Magic Bullet Denoiser 2 Trial and Comparison

With the launch of the new Red Giant Denoiser 2 I wanted to see how how this update was able to change the final picture compared to the original Denoiser. Though a RED MX is far from being considered noisy at 2000 iso, this is pixel movement and color shift on the pixel level. With the compression from 4.K WS down to 1080p, and then from RedRaw to ProRes 422 HQ, and finally to YouTube's compression, there is noticeable blocking of colors seen on large stretches of mid value colors. Specifically the orange wall and pylon skirts.

This was dramatically fixed in both the original Denoiser and its newest version.

However, for me the biggest test is in "Color Banding." This is when, through averaging, colors get clumped together into large blocks or bands of a single color. This is usually a result from the program working through and area that has no real detail or internal contrast. This can be from poor lighting, being out of focus, an already uniform color, or a combination of these contributors.

As best seen in the dark portions of the roof in wide shots. There is a fair bit of noise in the original unworked clips. The original denoiser has a hard time here with color banding as at the entrance settings are creating very large and noticeable, and ever shifting islands of color. The newest version of Denoiser handles this much better out of the gate. Though there is blocking of the colors, the size of these are significantly smaller than the previous iteration.

Now of course these issues can be massaged further with the tweaking of the individual settings that make up the "Correction Factor" of the program. But without spending the time to balance acceptable levels of blocking and noise persistence, I can assume the newer version DOES INDEED result in a significant improvement in the ability to correct noisy footage.


One issue I did face that I thought deserved mentioning was the horrific difference in time to render this new version of Denoiser. The original was able to rip the clips through After Effects in approximately 40 minutes to one hour. The new Denoiser Two (2) took over 3 and a half hours. Not sexy for a 30 second portion of a film. Though the results are there, I cannot say that it is 3 and a half times better of a product.

FURTHER TESTING in the render time is coming, please subscribe to learn more information once it is available.

Thank you for taking the time to learn more about this great product, and happy filming!