Betacams and VTRs

[audadvnc]

One of the great aspects of Betacam and Betacam SP is its archival quality. It's probably the best ever long term analog video storage medium, with bandwidth and color quality as good as any, and a solid tape medium to hold the signal. A tape flaw on Betacam will lose only the frame that is damaged, whereas a digital tape may have resync issues that cause you to lose additional information (or DVD's which may become completely unplayable if some data is lost).

[wado1942]

I thought the first post was a joke. Betacam is well known to be one of the best video tape formats ever, in many ways beter than MiniDV. Those decks cost about $50,000 when they were new, no joke either. They're STILL in use at many well established TV stations though they've largely been replaced by DVC-Pro and DigiBeta.

[ccortez]

Funny to find this thread in the archives... I just stumbled into possession of a Sony Betacam SP UVW-1800. I can imagine mastering SD WP transfers to it, but I certainly haven't thought 10 minutes about what I might use it for, since I just found it 5 minutes ago.

[Super8rules]

Wow, this thread is making me feel really old and I am only 43.....

I remember editing 1 inch Ampex tape by actually cutting it and splicing it together....and I thought that was high-tech...

[wado1942]

Hey, I'm 27 and I hand splice 1". I hand splice analogue video also, though I try to avoid it, sometimes it's the best, least lossy way to get a job done.

[Will2]

I thought the post was a joke too... I've probably got 300 to 350 betacam tapes in my closet of projects over the last 12 years. Going to digiBeta mostly now for SD projects.

[mattias]

A tape flaw on Betacam will lose only the frame that is damaged, whereas a digital tape may have resync issues

a tape flaw on analog will lose a couple of scanlines, on digital usually a block but sometimes the entire frame, and as long as there's timecode on the tape you shouldn't get sync issues. you're thinking of interframe codecs like hdv and dvd? those are a pain when it comes to dropouts for sure.

beta sp is obsolete. i see no reason for you or me to use it whatsoever. tv stations still do because they have invested in the decks, that's all. you get the same image quality on minidv and the same archival quality on digibeta.

/matt

[VideoFred]

Talking about digibeta.... The tapes... How long will they last?

Fred.

[wado1942]

you get the same image quality on minidv

By noise and resolution maybe, but Beta SP has better color space.

Fred, the average lifespan of analogue helical scan tape is 15 years. But its shelf life is more like 25. I don't know how Digibeta does with time but MiniDV has an average lifespan of about 5. No doubt Digibeta using 1/2" tape running much faster using much more robustly built cast iron deck chasis would live much longer. Probably pretty close to analogue video. Now the average lifespan of 1" and 2" video are more like 35 years.

[MovieStuff]

... the average lifespan of analogue helical scan tape is 15 years. But its shelf life is more like 25. I don't know how Digibeta does with time but MiniDV has an average lifespan of about 5. No doubt Digibeta using 1/2" tape running much faster using much more robustly built cast iron deck chasis would live much longer. Probably pretty close to analogue video. Now the average lifespan of 1" and 2" video are more like 35 years.

Physically, the shelf life of the binders on tape is that long but you need to alternate heads-out-tales-out on any tape about once every 6 months for long term storage or you get ghosting on the signal. The proper way to archive analog video or audio tapes (cassette or reel to reel) is to store them with heads out for only about 6 months then fast forward them to a tails out position and mark as such then, 6 months later, reverse the process and store them heads out. I had a farily priceless, 10 year old 1 inch master that was ruined by a duplication house because they were not doing that on a regular basis. The last time they handled it was 10 years before and when I decided to get DVcam submasters made from all my 1 inch archived material, it was ruined with ghosting from the layers of magnetic material being compressed against each other over a long period without moving. It is impossible to restore. I finally settled with them because I figured they had a whole warehouse full of other client's tapes that suffered the same fate. Since we were all paying an archiving fee, they didn't want to stir up trouble with their other clients. But, eventually, they'll have to come clean about it.

Roger

[mattias]

By noise and resolution maybe, but Beta SP has better color space.

true but the difference is tiny, especially in pal, and it's not nearly enough to lift the quality above dv. all in all the perceived quality is almost exactly the same. both are way better than for example umatic or vhs, and neither is nearly as good as digibeta, or even dvcpro and imx.

(btw i try not to use the term color space since that means something else. they both use the same yuv color space. i like the term chroma (sub)sampling)

/matt

[wado1942]

Well yes, they're YUV. But there's also the 8-bit per channel of DV vs the indefinite gradations of Beta SP. I'll tell you one thing, it's a heck of a lot easier to do composite work on Betacam and Betacam SP than DV. Though if I had my druthers, I'd rather work with D1 for composite work.

[mattias]

Well yes, they're YUV. But there's also the 8-bit per channel of DV vs the indefinite gradations of Beta SP.

there's no such difference. one has quantization noise and one, well, noise, and both limit the signal the exact same way. dv has a s/n ratio of about 60db, beta sp around 50. people look at schematic figures of what a digtal signal "looks like" and jump to conclusions. those figures are created using nearest neighbor algorithms that cause high frequency aliasing high above what fits in the bandwidth of video, hence it simpy doesn't exist. you can upsample dv to 16 or an infinite number of bits using dithering, which is one way of distributing the quantization noise uniformly thus creating a continuous signal.

I'll tell you one thing, it's a heck of a lot easier to do composite work on Betacam and Betacam SP than DV.

that's a myth from the early days of dv. you just need to resample the chroma properly. you can't use a shortcoming of dv alone to argue the superiority of beta sp. neither format is ideal for keying.

/matt

[wado1942]

you can upsample dv to 16 or an infinite number of bits using dithering, which is one way of distributing the quantization noise uniformly thus creating a continuous signal.

Perhaps but dithering is most efficient using uncompressed signals. When you start using DCT and neighboring pixels get averaged together, dithering just becomes part of the encoding problem.

As for noise figures, yes Beta has somewhat higher noise. But as I've learned in years of experience with audio, people percieve detail well below the noise floor. In the case of digital, the noise floor is also the limit of the dynamic range. Of course dither changes this phenomenon drastically, as I previously stated, the compression makes dither a moot point. Regardless, the noise floor of the lumanance channel has a theoretical limit of 48dB as per the maximum allowance of an 8-bit channel. In practice, internal noise of the camera circuitry prevents true pure black from ever happening. The least significant bit is always toggling so that puts the practical noise floor at 42dB unless of course you're using a computer generated image. The luma channel is the part onto which our eyes lock for percieved sharpness and noise levels.

Anyways, I know you can resample DV to 4:4:4 using filtering that'll smooth the transition from one chroma sample to another but it's not the same as sampling the chroma at a higher rate in the first place. I've composited both digital and analogue video live and in post. I've always had an easier time with the analogue video. Of course I was using hardware made specifically for that job rather than software so that may have something to do with it also.

Regardless, I don't want to make a huge debate between DV and Beta because they're just 2 completely different animals. I stick to my preference of D1. I've never worked with DigiBeta though. Rumor has it DigiBeta's claim of being 10-bit is false. Supposedly the quantization matrix is 10-bit but the signal is dithered to 8-bit for storage. Since Digibeta only compresses about 2:1, the dither would be much more effective in preserving the dynamic range of the matrix. Still, I'd like to work with it some time. Never worked with Blackmagic either, but the ability to record true 10-bit uncompressed would have to put it at the top of the ranks.

[mattias]

Anyways, I know you can resample DV to 4:4:4 using filtering that'll smooth the transition from one chroma sample to another but it's not the same as sampling the chroma at a higher rate in the first place

of course not, but the nearest neighbor resampling is the reason people think you can't key dv. if you use for example a bicubic scheme or even basic chroma smoothing you get results that are very close to sp and since there's less noise i'd say they become about equivalent.

Regardless, I don't want to make a huge debate between DV and Beta because they're just 2 completely different animals.

i agree, but when it comes to image quality they play in the same division. the main reason to chose either is what equipment you own and who you deliver to. investing in beta sp gear is still much more expensive than dv.

/matt