I am doing a LAB to RGB conversion.<div><br></div><div>If my LAB matrix is float64, the bounds for the axes must be -100, 100. &nbsp;Upon analysis of the output RGB image, the results look correct.</div><div><br></div><div>However, if the LAB matrix is float32 the bounds are checked and forced to -1 1 which&nbsp;yields&nbsp;an RGB image that is incorrect.</div><div><br></div><div>Upon further inspect, the problem looks like it comes from the convert function in dtype.py.</div><div><br></div><div><div>&nbsp; &nbsp; image = np.asarray(image)</div><div>&nbsp; &nbsp; dtypeobj = np.dtype(dtype)</div><div>&nbsp; &nbsp; dtypeobj_in = image.dtype</div><div>&nbsp; &nbsp; dtype = dtypeobj.type</div><div>&nbsp; &nbsp; dtype_in = dtypeobj_in.type</div><div><br></div><div>&nbsp; &nbsp; if dtype_in == dtype:</div><div>&nbsp; &nbsp; &nbsp; &nbsp; if force_copy:</div><div>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; image = image.copy()</div><div>&nbsp; &nbsp; &nbsp; &nbsp; return image</div></div><div><br></div><div>at the "if dtype_in == dtype:" the types are float32 and float64 causing the if to be skipped. &nbsp;Eventually the code reaches the point:&nbsp;</div><div><br></div><div><div>if kind_in == 'f':</div><div>&nbsp; &nbsp; &nbsp; &nbsp; if np.min(image) &lt; -1.0 or np.max(image) &gt; 1.0:</div><div>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; raise ValueError("Images of type float must be between -1 and 1.")</div><div>&nbsp; &nbsp; &nbsp; &nbsp; if kind == 'f':</div><div>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; # floating point -&gt; floating point</div><div>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; if itemsize_in &gt; itemsize:</div><div>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; prec_loss()</div><div>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; return dtype(image)</div></div><div><br></div><div>and the bounds are checked, outside of [-1 1] and an error is thrown.</div><div><br></div><div>Isaac</div>