Go to Home Page

Click Here to Go to All Video Accessory Reviews

 

Product Review
 

Algolith DragonFly Digital Video Processor (De-interlacer/Scaler/Cadence Corrector)

Part II

July, 2006

Ofer LaOr

 

The version I tested was an advanced beta version, and had a few limits. One such limit was the ability to only route analog inputs into the analog output and force HDCP on HDMI content. This is one downside that Algolith has unofficially stated thkey will fix very soon in a future firmware release.

Many features that we've gotten used to from most video processors have not been implemented in the DragonFly, and I would hope that they get added soon. This includes custom output video timing (a critical issue for many display types that have non-standard native resolution timing), customized aspect ratios, underscan/overscan support, and test patterns. Algolith has stated that they will look into these features and may implement them in later firmware releases, but that they are focusing on functionality and stability in the next release.

The focus on projectors has the DragonFly making some interesting assumptions. First, there is clear focus on keystone correction, angle adjustments, and rotation features. Keystone can be corrected digitally (i.e., by deforming the picture vs. trying to correct these issues optically) in a variety of ways either by moving each corner of the image to its dedicated location or by specifically setting vertical and horizontal keystone ratios. Obviously, correcting these issues optically is a better solution, but the ability to do this digitally, with very little visible artifacts as a result, is a nice addition. Color temperature and gamma adjustment features, however, are eye raising as they assume a base gamma curve and color temperature from the display that we haven't been notified about. These two features would make a lot more sense if the unit was bundled with a specific projector, but given the history of DragonFly and Silicon Optix demos, I have a feeling these features were optimized for the JVC HD2K projector. Fiddling through the options I found it hard to imagine how people might use them realistically for other displays.

From a PQ (Picture Quality) standpoint, HQV de-interlacing is delivered just as promised, and in my opinion, it clearly surpasses other de-interlacing algorithms I've tested. Realta/HQV seems to be in a completely new category. When compared with more recent efforts, like the ABT102, the differences are not as big, and although I saw more benefits for HQV (less combing, slightly better resolution), the main differences is a "different look" between the algorithms. i.e., in cases where one of the algorithms might work better, the other might be resolving slightly lower details. I would say that HQV is the king of the hill right now, but other new systems are very close on HQV's heels.

There are two benefits that DragonFly offers beyond other scalers at the moment. First, it does full HD de-interlacing (video and film). Except for the recently added competitor, the Crystalio II, no other processor has this feature at this level of quality. Testing HD sources was a real pleasure, and my usual test sequence (Lilu chase scene from The Fifth Element) in 1080i was unusually clear and detailed.

Movie content at 1080i uses 2:2 and 3:2 cadences with only occasional bad editing, but sports content filmed at 1080i video mode will likely benefit greatly from this feature.

The second unique feature offered by the DragonFly is the Noise Reduction Option. Essentially, this is a mini version of the Mosquito for SD content running through the DragonFly. Personally, I find it difficult to watch digital SD content from satellite sources without it running through the Mosquito. Whereas the Mosquito focuses on two different MPEG artifacts: mosquito noise (hence the name) and block/macroblock noise (BAR = Block Artifact Removal), the DragonFly only focuses on mosquito noise in SD content. This algorithm works quite nicely, but realistically, if MPEG noise is seriously bugging you, you should consider the Tick or Mosquito (other Algolith processors) for a more complete solution!

I found that the DragonFly variation removed a bit too much details for my taste, as the image just looked too "clean" and without enough details. This can be countered by DragonFly's detail enhancement feature that causes a lot of "oohs" and "wows" in demos, but adds too much EE (Edge Enhancement) sharpness artifacts for my own preference.

In terms of scaling, the Realta is supposed to have a very powerful 1024 tap solution, but I did not find it significantly more detailed than other solutions already available today. During testing, I found that the digital output on the DragonFly produced dramatically better results than the analog output, but this might be a result of my display's analog input resolution, I haven't needed to use analog inputs for processor testing for the last few years, as most processors provide DVI or HDMI outputs that can route both analog and digital inputs.

The DragonFly's remote is a dramatic leap from the Mosquito's original huge remote. Calling it tiny would be a serious understatement. It is miniscule and is clearly intended to be used as an interim solution, as such a processor would usually be integrated into a controlled environment with either a macro-enabled remote or a full Crestron/AMX style control system.  Still, I found the remote's range and usability very neat, and beyond ensuring that the remote doesn't get lost, it's a satisfactory item.

Conclusions

From a video processing standpoint, the Algolith DragonFly is a great de-interlacer/scaler. It has many de-interlacing features and ultimately resolves a near optimal picture. Functionally, however, this scaler still has a ways to go to meet the level of features that we expect from such products. High-end projector owners (1080p) will like many of the features on this processor, and although it may not do everything they want from a feature standpoint, it can do a lot of things other scalers are incapable of performing.
 

- Ofer LaOr -

Copyright 2006 Secrets of Home Theater & High Fidelity

Go to Table of Contents for this Issue

Go to Home Page

 

About Secrets

Register

Terms and Conditions of Use