pil.save()不能保存图像的不同质量版本

Question

我想将图像保存为具有不同质量设置的" JPEG"，然后再次加载它，并读取量化矩阵。对于每个质量参数，它们的矩阵应该不同。

以下代码应该这样做：

from PIL import Image
im_path = '~/Pictures/Origianl.JPG' # path to original image
im_path_tmp = '~/Pictures/DifferentQuality_' # trunk of path to images of lower quality
im = Image.open(im_path) # load original image
qs = {} # empty dictionary to save quantization matrices
qs['orig'] = im.quantization[0] # also remember q-matrix of original image
for i, qual in enumerate(range(100,-1,-10)): # run over 10 different quality parameters
    curr_path = im_path_tmp + str(qual) + '.JPG' # each image gets its own path
    print('iteration: {:2d}, curr_quality={:03d},curr_path={}'.format(i,qual,curr_path))
    im.save(curr_path, 'JPEG', qualilty=qual) # save image: quality-parameter is set!
    tmp = Image.open(curr_path) # load image again
    qs[qual] = tmp.quantization[0] # read out q-matrix and save to dict
    del tmp
    del curr_path
print()
for key, value in qs.items(): # show all q-matrices
    print('{}:nt{}'.format(key,value[0:32]))

我预计现在有10种Q- matrices。尽管它们都是相同的：

orig:
    array('B', [2, 2, 2, 2, 2, 2, 3, 2, 2, 3, 5, 3, 3, 3, 5, 6, 5, 5, 5, 5, 6, 8, 6, 6, 6, 6, 6, 8, 10, 8, 8, 8])
100:
    array('B', [8, 6, 6, 7, 6, 5, 8, 7, 7, 7, 9, 9, 8, 10, 12, 20, 13, 12, 11, 11, 12, 25, 18, 19, 15, 20, 29, 26, 31, 30, 29, 26])
90:
    array('B', [8, 6, 6, 7, 6, 5, 8, 7, 7, 7, 9, 9, 8, 10, 12, 20, 13, 12, 11, 11, 12, 25, 18, 19, 15, 20, 29, 26, 31, 30, 29, 26])
80:
    array('B', [8, 6, 6, 7, 6, 5, 8, 7, 7, 7, 9, 9, 8, 10, 12, 20, 13, 12, 11, 11, 12, 25, 18, 19, 15, 20, 29, 26, 31, 30, 29, 26])
70:
    array('B', [8, 6, 6, 7, 6, 5, 8, 7, 7, 7, 9, 9, 8, 10, 12, 20, 13, 12, 11, 11, 12, 25, 18, 19, 15, 20, 29, 26, 31, 30, 29, 26])
60:
    array('B', [8, 6, 6, 7, 6, 5, 8, 7, 7, 7, 9, 9, 8, 10, 12, 20, 13, 12, 11, 11, 12, 25, 18, 19, 15, 20, 29, 26, 31, 30, 29, 26])
# and so on ...

请注意，它从Orignal更改为其他所有。如果我像在python-Interpreter中一样手动保存它，则出于某些奇怪的原因有效：

In [90]: im.save('~/Pictures/manual.JPG', 'JPEG', quality=70)
In [91]: tmp = Image.open('~/Pictures/manual.JPG')
In [92]: tmp.quantization[0]
Out[92]: array('B', [10, 7, 7, 8, 7, 6, 10, 8, 8, 8, 11, 10, 10, 11, 14, 24, 16, 14, 13, 13, 14, 29, 21, 22, 17, 24, 35, 31, 37, 36, 34, 31, 34, 33, 38, 43, 55, 47, 38, 41, 52, 41, 33, 34, 48, 65, 49, 52, 57, 59, 62, 62, 62, 37, 46, 68, 73, 67, 60, 72, 55, 61, 62, 59])

如果手动完成行为，为什么行为会有所不同。我最好的猜测是，口译员优化了某个人。如果是这样，我该如何关闭？我正在使用Anaconda Python 3.6.1作为解释器运行，这是任何帮助。

Answer 1

感谢 @ch7kor，他指出，保存功能的签名中有一个错别字。我通过" Qual il ity"而不是"质量"

解决我的实际问题。

提出了下一个问题，为什么没有丢弃错误。我怀疑该函数签名中存在此参数（带有错字）...