happening. If it’s more scripted and you can follow the action they make sense. If you’re running and gunning sometimes they don’t make sense because the added setup time is too long. I have one from Cinevate that is great, but we often don’t use it just because of setup time (set the ring on the lens, adjust it, add rails to setup, rearrange setup on tripod or shoulder rig to accommodate follow-focus, etc.).

CF cards…I pretty much swear by Sandisk cards. In 5 years of using them they’ve never failed on me, ever. I would purchase the Sandisk Extreme Pro cards in either 32GB or 64GB flavors. However, if you’re thinking of running the Magic Lantern hack on the camera to shoot RAW video then these cards don’t have the necessary speed to write that data. The common consensus there is a company called KomputerBay which has 1000x cards….but they are notoriously inconsistent straight from the factory. If you’re not thinking of the ML hack, then stick with Sandisk.

As for bags, I use Think Tank because I have to travel with the gear and they do great for security screenings, overhead bins, or just pulling around on set – https://www.thinktankphoto.com/categories/rolling-camera-bags.aspx

Batteries are cheap for the 5DIII so stock up. I tend to only by the Canon brand ones as I’ve heard too many bad stories about third party batteries failing, not charging, failing to hold a charge. If you’re shooting or teaching people to shoot you can’t have your power source dying in the middle of a production. Pay the $59 and get the goods. I would also recommend additional chargers. The 5DIII only has a single battery charger so you’re recharging 1 at a time. Something like this lets you charge 2 at a time: https://www.amazon.com/PhotoPro-Charger-Canon-Camera-Batteries/dp/B007109ZA8/

My team never shoots a DSLR without having a bigger, easier to see screen attached. The on-board LCD is only 3-3.5″ so it can be difficult to determine focus, especially if you’re running your lens wide open. I recommend SmallHD screens as a necessary accessory for any DSLR shooter – https://www.smallhd.com

The DP4 is cheap and easily portable (runs off LP-E6 batteries just the 5DIII). Or you can go to the AC-7 and really pickup some real estate. I currently have 2 DP4’s, 1 DP6, 1 AC-7 and I love them all! 🙂

I’m assuming you’re getting rails as for this so you might consider some type of shoulder rig unless all you do is tripod type work. Having a way to hand hold it more steadily can be important. Or you can even get a monopod so the camera has some float to it but remains steady overall (much more than trying to handhold it).

Just my $.02.

Ryan Holmes
http://www.ryanholmes.me
@CutColorPost

[Ryan Holmes]

You probably mean IBP and IBBP….you can change the intraframe compression scheme in the dialog box in either Premiere or Media Encoder. Within the export dialog box scroll down and you should see a drop down that offers you different options for compression. Select the one that matches what you need to deliver. You can also change how often the I-frame appears (commonly 12 or 15 frames in NTSC lands, different if you’re delivering PAL).

I know this link if for Final Cut, but it does a really good job of explaining what GOP structure is and how the IBP/IBBP formats work: https://documentation.apple.com/en/finalcutpro/usermanual/index.html#chapter=C%26section=12%26tasks=true

Ryan Holmes
http://www.ryanholmes.me
@CutColorPost

[Ryan Holmes]

What version of Adobe are you running (CS5.5, CS6, CC?)? Up to date?

Are you pulling the files off the CF card before running them through Prelude (i.e. copying to a hard drive)? I’ve found problems with big batches if I attempt to access the CF card directly. Copying them onto a hard drive before transcoding seems to work better.

Have you tried skipping Prelude and going directly into AME or Premiere Pro with the MXF files? Technically, you don’t need to transcode to ProRes if you’re staying within the Adobe universe.

Ryan Holmes
http://www.ryanholmes.me
@CutColorPost

[Ryan Holmes]

What settings are you using to encode your file (resolution, frame rate, data rate, audio settings, etc.)?

You’re not the first person to complain about this with an Android device. Not every manufactured Android device has a dedicated h.264 decoder chip on it. So many have to rely on the CPU to decode and it is a resource intensive decode. You may have better luck transcoding to a WebM or Ogg Theora format for use on Android devices (again…depends on the device).

Not to make this an iPhone rally-cry post more to highlight a difference, but Apple in 2008/2009 started to embed a h.264 decoder processor on the main board of the iPhone. So instead of using CPU cycles to decode the video (battery intensive) the dedicated chip handles the h.264 decode. In that area the iPhone has a great advantage.

Ryan Holmes
http://www.ryanholmes.me
@CutColorPost

[Ryan Holmes]

In the future you’ll need to verify that the lens your buying will properly cover a full-frame sensor. The lens you bought is appropriate for a crop-sensor camera – like on a 60D, 70D, 7D, t4i, etc. you won’t have any problems. In post you could scale up the footage to get rid of the black ring.

Or there’s the reshoot filter! 😀

Ryan Holmes
http://www.ryanholmes.me
@CutColorPost

[Ryan Holmes]

[Suzin Daly] “On another note, I again thank you for your response and taking the time to give guidance. It is truly very much appreciated. “

No problem. Glad to help.

You may look into getting out of the Canon consumer DSLR world – t3i, t4i, etc. If you move up into their “professional” bodies I don’t think you’d have these problems….at least in my experience I haven’t had them. I use a Canon 7D and 5D Mark III. But there’s also the 6D (and maybe the 60/70D would work better). Just something to consider if you want to go further down this road….

Ryan Holmes
http://www.ryanholmes.me
@CutColorPost

[Ryan Holmes]

Suzin – You have to keep the CF card structure intact when you transfer it over to your hard drive. So you can’t just grab the .MOV files and expect FCP L&T to see them. You have to bring the whole card over with its various folders and files.

But again a program like MPEG Streamclip or 5DtoRGB will just grab the .MOV files and convert them to whatever you specify.

Ryan Holmes
http://www.ryanholmes.me
@CutColorPost