Skip to content

Unknown error when reading an input .mov file

0

Hi everyone, First time posting a question around, I hope it fits the place; Since end of January, we have started to see quite some jobs failing when using the AWS Elemental MediaConvert service. Only happens for .mov files, and all we can see is this error in the job details (I have removed the sensitive url of the media file):

Unable to open input file [{url}]: [Failed probe/open: [no moov box found in file]]

In the job details the error code is the 1075: Demuxer Parser Error: "MediaConvert couldn't recover from a problematic file when demuxing. Verify that you have provided a supported input. If MediaConvert supports your input, open a case with Support." https://docs.aws.amazon.com/mediaconvert/latest/ug/mediaconvert_error_codes.html

I have checked manually more than 25 failed files but I can't understand why they are failing on AWS Elemental MediaConvert side. We thought that might be related to some files lacking the moov_pos atom metadata, but we found some other .mov files without this atom that went correctly through the conversion flow. We also made sure that the files are fitting the supported containers for the .mov extension https://docs.aws.amazon.com/mediaconvert/latest/ug/supported-containers-codecs-details.html#container-mov Some ffmpeg outputs of the file streams that failed:

# AVC (h264)
Stream #0:1[0x2](und): Video: h264 (High) (avc1 / 0x31637661), yuv420p(tv, bt709, progressive), 2160x3840, 91330 kb/s, SAR 1:1 DAR 9:16, 50 fps, 50 tbr, 5k tbn (default)
Stream #0:0[0x1]: Video: h264 (High) (avc1 / 0x31637661), yuv420p(tv, bt709, progressive), 2160x3840, 5029 kb/s, SAR 1:1 DAR 9:16, 24 fps, 24 tbr, 12288 tbn (default)

# HVC1 (HEVC main 10)
Stream #0:0[0x1](und): Video: hevc (Main 10) (hvc1 / 0x31637668), yuv420p10le(tv, bt2020nc/bt2020/arib-std-b67), 1920x1080, 13222 kb/s, 59.96 fps, 59.94 tbr, 600 tbn (default)
Stream #0:1[0x2](und): Video: hevc (Main 10) (hvc1 / 0x31637668), yuv420p10le(tv, bt2020nc/bt2020/arib-std-b67), 1080x1920, 20709 kb/s, 60 fps, 60 tbr, 600 tbn (default)

# HVC1 (HEVC)
Stream #0:0[0x1](und): Video: hevc (Main) (hvc1 / 0x31637668), yuvj420p(pc, bt709), 1440x2560, 11131 kb/s, 30 fps, 30 tbr, 600 tbn (default)
Stream #0:0[0x1](und): Video: hevc (Main) (hvc1 / 0x31637668), yuv420p(tv, bt709), 2160x3240, 9638 kb/s, 24 fps, 24 tbr, 1000k tbn (default)

The files are publicly accessible. Does anyone have an idea why the service is not able to even opening the file? Did something change on AWS Elemental MediaConvert?

asked 7 months ago190 views
3 Answers
1
Accepted Answer

Contacting with support we reached the following conclusion:

  • When we provide a link containing the video to be converted, and it's not located in a S3 bucket, AWS MediaConversion service performs HTTP Range requests for fetching the data. Our server was not handling the response codes correctly (206 Partial content), was returning a 200 success. As a result, the service tried to download in a Range, but our server replied with 200, making believe the client that the media was completely download. Because the moov atom sometimes is at the end of the video file, the AWS MediaConversion service was effectively returning the error "no moov box found in file".
  • We will fix our server to handle this HTTP Range requests correctly.
  • Because our media is already in S3 buckets, we will change the URL we provide to AWS MediaConversion service to always provide a resource from S3 bucket, not from our own server. That way, AWS already has the media and doesn't need to travel or Range download it from outside.

Thanks for the support

answered 6 months ago
0

Try to transcode files that were successfully transcoded before end of January. If it will not work, open support case in AWS support (might be regression).

AWS
answered 7 months ago
-1

The issue you're experiencing with AWS Elemental MediaConvert failing to open certain .mov files is likely related to the structure of these files, specifically the absence or incorrect placement of the 'moov' box (movie metadata).

In QuickTime (.mov) files, the 'moov' box contains important metadata about the video, including information about tracks, duration, and other essential details. When MediaConvert reports "no moov box found in file," it suggests that this crucial metadata is either missing or not in the expected location.

This problem can occur for several reasons:

  1. Incomplete file transfers: If the file upload to S3 was interrupted, the 'moov' box might be missing.

  2. File structure: Some encoding tools place the 'moov' box at the end of the file instead of the beginning, which can cause issues for certain readers.

  3. File corruption: The file might be damaged, preventing MediaConvert from locating the 'moov' box.

  4. Incompatible encoding: While the codecs (H.264 and HEVC) you mentioned are supported, there might be specific encoding parameters or settings that are causing issues.

To troubleshoot this problem, you could try the following:

  1. Use a tool like AtomicParsley or FFprobe to examine the structure of the problematic files and compare them with files that work correctly.

  2. Try remuxing the files using FFmpeg to ensure the 'moov' box is at the beginning of the file: ffmpeg -i input.mov -c copy -movflags +faststart output.mov

  3. Verify that the files are completely and correctly uploaded to S3.

  4. If possible, try processing these files through MediaConvert in a different AWS region to rule out any region-specific issues.

  5. Contact AWS Support with specific examples of failing files, as they may be able to provide more detailed insights or identify if there have been any recent changes to the service that could be affecting your workflows.

Remember that while MediaConvert supports a wide range of input formats, there can be subtle differences in file structures that cause issues. It's also possible that recent updates to MediaConvert have changed how it handles certain file structures, which could explain why you're seeing this issue more frequently since the end of January.
Sources
Error codes - MediaConvert
Troubleshooting - MediaConvert
Supported input formats - MediaConvert

answered 7 months ago
  • The given solution does not fix the issue. Reason 1: file wasn't uploaded to any S3 storage, it's publicly available in our servers. Reason 2: AWS Elemental MediaConvert is able to parse files without a moov atom (I double checked some files lacking this atom and they succeed, others failed) Reason 3: files are not damaged, tested consistently with bunch of different uploads that contained the same original file. Reason 4: this could be the most relevant one, but the AWS Elemental MediaConvert docs doesn't give more details about which parameters might be the conflicting ones. Looking the files closely with ffmpeg I wasn't able to spot a pattern neither.

    About troubleshooting: Troubleshooting 1: already did that unfortunately without a strong pattern to detect why some .mov files are failing and why some others aren't Troubleshooting 2: if I need to remux the file, what's the point of using a Conversion tool like AWS Elemental MediaConvert? We don't have strong GPU servers to remux these videos ...

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.