As you probably are aware of by now, the upcoming April firmware upgrade for the Canon 5D MK III will enable clean HDMI output from their flagship HDSLR.
In anticipation of this release, we examined four External Video Recorders, each of which offer different capabilities at multiple price points in order to identify the most appropriate product for a 5D MK III HDSLR post production workflow.
The four External Video Recorders reviewed were:
Sound Devices PIX 220i ($2395)
AJA Ki Pro Mini ($1695)
Atomos Ninja 2($995)
Blackmagic Design HyperDeck Shuttle 2($345)
While these four represent a fairly broad spectrum of devices on the lower end of the pricing scale, there are other alternatives on the market. One of the higher end devices is Convergent Design’s Gemini 4:4:4 model, with a similarly high end price tag of $6000. Add another $2k and the unit becomes a functioning 3-D Recorder.
Although there are many factors one should take into consideration (other than price) prior to purchasing a Recorder, this discussion would be incomplete without reviewing some of the principal issues at hand:
Compression
Obviously, there are multiple positive aspects in recording video externally from your HDSLR; Most notably, the ability to record in an edit friendly, quality codec best suited for your needs is certainly at the top of the list.
According to Wikipedia:
“In the field of video compression a video frame is compressed using different algorithms with different advantages and disadvantages, centered mainly around amount of data compression.”
The Canon 5D MK II offered only the single IPP “Intraframe” method, where each group of 15 frames (at 30 fps) is stored in what is known as a Group of Pictures (GOP).
The 5D MK III changed that approach and instead now offers the “Interframe” IPB and the ALL-I formats. In both codecs, the video file is wrapped within a H.264 Quicktime (.mov) container in 4:2:0 color space.
Unlike the IPP method, IPB (“B” in IPB stands for Bidirectional compression) is designed around “frame predictors”; In other words, the prediction of the content of future frames — with reference to both previously captured frames and subsequent frames — resulting in a more efficient compression format.
In the ALL-I or “Intra-coded” Frame method, each frame is compressed and treated as a separate, single image.
There are benefits to using either IPB or ALL-I compression configurations. For one, IPB results in a file size of approximately 1/3 of an ALL-I file and is most suitable for static shots, such as locked-down, documentary style interviews.
ALL-I compression lends itself to faster editing on the typical NLE compared to the IPB format as it requires less processing power … not to mention the fact that it replays smoother on lower end computers.
ALL-I is best suited for is high-end editing systems with higher storage capacities, as well as users looking for the highest quality possible.
IPB compression: 1080p results in approximately 13 minutes of video for 4GB file
ALL-I compression: 1080p results in approximately 5 minutes (for each 4GB file)
Courtesy Sareesh Sudakaren
However, when choosing to record a video signal on an External Recorder, compression alternatives are usually greatly increased (depending upon the Recorder), such as:
• ProRes 4444 – 330Mbps: Faithfully preserves motion images sequences originating in either 4:4:4 RGB or YCbCr color spaces
• ProRes HQ – 220Mbps: Visually lossless preservation (through multiple generations of encoding and decoding) of the highest quality professional HD video that a single link HD/SDI signal can carry
• ProRes 422 – 150Mbps: Offers nearly all the benefits of “HQ” but at a significantly lower data rate
• ProRes LT – 100Mbps: Ideal for productions which require a large amount of video storage
• Avid DNxHD – 220X: (220Mbps 10-bit 1920×1080 30fps) Superior quality in YCbCr color spaces
• Avid DNxHD – 220: (36 Mbps 8-bit) Highest quality in 8-bit color sources
• Avid DNxHD – 145: (36 Mbps 8-bit) High quality mastering when using lower 8-bit data sources
[Note: Data rate is dependent upon image size, complexity, frame rate and quality settings]
Some may argue a significant increase in image quality when recording with one of the above codecs compared to the native ALL-I or IPB formats, but in reality the improvement is ultimately dependent upon the intended workflow.
Typically, if the footage will be used for the purpose of multi-pass VFX or green screen chroma key production, a high quality intermediate codec would be beneficial.
Apple ProRes 422 (HQ) typically offers more robust footage which is better suited in productions where heavy compositing or color correction will be performed in post.
In most cases, I-frame codecs such as Apple ProRes generally produce better results at data rates of 100Mbps and higher. Apple ProRes HQ (as well as 422) are typically classified as “visually lossless”.
In general, the more frames tends to change, the higher the data rate that most likely would be required in order to faithfully reproduce the scene.
Record Duration
One of the current limitations of HDSLR’s is the maximum file length of :30 minutes. This artificial ceiling of duration was established by the World Trade Organization (WTO).
In essence, it requires HDSLR manufacturers to pay a surcharge on devices which record in excess of :30 minutes.
Nikon and Canon, for example, have avoided the tariff by restricting their HDSLR’s to stop recording at :29 minutes, 59 seconds. As a result, they sidestep a fairly significant 5.4% tax. However, if the WTO’s Information Technology Agreement is revised (a proposal which is currently under consideration), this problem would go away.
Subsequently, abolishing this tax would prompt camera manufacturers to eliminate any length restriction which they currently apply to their HDSLR’s.
In any event, an External Video Recorder workflow makes this a non-issue, since video files are not internally captured by the HDSLR but rather on the external device (not limited by the camera’s configuration).
Media
External Video Recorders also pave the way for longer record times by virtue of higher capacity and cheaper media.
Both Hard Disc Drives (HDD) and Solid Sate Drives (SDD) are available in larger storage capacities, usually at more cost efficient scales (per gigabyte) compared to high capacity, high definition video qualified Compact Flash (CF) or Secure Digital (SD) cards used in HDSLR’s.
In addition, the necessary process of transferring (and sometimes transcoding) files from internal SD or CF cards onto the editing system’s hard drive would also not be necessary. Files recorded to HDD or SSD media can be accessed immediately as NLE externally mounted drives.
And those same files can be recorded in an edit-friendly, intermediate codec such as Avid DNxHD or ProRes, depending upon the model of recorder used.
Connectivity
This is one feature that separates the men from the boys, so to speak.
HDMI is, was, and always will be a consumer grade connector, never intended for use in professional workflows. SDI, on the other hand, is designed for professional video applications due to its locking mechanism and its’ ability to deliver timecode signal.
Timecode via HDMI connectivity is not a given, although the Ninja-2, for example, can see the added timecode in the HDMI stream on Canon and Sony cameras that have it.
Although HDMI and HD-SDI will pass along identical video signals, the differences between the two are primarily centered around:
• Versatility with other devices
• Whether the connection supports 8 channels of audio as well as SMPTE timecode
• The physical configuration of the connector itself: Locking SDI vs. Fitted HDMI
No Audio to Video Syncing Required
Another post production benefit in using an External Recorder is the simultaneous mastering of audio with the video signal in one single file. The formerly necessary, cumbersome step of matching a separately recorded audio track to a video file suddenly becomes a non-issue. Audio can be passed along with video through the HDSLR’s output to the Recorder, or the camera can be bypassed altogether by plugging audio source(s) directly into the Recorder.
The higher priced Recorders typically include professional XLR mic inputs and contain advanced audio circuitry compared to a HDSLR. Feeding directly to the recorder also eliminates the need for a mic preamp (such as Sound Devices MixPre-D, Juiced Link, or Beachtek) which are typically used when recording sound direct-to-camera.
Elimination Of The External Monitor
Although adding an External Video Recorder also adds the amount of hardware to deal with (and also increases your battery, mounting and cable inventory), there is another heretofore necessary device you may be able to do away with.
Models such as the Atomos Ninja-2 and the Sound Devices Pix 220i include on-board monitors which have the professional calibration features typically available on high quality, stand-alone monitors (features such as false color, zebras, peaking etc.).
If you have yet to invest in a stand alone monitor, you can apply the savings towards the Recorder. For example, the MSRP of the smallHD DP4 is $499. Purchase the Ninja-2 instead and you are effectively paying half price for the Recorder, while still gaining the important asset of a confidence monitor.
Timecode
Depending upon their price point, External Recorders can also record timecode, making multi-camera shoots easier to assemble in post. In addition, timecode allows for more accurate clip logging as well as more accurate frame editing.
Another advantage is syncing independently recorded, timecode matched video and audio sources.
Other Uses
In addition to recording video direct from your camera, External Recorders typically accept other HDMI or HD/SDI sources, such as hand-held devices like an iPhone and iPad. This extends your investment and broadens potential applications for multimedia production.
[The perfunctory word to the wise: Please be sure you are recording non-copyrighted material in order to respect the rights of other content creators.]
And don’t forget that an External Recorder also comes in very handy for replaying finished productions on projection devices. The ability to feed content from a recorder onto a projector or monitor for client review is an added bonus.
Conclusion
Needless to say, these are exciting times in the External Video Recorder product sector, with many companies upgrading their units and pushing the industry forward regarding product development.
For instance, Sound Devices improved the functionality of their on-board monitors in the latest “i” series of recorders (the 220i and the 240i).
Of course, the big news from Sound Devices most recently is opening up access to 3G-SDI by offering 12-bit ProRes 4:4:4:4 (rgb/a) codec up to 30fps — with an image quality that is said to be “indistinguishable” from the source material.
This is a significant step forward for such an economically priced device: This free firmware upgrade gives the 240i a “leg up” upon other 4:4:4 units like the Convergent Design Gemini 10-bit recorder (retailing at double the price).
And not to be overlooked is Sound Devices’ recent change in their warranty policy … from one year to two years. The company made this change retroactive for all of their Recorders, demonstrating an impressive commitment to their entire product line.
Other Recorder manufacturers are also busily innovating: Atomos’ latest iteration of the Ninja — the Ninja-2 — added HDMI output, improved its’ on board monitor capabilities, and added Avid DNxHD codec.
Not to be left behind, Black Magic’s bargain basement HyperDeck Shuttle 2 added ProRes HQ which significantly reduces the size of uncompressed HD video files while preserving full frame, 10-bit 4:2:2 signal quality allowing users to record up to six times longer than before.
Not bad for a unit which sells for $327.
As technologies continues to advance and move forward, high quality acquisition devices will become more and more accessible, making for a much more level playing field in the world of digital video production.
Daniel Freytag’s comparison of the Ki Pro and the HyperDeck Shuttle 2
We invite you to follow us on Twitter
























