Compared to other lossless video codecs, such as Animation, PNG, and LZW, SheerVideo is designed to be very robust, so that it compresses well even for natural images with very high detail. And even in the most extreme cases, a Sheer-encoded frame never takes up more than 1 bit per scanline more space than the uncompressed original.

Nevertheless, random noise is completely uncompressible, even in theory, so the presence of noise reduces even SheerVideo's compression power. Here is a list of some common sources of noise appearing in video. In order to ensure the compressibility of your image data, you should strive to minimize the noise contributed from these and other sources. If any type of noise is essential to the look of your compositions, you should consider adding it as late as possible.

• Illumination fluctuations: Noise caused by statistical fluctuations in the amount of light reaching a photosite for a static image of a statically lit static scene. Illumination noise be ameliorated by using cameras with larger image sensors, using longer exposure times, and lighting the scene more brightly.
• Thermal noise: Noise caused by thermal motion of electrons in the image sensor. Thermal noise can be reduced by cooling the sensor.
• Fixed pattern noise: Noise due to the unequal response of individual photosites in an image sensor. Better cameras compensate for this by subtracting and normalizing. Keeping the camera away from radiation sources, such as cosmic rays at high altitudes, helps prevent worsening of a camera's fixed pattern noise.
• Film grain: Noise due to the random sizes, shapes, and orientations of photosensitive grains in film. When desired, a grainy look can be added back in afterwards, yielding a more consistent look in compositions with computer-generated effects.
• Print-through: Noise due to the cross-magnetization of adjacent layers of tape that occurs during storage. Print-through can be reduced by increasing the coercivity of the magnetic particles on the tape, storing the tape at low temperature, and periodically rewinding the tape. Modern digital tape formats overcome print-through additionally with spatially-distributed error-correcting codes.
• Tape noise: Noise due to variations in the size and distribution of magnetic particles in magnetic tape. Tape noise can be reduced by distributing the particles more uniformly and by using particles of smaller size. Common magnetic pigments, in order of increasing particle size and therefore increasing noise, are: barium ferrite, metal, chromium dioxide, iron oxide. Modern digital tape formats overcome tape noise with spatially distributed error-correcting codes.
• Dithering: Noise purposely added to an image before reducing the precision, in order to preserve gradations over larger areas, thus avoiding Mach bands.
• Effect noise: Randomness deliberately introduced as part of an effect to make it appear more natural. SheerVideo looks for local coherence in the image, so it is more strongly affected by small-scale randomness than by gradual variations over a larger area.

If you are working in 8 bits per channel, and if the material in your project has predominantly coarse detail and very little noise, you may find that the Apple Animation codec works better for you than SheerVideo. When set to Millions of Colors or Millions+ of Colors and Best Quality, Animation losslessly codes RGB[A] 8bf images. However, Animation does not support RGB[A] 10bf or any precision greater than 8 bits per channel. The Animation codec is designed for traditional flat constant-shaded two-dimensional animations, and uses run-length coding, so if your images have large areas of exactly the same color, Animation will compress them very well. In contrast, the Animation codec typically only compresses real-world images by 5%, if at all. In terms of speed, SheerVideo encodes faster than Animation, but sometimes decodes slower.