FINAL mastering to Digi-Beta optimizing

[Rory Brennan]

It’s always sad to hear of someone shooting film and not making HD transfers… I’m sure of course that finances were a reason.

Anyway, given that you have captured the footage at DV, the best quality master you can making is DVCam. You will gain no quality benefit going to a better format. You will not lose quality if you had the DVCam dubbed to Digibeta (digital process) but you will lose a barely noticeable amount going to BetaSP (analogue process).

RB

Rory Brennan
Editor
New York City

[Alan Okey]

Since your original film footage was transferred to DVCAM, any shots using unaltered content will never look better than DV, even if dubbed to DigiBeta. Any successive change of format can potentially further degrade the image, so it’s best to avoid it unless a different format is required. There are caveats, however.

If your project has extensive compositing and graphics, it could potentially benefit from being edited in an Uncompressed 10-bit 4:2:2 sequence. This wouldn’t improve the quality of the original DV footage, but it would avoid further degradation of any graphics or composited elements added to the project.

Compared to Uncompressed 10-bit 4:2:2 video formats, DV is noticeably lower in quality. DV has a 5:1 compression ratio, 4:1:1 color sampling and 8-bit color depth. The artifacts of DV compression are visible as banding, blocking and mosquito noise. It’s best to avoid adding DV compression to graphics, composited elements and effects footage if possible.

If you had graded your project in Color, you’d also want to avoid going back to DV. Color allows you to work in 32-bit floating point RGB color space. Depending on the extent of color manipulation, rendering Color output as Uncompressed 10-bit 4:2:2 and then sending the result back to an Uncompressed 10-bit 4:2:2 FCP sequence could potentially look better than rendering to DV, given the higher compression, reduced color sampling and lower bit depth of DV.

As for effects shots, exporting from AE using the Uncompressed 10-bit 4:2:2 or Animation codec and then importing them into an Uncompressed 10-bit 4:2:2 FCP sequence would similarly avoid the image degradation that would be introduced by bringing them back into a DV sequence.

In summary, if your project makes significant use of additional graphics/compositing/effects shots, or if you did extensive color manipulation when grading in Color, you would gain tangible benefits by working and finishing in Uncompressed 10-bit 4:2:2 versus DV.

DigiBeta is a 10-bit 4:2:2 format with a fairly low compression ratio, so it would preserve the quality of your finished project better than DVCAM, which is 8-bit 4:1:1 with a 5:1 compression ratio. If your project has little in the way of additional graphics, compositing or significant color grading, then staying in DV all the way through the pipeline would probably be fine.

[Alan Okey]

I forgot to mention that working in ProRes or ProRes HQ is a perfectly viable alternative to working in Uncompressed 10-bit 4:2:2, as ProRes is (reportedly) visually lossless when compared to full uncompressed. ProRes also has lower disk bandwidth and storage space requirements relative to uncompressed.

[Rafael Amador]

I agree with Rory and Alan about that you have made a big mistake when transferring 16mm to video.
You should have gone HD. And if you’d go SD you should have captured at least at 8b Unc.
But I don’t agree about that there is nothing to do to improve your DVCam footage.
If some time your movie will go to DigiBeta, start working in a sequence 720×486.
Set the sequence codec to 10b Unc or Proress.
Edit, CC etc and get your self the “Nattress Chroma Smooth/Sharpening” filter.
When everything ready, set your sequence to “Render all YUV material in High Precision YUV”.
Cheers,
rafael

http://www.nagavideo.com

[Leslie Tai]

Hi everyone,

I don’t know if it’s correct practice to post a new question in an ongoing thread, but it seems as if I’ve stumbled across the right crowd of experts for my related question, so I’m going to give it a try.

I have a project shot on Mini-DV that is now getting festival interest, so I need to figure out how to ultimately end up with it on Digibeta and HDCam. Most festivals desire Digibeta, so it is right that I want to make a Digibeta master?

The project uses footage shot from both a Panasonic DVC30 and a consumer Sony HandyCam. Most of the footage was digitized into FCP under the DV NTSC 48 kHz setting, and some of it was digitized in iMovie as a DV file, which was later exported out of FCP as Quicktime .mov files in order to work with.

The project has heavy (FCP) subtitling throughout, and no other graphics/compositing/anything.

Digibeta:
I’ve consulted a few of the more technically-inclined, and it seems the vote is split evenly between (1) outputing directly from my uncompressed Quicktime movie file to Digibeta (TRT: 103 min. so it won’t fit on a Mini-DV) and (2) re-digitizing all the clips in the FCP sequence under Uncompressed 10-bit settings before outputing to Digibeta.

Those advocating Option 2 tell me about a noticeable upgrade in image quality, and others tell me the difference is negligible. Clearly, I want to do what’s right for the film, and preserve it in the highest quality possible. I am lacking, however, the information that would help justify the tremendous time and energy commitment (2 weeks?) it would take to locate, redigitize, and match frame every single clip in my timeline for the sake of a potentially negligible improvement in image quality. (This is my first large project, and I regret all the many, many mistakes I made in media management. Total mess.)

HDCam:
The same dilemma applies (within my limited understanding). The dub house I spoke with told me they’d do a file to HD upconversion. Also mentioned the difference was negligible. Of course, I’m terribly aware that it’s going to look like crap whichever way I go, but would having a file that’s uncompressed 10-bit vs. DV make it any better?

This guy also offered that it might not even be possible, with my set-up, to re-capture the footage in Uncompressed 10-bit properly, which I guess might eliminate this entire dilemma completely.

I am working on a MacBook Pro, with FCP 5.0.4, absolutely no bells and whistles.

Thanks so much for any help you can give!

Leslie

[David Roth weiss]

[Leslie Tai] “I don’t know if it’s correct practice to post a new question in an ongoing thread”

Not a good idea if you want your question seen by the whole forum.

Rather than get into a whole technical explanation, which frankly is above your level of expertise at this time, let me simply advise you that, your first time out of the gate, you absolutely must enlist the help of a facility who do this kind of thing all the time. Make sure to get them to agree to allow you to watch everything they do, and that way you will have the knowledge to do most, if not all, of the process yourself the next time.

Hard drives are very inexpensive these days, and you’re working on a laptop as well, so find a good shop in your city that is very FCP oriented, and take your project to them. Ask them if they have a Kona or Blackmagic card, and ask to see what they’ve done on similar projects. You can report back here to the Cow with your findings and we’ll let you know if you’re heading to a facility that sounds good, before you go.

I hope this helps…

David

David Roth Weiss
Director/Editor
David Weiss Productions, Inc.
Los Angeles

POST-PRODUCTION WITHOUT THE USUAL INSANITY ™

A forum host of Creative COW’s Apple Final Cut Pro, Business & Marketing, and Indie Film & Documentary forums.

[Leslie Tai]

Thanks David.

I did a quick search on Kona and Blackmagic card, and I’m guessing that that has to do with color correcting in the HD upconversion.

I guess my main question is whether or not it’s worth it to re-digitize my footage as Uncompressed 10-bit for output onto Digibeta. Due to budget, timeframe, and aesthetic considerations, I won’t be doing any professional color correcting.

I can do all this at the Bay Area Video Coalition, which has all the facilities and technical support to walk me through the entire process. My options with them are (1) I can either output directly as is or (2) re-digitize all of the footage.

Two other local San Francisco dub houses can make the Digibeta for me, but neither of them recommends optimizing the footage by re-digitizing.

I guess I’ll have to do more research on the SD to HD conversion.

Thanks,
Leslie

[David Roth weiss]

[Leslie Tai] “I did a quick search on Kona and Blackmagic card, and I’m guessing that that has to do with color correcting in the HD upconversion. “

No, the Kona and Blackmagic cards are I/O cards , with I/O shorthand for “input/output.” So, they make it possible to input and output from and to the various tape decks for the formats that you have asked about. DigiBeta and HDCam do not have firewire I/O.

Now, as far as redigitizing, that’s not going achieve anything. Your video will never get any better that the original DV (except by good color correction), and anyone who told as you stated earlier that there would be “a significant improvement,” in your video was simply ill-informed and very wrong.

However, by simply going to Sequence>>Settings and changing the Compressor in the Quicktime Video Settings to any of the codecs that operate in 4:2:2 video space, such as DV50 or 8-bit uncompressed, and then re-rendering, you will substantially improve the quality of your text and graphics. DV compression absolutely hammers text and graphics with compression in the much worse 4:0:1 color space. Do keep in mind that your firewire drives will not be able to playback 8-bit uncompressed without dropping frames, so you would need to move your project to a faster raid at the facility where you would output to tape.

David Roth Weiss
Director/Editor
David Weiss Productions, Inc.
Los Angeles

POST-PRODUCTION WITHOUT THE USUAL INSANITY ™

A forum host of Creative COW’s Apple Final Cut Pro, Business & Marketing, and Indie Film & Documentary forums.

[Leslie Tai]

David, thanks for your clarification. I really appreciate all your help and your patience with such a newbie.

Please correct where I’m wrong, but my current understanding is, the 5:1 compression (which is what had me worried) occurs within the camcorder itself, so when I digitize into FCP it makes no difference whether I do so in the DV preset or the Uncompressed 10-bit preset, because the DV codec is essentially bringing in the footage at the highest quality possible, which is just what it is, in 4:1:1 8-bit format. Is this right?

And now to output to Digibeta, if I understand correctly, what I can do is, I open up my current FCP sequence, copy the entire timeline including subtitles and paste it into a new sequence in a new project with Uncompressed 8-bit preset, do all my basic color correcting in FCP here. Then render. Then export Quicktime movie using Uncompressed 8-bit settings (DV settings here would do the 5:1 compression again, thereby losing more quality?). Then bring it to the facility and output to Digibeta. Is this correct?

Now, may I ask, what is the difference between Uncompressed 8-bit and Uncompressed 10-bit? I’ve read above that Digibeta is a 10-bit 4:2:2 format. Do I want to do all of the above in Uncompressed 10-bit instead to make it more lovable to dear Digibeta?

Finally, I just have to ask to ease my conscience. I am mildly worried that some of the footage in my sequence, iMovie DV files which I had to recompress into Quicktime .mov files in FCP, experienced some kind of generational loss(?). The file size was the same, I don’t need to get into deep technical details, so if you think this is fine, I can forget about it.

Again, thanks for your help and understanding with such elementary questions.

Leslie

[David Roth weiss]

[Leslie Tai] “when I digitize into FCP it makes no difference whether I do so in the DV preset or the Uncompressed 10-bit preset, because the DV codec is essentially bringing in the footage at the highest quality possible, which is just what it is, in 4:1:1 8-bit format. Is this right? “

Exactly.

[Leslie Tai] “And now to output to Digibeta, if I understand correctly, what I can do is, I open up my current FCP sequence, copy the entire timeline including subtitles and paste it into a new sequence in a new project with Uncompressed 8-bit preset, do all my basic color correcting in FCP here. Then render. Then export Quicktime movie using Uncompressed 8-bit settings (DV settings here would do the 5:1 compression again, thereby losing more quality?). Then bring it to the facility and output to Digibeta. Is this correct? “

No. As I mentioned before, simply change the compressor in Sequence Settings to 8-bit uncompressed and re-render. Then, either export a Quicktime Movie using current settings or bring the entire project to the facility where you will print to tape. The I/O card will add the necessary 6 scan lines to convert from 720×480 to 720×486 on the fly. There is no additional 5:1 compression.

[Leslie Tai] “Now, may I ask, what is the difference between Uncompressed 8-bit and Uncompressed 10-bit? I’ve read above that Digibeta is a 10-bit 4:2:2 format. Do I want to do all of the above in Uncompressed 10-bit instead to make it more lovable to dear Digibeta? “

That extra 2-bits of information equals nothing but bigger files and the potential for dropped frames for you. Since your original was DV you’re just not going to get any real benefit from the extra bits. If you had extensive graphics with expansive flat areas subject to banding, such as blue skies, then you might want to change the compressor to 10-bit.

For the record, because you are still using FCP 5.1 you are missing out on the ability to use ProRes 422, which is 10-bit, but without the huge file size.

David Roth Weiss
Director/Editor
David Weiss Productions, Inc.
Los Angeles

POST-PRODUCTION WITHOUT THE USUAL INSANITY ™

A forum host of Creative COW’s Apple Final Cut Pro, Business & Marketing, and Indie Film & Documentary forums.