High ISO test for Canon EOS 600D, Canon EOS 650D / Rebel T4i, Nikon D3200. This comparison test was published by FOCUS NUMERIQU, the noise of Canon 600D, Canon 650D and Nikon D3200 is very low, and not visible at up to ISO 1600, so we have selected images after ISO 1600 to 12800.
| Store | Price | Cheapest Price |
|---|---|---|
| Canon 600D / T3i | $579.00 | Buy Now |
| Canon 650D / T4i | $849.00 | Buy Now |
| Nikon D3200 | $699.00 | Buy Now |
Conclusion – Translated by Google:
Typically for an APS-C CMOS sensor, noise is well contained 100 to 800 ISO with a grain almost invisible and very gentle smoothing at 800 ISO. A quantum leap is taken at ISO 1600 with a granulation becomes visible and more blunt for details, but the image remains a draw for great classic between A4 and A3. A cut above (3200), the grain is a little more visible and it just nibble the finest details. A new line has been crossed in 6400 with a slight color drift: the colors become pale and lose intensity. On flat colors, the grain becomes visible and provides a geometric structure unattractive. On the black background, it is possible to see purple veins. In ISO 12800, the dynamics decrease, the grain becomes coarse in fact eliminating the fine details. It is, as proposed Canon, ISO sensitivity up really useful because, let’s say it bluntly, the H mode (ISO 25600) is frankly not good. The rendering is very washed out with a little aesthetic granulation.
Follow Us:
Utterly pointless test. Canon applies lots of noise reduction in camera, you even get to work with the finished image (JPG or RAW) so a much better test would be to take these results and use well processed images with moderate amounts of noise reduction in Photoshop using something like Topaz denoise for the Nikon images. Anyone who thinks 25600 ISO images from these el-cheapo bodies would look anything other than "poor" is delusional.
Utterly pointless test. Canon applies lots of noise reduction in camera, before you even get to work with the finished image (JPG or RAW) so a much more appropriate test would be to take these results and use only moderate amounts of noise reduction in Photoshop using something like Topaz denoise for the Nikon images. Anyone who thinks 25600 ISO images from these el-cheapo bodies would look anything other than "not good" (that's a direct quote) is more than a little bit delusional.