PowerShell 和 ffmpeg : No such file or directory

Question

我正在尝试在 Python 脚本中使用以下命令，但是我发现这是一个 PowerShell 问题，因为它似乎无法找到我指向的视频文件。
我在云盘上有一个视频文件Z:在以下位置(为了重现性，将其更改为您想要的包含视频文件的任何路径):Z:\Udemy_And_Misc_Downloads\TensorFlow Developer Certificate in 2021 Zero to Mastery\3. Neural network regression with TensorFlow\18. Setting up TensorFlow modelling experiments part 2 (increasing complexity).mp4请注意文件名中的空格和特殊字符，以防万一。
我想要做的是使用以下命令将该视频文件缩减为更小的尺寸:ffmpeg -i "Z:\Udemy_And_Misc_Downloads\TensorFlow Developer Certificate in 2021 Zero to Mastery\[TutsNode.com] - TensorFlow Developer Certificate in 2021 Zero to Mastery\3. Neural network regression with TensorFlow\18. Setting up TensorFlow modelling experiments part 2 (increasing complexity).mp4" -y -vcodec libx264 -acodec ac3 -threads 1 "Z:\Udemy_And_Misc_Downloads\TensorFlow Developer Certificate in 2021 Zero to Mastery\[TutsNode.com] - TensorFlow Developer Certificate in 2021 Zero to Mastery\3. Neural network regression with TensorFlow\18. Setting up TensorFlow modelling experiments part 2 (increasing complexity)DOWNSAMPLED.mp4"我知道它很长，因为它主要被文件名消耗，但我得到的错误是:Z:\Udemy_And_Misc_Downloads\TensorFlow Developer Certificate in 2021 Zero to Mastery\[TutsNode.com] - TensorFlow Developer Certificate in 2021 Zero to Mastery\3. Neural network regression with TensorFlow\18. Setting up TensorFlow modelling experiments part 2 (increasing complexity).mp4: No such file or directory路径是正确的，因为它是从文件夹的 URL 栏中直接复制粘贴。
我尝试过的事情包括:

在文件名

周围使用引号

不在文件名周围使用引号

在引用的文件名前加上“r”(我收到消息说“你的意思是文件:r 吗？)

在引用文件名前加上“file:r”

在文件路径中使用双斜杠\

有没有powershell pro可以告诉我为什么它不能“找到”文件？
更新
如果我找到具有以下内容的文件名，则该命令运行:

$oldvids = Get-ChildItem *.mp4, *mov, *wmv, *avi -Recurse
foreach ($oldvid in $oldvids) 
{
    $newvid = [io.path]::ChangeExtension($oldvid.FullName, '_.mp4')
    ffmpeg -i $oldvid.FullName -y -vcodec libx264 -acodec ac3 -threads 1 $newvid
}

当我打印 $oldvid.name和 $oldvid.FullName ，路径/文件名如下所示:

Z:\Udemy_And_Misc_Downloads\TensorFlow Developer Certificate in 2021 Zero to Mastery\5. Computer Vision and Convolutional Neu
ral Networks in TensorFlow\35. Multi-class CNN's part 9 Making predictions with our model on custom images.mp4

35. Multi-class CNN's part 9 Making predictions with our model on custom images.mp4

...第一个是我在发布的代码中尝试使用的内容的精确副本。那么有什么区别呢？？
第二次更新
我认为问题在于文件位于 C: 以外的驱动器上。 .只需将文件移动到我的计算机上并运行命令对其进行测试，它就可以工作。所以我想，有没有办法告诉 PowerShell 使用 Z:驱动器找到文件？

Answer 1

我发现问题在于某些文件名中包含特殊字符(可能来自另一个操作系统)，例如 ' .一旦我删除了这些，问题就消失了。我的错。
但是，我创建了一个简化的脚本，如果他们想要的话，可以帮助其他人。我会通过添加一些错误检查来改进这个脚本，检查已经下采样的视频文件(我在文件名中用 ._ 表示，但可以使用任何东西)等等，但是在这里你去:

import subprocess
import os.path
import os

# ------------------------------------------------------------------------------------------------ #
#                                         GLOBAL VARIABLES                                         #
# ------------------------------------------------------------------------------------------------ #
types = (".mp4", ".mov", ".wmv", ".avi") # the tuple of file types
directory_to_recurse = r"Z:\Python_Projects\Misc\Video_Reformatter"

# ------------------------------------------------------------------------------------------------ #
#                            HELPER FUNCTION - RUN A POWERSHELL COMMAND                            #
# ------------------------------------------------------------------------------------------------ #
def run_powershell_command(cmd):
    completed = subprocess.run(["powershell", "-Command", cmd], capture_output=True)
    return completed

# ------------------------------------------------------------------------------------------------ #
#                                          SIMPLIFIED TEST                                         #
# ------------------------------------------------------------------------------------------------ #
# Recursively iterate through all folders in "directory_to_recurse", finding video files
# along the way
for root, dirs, files in os.walk(directory_to_recurse):
    # Inspect
    print(root, "\n", dirs, "\n", files, "\n")
    for file in files:
        # If we've found a video file...
        if file.endswith(types):
            # ...parse some data
            full_path_and_file = os.path.join(root, file)
            filename_only = os.path.splitext(file)[0]
            extension_only = os.path.splitext(file)[1]
            # Inspect
            print(full_path_and_file,"\n", filename_only,"\n", extension_only,"\n")
            # Run the ffmpeg downsampling command
            result = run_powershell_command("ffmpeg -i '" + full_path_and_file + 
                                            "' -y -vcodec libx264 -acodec ac3 -threads 2 '" + 
                                            str(root) + "\\" + str(filename_only) + "._" + 
                                            str(extension_only) + "'")
            # Inspect the result object (if desired)
            print("result:", result)
            # If we receive a return code of anything over 0, exit
            if result.returncode >= 1:
                print("ERROR!")
                exit(result.returncode)
            # Once we've successfully downsampled the video, delete the original
            os.remove(os.path.join(root, full_path_and_file))
            continue