When it comes to providing great video signal processing, You’ll Find
no “FLIES” with this combo
by David McCallum
A year ago I began working with Algolith, a Canadian company that
specializes in video signal processing, which lead to my review of
the Flea, Algolith’s entry level video enhancer/artifact reducer
(Vol. 17, No. 1). While I tested the Flea, Algolith’s Robert Young
(vice president of sales and marketing) and Jacques Patry (product
manager) exposed me to problems inherent in digital video
distribution, either via broadcast or DVD, and how their products
attempt to address these issues.
After writing the review for The Inner Ear, I continued testing the
Flea at my post production studio in various professional
applications, working with compressed video signals displayed on both a plasma display and projection screen. I also started working with Algolith as a Beta tester for the Dragonfly, their forthcoming
external video processor.
I must confess that I didn’t contribute much as a Beta tester, but
shortly after the Dragonfly was released, a sample arrived for
review. Because of my growing interest in video artifacts, I asked if
I could pair the unit with the Mosquito, a more powerful noise
reduction unit, and continue my testing with these upgraded products
while working on a joint review of the two units. However, before
beginning the review I’d like to go into a bit of detail about video
artifacts and some of the issues I’ve become more cognizant of over
the past year.
The advent of digital video transmission has greatly enhanced the TV
viewing experience with DTV and HDTV broadcast now commonly
available. However, the substantial increase in the size of the media
now passing through transmission lines introduced new challenges,
namely how do you fit it all in?
DTV created the opportunity for broadcasters to fit multiple channels
through the same 6MHz bandwidth used for analogue television.
Simultaneously, the resolution of HDTV signals is approximately five
times that of a typical NTSC analogue signal. So while we now have
access to much better images and features through our television,
there remains a limit to the amount of available bandwidth to
transmit all this extra picture resolution and other features.
The solution to this dilemma is compression. The most commonly used
form of compression is MPEG-2, a complex process of image analysis
and detection that reduces the size of an incoming image and enables
it to be rebuilt after transmission. The MPEG-2 process, or pretty
much any DCT (discrete cosine transform-based compression) has
visible side effects in the video signal called mosquito noise and
blocking artifacts. In an article published on the website Video
Imaging Design Line, here’s how Alogolith’s Jacques Patry describes
mosquito noise and blocking artifacts. “The VIRIS project (a Video
Reference Impairment System) defines mosquito noise as a form of edge busyness distortion sometimes associated with movement, characterized by moving artifacts and/or blotchy noise patterns superimposed over the objects (resembling mosquitoes flying around a person's head and shoulders). It occurs when reconstructing the image and approximating discarded data by inversing the discrete cosine transform (iDCT) Mosquitoes’ can also be found in other areas of an image. Generally, compression engines are less efficient in compressing high frequency contents such as textures or film grain. Therefore, at playback, these areas are tainted with mosquito noise. The result will be somewhat similar to random noise; the mosquitoes will seem to blend with the texture or the film grain and may look like original features of the picture. But when the mosquitoes are removed, the picture looks much crisper.
“Blocking artifacts, as the name suggests, manifest themselves as
objectionable and unnatural blocks within an image. Sometimes
referred to as macro-blocking, it is a picture distortion characterized by the underlying block-encoding structure becoming visible.
“When pushing the limits of the encoder, the blocks are rather
roughly averaged, making them appear as one big pixel. From block to
block, the average calculated can vary, and thus creates these well-
defined borders between blocks.
“This effect becomes even more pronounced when there’s some fast
motion or quick camera movement. Probably the best example for this
is during NFL telecasts, where the player carrying the football can
quickly turn into some form of low-res, blocky, pixilated Mario Bros.
look-alike from the old Nintendo days.”
After spending the past year testing heavily compressed video images
with Algolith products, I’ve become acutely aware of how much of an
impact these artifacts have on the displayed picture. Once the
problems became recognizable, it became easier to understand why and how Algolith addressed them, and also how good their products are at repairing the damage.
Appearance & Connectivity
All right, let’s get on with the review. What we have here are two
identical looking single rack-sized units. Both products have a soft
silver finish and a clean front panel, which includes only the product name, the Algolith name and logo, and an IR receptor and blue LED light indicating when the unit is in operation.
The Mosquito HDMI offers two component and four HDMI inputs capable of receiving 480i/p, 576i/p, 720p50/60 and 1080i50/60. All inputs include automatic detection of interlaced/progressive signals and SD/ HD signals. HDMI and component are the two output options and, while the component input signals can be routed to the HDMI output, the HDMI input signals will only pass via the HDMI output.
The dragonfly comes equipped with one input each of composite, S-
Video and component and two HDMI inputs. Outputs include single
component, RGBHV via a 15pin D-sub & HDMI. Similar to the Mosquito,
the analogue inputs are available via the HDMI output, but the two
HDMI inputs will not pass via the analogue outputs.
Overall, both products are sleek and simple. While not overloaded
with input options, the result is a simple and uncluttered set-up
that’s efficient and clean and, when paired together, offers plenty
of flexibility and choice. One final note: When using the two units
together, it’s best to connect first to the Mosquito, cleaning up the
signal as much as possible, then to the dragonfly for processing and
Technology & Performance
While the Mosquito and Dragonfly look the same, they perform
completely different functions. The Mosquito addresses the mosquito
and blocking artifacts explained earlier, while the Dragonfly
provides processing, de-interlacing and scaling functions. Paired
together, all aspects of video enhancement are covered. Here’s how
they do it.
The Mosquito focuses on image noise reduction and filtration via four
user settings: 2D Reduction, 3D Reduction, Block Artifact Reduction,
and Detail Enhancement.
The first step in the process is filtering, eliminating bad reception
and low light noise. Next, mosquito noise is removed via proprietary
pixel-based algorithms followed by block artifact reduction, which
blends and diminishes 8x8 pixel block structures of the MPEG-2
compressed images. The last step is the image enhancement, which
helps restore natural sharpness and noise-free images.According to Algolith, the Mosquito maps out the problem areas of the image with edge, detail and artifact regions analyzed separately. The processor then focuses on the visible artifacts and eliminates them. The Mosquito does not invent or recreate original data; rather, it attacks the specific areas where artifacts are located and the rest of the image is left alone.
The final step, detail enhancement, is a much more sophisticated
feature than your display’s sharpness control as the Mosquito
processes edge transitions in both horizontal and vertical dimensions.
The Dragonfly is a more traditional video processor. Using the HQV
processing chip from Silicon Optix, the Dragonfly provides true de-
interlacing using the full 4:4:4 processing window for HD video de-
interlacing, automatic film/video detection and cadence processing
and finally, SD to HD detail enhancement. The processor is able to
scale incoming signals of any available output signal to 1080p and
also includes audio sync delay adjustment inputs to compensate for
the time delay in video processing. Other adjustment features include
brightness, contrast, sharpness, black level, color saturation and
temperature, and hue.
I know all of that sounds technical, but it’s important to have a
full understanding of what and how video processing works in order to ensure that your purchase is operating properly. However, once you’ve configured the Algolith using the appropriate adjustments, you really don’t need to go back into the set-up window.
Used primarily at my post production company in our mixing/screening
theater, I experimented extensively with the units under challenging
conditions. Working with mostly unfinished master tapes of films &
television projects, the material often consisted of heavily compressed 480i DVCAM tapes, connected via the analogue inputs, then
output to a Christie Digital LW40U LCD projector and displayed on a
80” diagonal Screen Research projection screen.
Rather than feeding high-resolution HD images, my objective was to
see just how good I could get these problematic sources to look. I
could go into detail about what I thought; however, after a recent
series of screenings of one of our client’s films, the director, producer, editor and assistant editor, none of whom were aware of my testing, all told me that the screenings were the best they had ever seen from an AVID output. They were thrilled to see their film look so good. That kind of independent reaction offers a much stronger conclusion than I could.
While you can’t turn mud into magic, the features and functions of
these Algolith products provided tremendous improvements to any
signal I fed though them. Obviously, the better the source signal the
better the image, but because our Christie projector operates in
native resolution (meaning it projects the image in the native rate
of the signal) by first performing the mosquito’s noise reduction
then scaling the 480i signal to 1366x768 (the Christie’s native scan
rate) great improvements were realized enabling all products in the
video chain to function at their optimum ability.
Synopsis and Commentary
I’ve spent a year with Algolith products and what I’ve found most
interesting, is how much more educated I’ve become about the inherent problems in video distribution by working with all three units. I’ve done my best to outline what Algolith has set out to tackle with
their product line. However, seeing is really believing, and when you
sit and experiment with these products it becomes clear how sophisticated they are, and how strong an understanding the engineers and developers at Algolith have for video signal processing.
I’ll say that the real jewel of their development is the Mosquito,
which I consider a must for any serious videophile, especially one
who is using a projector (if you’re not into projectors, please read
“Projecting The Future, part two,” in this issue—you might change
your mind). Neither the Mosquito nor the Dragonfly were designed to
be used exclusively in combination with the other, and each unit functions exceptionally well on its own. However, once paired together, all aspects of video signal processing are covered, and to
the best of my knowledge, no single video processor on the market
offers comparable features to the Mosquito and Dragonfly combination