DVCPRO HD Criticisms

I've been a huge fan of the 1200 deck/FCP HD solution, but some of my competitors and a few clients have been critical of it. I'm eager to provide some rebuttals but wanted to get "backup" from some of you gurus out there with hard facts. Here are a few of the slams people have mentioned:

1) AV Video MMP magazine wrote an article that says capturing over firewire with the 1200 will not result in compression/decompression cycles if you stay in FCP/Motion, but going to After Effects results in compression issues that make the footage noisy and generally poor quality. Is this true?

2) Many believe that editing from the native Codec of DVCPRO HD is not as good as capturing via HD-SDI thru a capture card. I've been told that as long as you edit in an uncompressed HD sequence with Firewire HD captured footage, you will not have any issues. What's the truth?

3) People complain that they see a greater amount of Banding in DVCPRO HD footage shot on Varicam. This is one of the biggest issues thrown at me from Sony HDCAM loyals. I have yet to see this, but how about others?

Thanks for taking the time. I'm grateful.

1) Capturing DVCProHD over firewire does a data copy to your hard drive. If you do a cuts only edit there is no decompression / recompressed and the result back to DVCProHD tape is losslessly identical to the original footage

If you do an effects edit in FCP with colour correction, dissolves etc. Then there is one decompression and one recompression to DVCProHD. This is not bad.

If you export a segment of the timeline to After Effects and back to FCP there is also just one decompression recompression cycle. If you're doing effects in FCP and AE on the same clip, export from FCP uncompressed, work on it in AE uncompressed and export back to FCP uncompressed. Then you'll still get just one decompression and recompression.

2) Capturing over SDI and going back to the deck ALWAYS involves one decompression and recompression. You're editing uncompressed and that gives NO advantages over method 1. There is still one decompression and one re-compression. As there is no uncompressed HD deck to go back to, there's always a final compression step. Even if there was an uncompressed HD deck, or you use a more lightly compressed deck like HDCAM SR, you cannot avoid the DVCProHD decompression. You can avoid recompression to DVCProHD using either workflow 1 or 2. In 2 it's obvious as you're editing uncompressed. In workflow 1 you just edit in DVCProHD until you want to output to your better deck, at which point you change the timeline settings to uncompressed, you re-render and output over SDI.

3) Seen no banding in either. Both the HDCAM and Varicam are good cameras. End of story. They have pros and cons each.

So yes, the AJ1200 workflow is great. I still think you need something like the Decklink HD card for monitoring out to a HD monitor though, and for SDI input output with other HD decks.

With HDCAM you're forced into workflow 2. No choices. With DVCProHD you can choose 1 or 2, or even, as I suggest start in workflow 1, and render and output using 2 if going to a better than DVCProHD deck or HDCAM. Either way, quality is the same as decompression and recompression are matched.

However, if you want to do a cuts only edit in workflow 1 it's totally and utterly lossless. Only way to do that on HDCAM is with a Sony Xpri edit system as with that you can edit HDCAM native as you can in FCP edit DVCProHD native.

OK??

Graeme

Graeme,

As always, I'm both grateful and impressed with your thorough clarity. Thanks. This is excellent ammunition for future talks.

David

Graeme,

Excellent review.

That is the truth.

Mike
Midnight Productions
www.midiproductions.com

Thanks!

Graeme

Very cool, Graeme. Can we extend the love? Where do you fit HDV in the masonry-- especially Sony's implementation with their 3-chip HDV camera?

- Loren
Today's FCP HD 4.5 keytip:
Depending on selection, Command-Option-L for video opacity or audio gain!

The FCP HD KeyGuide?: your power placemat.
Now available at KeyGuide Central
www.neotrondesign.com

HDV is a very different situation. 2 formats, 720p30, and 1080i60 (and 1080i50 for PAL users). The JVC's are 720p and the Sonys are 1080i.

The 720p format is a full1280x720, but the 1080i format is reduced to 1440x1080 (like HDCAM), but the FX1 only has 960 x 1080 pixels on the CCD so it uses pixel shift technology to fill in the gaps.

Recording format is long GOP MPEG2. The JVC offerings are 6 frame gops, the Sony about 15 frames. Data rates are 19mbps and 25mbps.

So, can you edit MPEG2 native? Welll, yes you can, theoretically, but it's pain ful MPEG2 with long gops was never designed for that. Personally, I think the route to any quality off these cameras is to edit them uncompressed.

What Apple could (should) do, if they are reading this, is allow timelines to have multiple codecs without rendering. Then, the MPEG2 could be dropped on the timeline, but whenever an effect or edit is done, it could render to a good proxy codec, or uncompressed. Finally, another setting for the timeline could be your output to tape codec, have another render to MEPG2 and out you go.

Graeme

""What Apple could (should) do, if they are reading this, is allow timelines to have multiple codecs without rendering""

OH YEAH! (Think- Ferris Buellers Day Off - (Chicka-Chick-ahhhh!)

Please Apple - wake UP to reality! Having to render to one codec is so 90's.

Then again, I wonder about Sony's insistence to use only the MPEG-2 format for recording. With H-264 on the near horizon, MPEG-2 seems so 95's!

Yes, MPEG2 is so out of date it's not funny. Yes, HDV is deliberately MPEG2 because it's old and out of date. HDCAM SR uses MPEG4!! Now that's a lot better!!

As for what Apple needs to do with FCP:

1) write a decent object / clip / sequence database

2) allow access to the database via scripting controls, with a nice scripting language like TCL or LUA

3) Take the existing FCP gui and tie it to the scripting language. Change some of the GUI elements so that they can be generated / controlled via scripting.

4) make the effects use Image Units / Video units from Tiger. Make a development environment for them inside FCP like Shake.

5) allow you to make compiled FCP plugins that use OpenGL for the GUI, with scripting controls, like TCL/TK that access the underlying scripting architecture

WHY??

The current object database is broken. That's why media manager is so crap. Building a decent one is the only way forwards. Adding controls to that through scripting is THE route to total application stability. Scripting archictecture that's open to developers means new features fast, and any feature you can think of could be done by Apple or external developers. Look at the fantastic features that Vegas has due to it's scripting architecture. Video units and GPU processing are the way of the future - embrace it! Move to a new PCI architecture will mean pulling data back from the card is as fast as putting it onto the card - think power!!

That's my recipe, Apple, if you're reading for future success for FCP. If you keep going with FCP like you are now you do not have the programming power to take on the new features that are cropping up in the competition. IF you do not open up the base architecture to external developers you'll never be able to beat the competition!! If you don't move to a new, stable data base, you'll never take on Avid in the high end.

I don't mean to be a doom and gloom merchant, but if these things are not done, another application will come along that DOES do these things, and then FCP's days of glory are over. Just as we all moved from Premiere to FCP, we'll move from FCP to newApp. Seriously.... Are you listening, Apple??

You've gone as far as you can with FCP as it is. You need to rebuild from the ground up. Do it. Do it now. Tell people you're doing it and get the community involved. You know it makes sense.

Graeme

Hi Graeme,

Following up on DVCPROHD sequence > 10-bit onlining procedures: suppose one is mastering to HDCAM SR. My client wants a U.S. broadcast standard, 10bit, 1920x1080, 60i, master.

When you take a DVCPROHD sequence and 'drop' (nest) it into a 10 bit, 1920x1080 sequence, OR if you change the settings of your DVCPROHD sequence to 10 bit, you not only have to re-render, but also have to upscale all the material (200%), do you not?

How is it best quality to handle the necessary upscaling of the 960x720 footage? 1) In FCP, as described 2) By redigitizing the footage via HD-SDI or 3) is it possible to have, eg, a Decklink Pro card record a 960x720 10 bit sequence to the HDCAM SR tape?

The issue of compression (DVCPRO vs 10bit) seems to be a separate issue from that of pixel dimensions (960x720, or other vs 1920x1080). Comments?

Many thanks for the all around intelligent discussion!

Josh