I currently help edit for a nationally broadcast show. We shoot HDV in 1080/60i and import HDV files into FCP 7. We edit in the same format & export a .mov file. We then pull that file into Adobe Media Encoder and output a MPEG2 file for network submission for SD broadcast.

My first priority is Quality for broadcast - best it can be in SD of course.

What we have now is working ok - but I'm always looking to improve.

Suggestions / Ideas for improved workflow ?

Thanks
D

HDV in......

MPEG2 out : [ these settings are already saved in Adobe Encoder ]

Format: MPEG 2 video
Format profile: MP@ML
Format settings: BVOP: Yes
Format settings: GOP M=3, N=15
Bit rate mode: constant
Bit rate: 8000 Kbps
Width: 720 Height: 480
Display aspect ratio: 4:3
Frame rate: 29.970 fps
Chroma subsampling: 4:2:0
Standard: NTSC
Scan type: Interlaced
Scan order: top field first

Audio:
Format: MPEG audio
Format version: Version 1
Format profile: Layer II
Bit rate mode: constant
Bit rate: 224 Kbps
Channels: 2
Sampling rate: 48.0 KHz
Sample size: 16 bits

Is there a reason you don't export MPEG2 directly from FCP?

Final Cut doesn't transcode as efficiently as Media Encoder or Compressor. FCP 7 is still 32-bit architecture and so doesn't take advantage of additional RAM. Also, FCP doesn't recognize multiple or multi-core processors. It's slow at transcoding and gives you lower quality video than the other two.

Encoder does move along quite fast. That's the reason I was given myself to use it instead of the others.

I'm looking for better workflow back up the line from there. For instance: I've been told that there's an advantage in quality in exporting the mov file from a 720p timeline vs a 1080 timeline. I've tried exporting from 720,1080 and tried both in mov current settings & Pro Res 422 HQ. Here's a screen shot of all 4.

The only way I see this going faster is if, instead of a mov file in the intermediary step, you export/render to an image sequence. This will speed up your second transcode to MPEG2. The other upside is if your system crashes during the first render, you can start from where it crashed, rather than from the beginning with mov.

The downsides are you'll need more hard disk space, and you'll need to layer in audio separately again.

The fastest workflow in your case is to shift to Adobe Premiere. You can edit HDV natively (without the MOV wrapper). You can export directly to MPEG-2. You can stick to the 1080 60i timeline.

Well he's not looking for a faster workflow but a better (higher quality) workflow...

Sareesh hit the nail on the head (as usual).

It looks like the conversion to ProRes either adjusts the gamma or clips the video from 0-255 to 16-235.

The best quality workflow is to export to uncompressed from FCP and use AME to encode from that.

Btw, exporting to NTSC MOV is much lower quality than using MPEG2 CBR. I tested this a while back before uploading some 30s spots.

Steve -

The final out product from AME is MPEG2 / CBR.

If I should export as uncompressed - Should I export from FCP as 8 bit or 10 bit uncompressed NTSC ?

Then , I assume process that thru AME for the required final .

Thanks !

Hi Darin,

8 bit uncompressed is fine. Also, for best results, make sure you have MRQ checked in AME (Max Render Quality). It will add significant time to the encoding but makes a noticeable difference, especially when down-scaling. Probably the best method would be to export as HD from FCP and use AME to down-scale. With MRQ enabled on a CUDA enabled machine, AME uses the same Bicubic Sharper algorithm as Photoshop which is very very good.

Better still, why not bypass the mpeg2 standard and go directly to h264 encoding instead? If the station will accept this then it's a no-brainer.

Whether they are digital or analogue broadcasting, chances are that your programme will benefit. Even for digital broadcasting, I'd expect that your mpeg2 file will be re-encoded for DTV transport stream output (which isn't a regular mpeg2 stream).

Andrew