Analog to DV Tape vs. Analog directly to computer

[jlorelle]

I am capturing analog videos (VHS) via a Sony MiniDV camera. I can either capture directly to the computer to a raw DV file or capture first to the DV Tape and then to the computer. I had read that it was a good idea to go to tape first because of timecodes, etc. Is that an advantage and are there others? I would like to avoid the two-step process if possible. My long term goal is preserving these family videos forever in a raw (mostly) uncompressed format. Thanks for your help.

[minidv2dvd]

it's your choice. the 2 step would give you minidv tapes you could save and use later if you needed to transfer again. it's also a good storage media for DVavi video.

[JohnnyMalaria]

If the leaning towards tape is for the timecode reason, you should be aware of a weird quirk that affects all the Sony cams I have had. You could capture to the computer while recording to tape but, believe it or not, the timecode that goes on the tape is not the same as that sent via FireWire! It is off by about 3 frames.

If you are only using the tape option to get timecode into the DV stream and, once captured, you won't use the tape ever again, our software (free for your needs) will let you capture and add timecode (based on the computer's clock) - that will save you a step. But, if you do record to tape, you might as well keep it...

[edDV]

If you capture to DV format, you can cut and archive as DV without further losses. There is no need to convert it to anything else.

[jlorelle]

@JohnnyMalaria

Thanks for the info - I have been using your software for capturing and its great. I will look into the timecode option in your app. On longer videos (1.5 hours) the audio tends to drift by about 5 seconds. This happens with any capture app - yours, windv and Pinnacle. It also happens on both Vista and XP. I can only assume the video is too long. I am going to try to capture and split up the dv files and see if that fixes it. Is there a way to stitch the dv files back together when I am done?

[JohnnyMalaria]

Thanks for the kind feedback.

Are you capturing to Type-1 or Type-2 DV AVI files? Type-2 is susceptible to drift because the video and audio are stored as separate streams in the file. In Type-1, they are stored exactly as they are in the original DV so it shouldn't(!) be possible for Type-1 to drift.

So, try the format you aren't using to see if that helps. Splitting and then rejoining the files may not achieve much and you probably need to use some simple command line tool to do it. I've used AVISplit (I think that's its name) to split 13GB files into 1GB chunks. A similar tool would be able to stitch them back together.

John

[jlorelle]

@JohnnyMalaria

Good information - Thank you. I was capturing to Type 2. I just assumed that was better (after all, 2 is better than 1, right?) I'll give Type 1 a try.

[guns1inger]

Also, passing through some cameras produces a time base correction effect which can smooth out capture.

[JohnnyMalaria]

Type-1 came first and simply stores the DV info in the AVI file exactly as-is. To do this, Microsoft had to flag this in the AVI header in a way that older software couldn't understand. Basically, software that uses the DirectShow framework can read them and those that use Video for Windows cannot. The latter has been around since the ark (well, Win3.x) and MS have unsuccessfully tried to banish it (it has lots of limitations). So, to address the incompatibility, MS created the Type-2 version. It still stores the DV as-is but it stores an extra copy of the audio.

[edDV]

Back around 2000-2002 there was much discussion of audio sync slip with long analog to DV captures. Canopus claimed to solve the problem with the ADVC line of DV capture devices and has a switch for "locked" and "unlocked" audio. I forget now exactly how this works but I've never had any audio slip with my ADVC-100.

This may be related to early versions of Sony camcorder DV chipsets. I haven't heard this complaint in recent years.

[jlorelle]

Thanks - is one format better than the other? I am capturing everything into raw DV for long term storage but I convert to Type 2 for working with in Windows. Is the raw DV on the PC the same as the raw on the Mac? GSpot gives different numbers for each.

Interesting about the drift on the Sony cams - mine is from around 2001.

Thanks again.

[minidv2dvd]

Type-1 came first and simply stores the DV info in the AVI file exactly as-is. To do this, Microsoft had to flag this in the AVI header in a way that older software couldn't understand. Basically, software that uses the DirectShow framework can read them and those that use Video for Windows cannot. The latter has been around since the ark (well, Win3.x) and MS have unsuccessfully tried to banish it (it has lots of limitations). So, to address the incompatibility, MS created the Type-2 version. It still stores the DV as-is but it stores an extra copy of the audio.

not sure who is wrong, but you might check this out.

http://en.wikipedia.org/wiki/DV_AVI#DV_AVI

type II is the original format according to those guys.

[JohnnyMalaria]

I can't really tell. MS don't really say anything about it in their 1997 white paper:

http://download.microsoft.com/download/1/6/1/161ba512-40e2-4cc9-843a-923143f3456c/DVAVSPEC.RTF

Within MS, I would have expected Type-1 to have come first because their were pushing ActiveMovie (later DirectShow) and desperately trying to deprecate VfW. The Type-1 is straightforward to handle but VfW can't cope with a single stream containing interleaved material. Type-2 would have been the obvious response to ensuring reluctant backwards compatibility. They really wanted to kill VfW but the likes of Adobe weren't ready or willing to use ActiveMovie etc.

Personally, in 1997 or so I switched from my analog system with Premiere to a great DV editor called EditDV which, oddly, was QuickTime based. So I can't remember the state of play w.r.t. Type-1 and Type-2.

Why's it always so complicated?!

[JohnnyMalaria]

I am capturing everything into raw DV for long term storage but I convert to Type 2 for working with in Windows. Is the raw DV on the PC the same as the raw on the Mac?

Raw DV (.dv files) used by our software are exactly that - a bit-for-bit copy of the DV coming in through the FireWire. There's no header information etc. Folk that have used our software to open Mac raw DV files haven't had problems. I do seem to remember reading somewhere that the frames in some raw DV files on the Mac are padded to an exact multiple of 512 bytes. Strictly, this is no longer raw in my book. I've also read the raw files have a QuickTime header which, again, seems to stretch the definition of raw!

When you say the GSpot numbers are different, can you post the info?

[edDV]

I don't recall much discussion of Type1 vs. Type2 in the audio sync discussions. I do remember that the product with greatest audio sync issues was the Pinnacle DV Moviebox. Other products mentioned were the first generation DataVideo and the Sony DV capture box that was quickly discontinued.

I came up with this discussion in Adam Wilt's DV-FAQ.
Scroll down about half way to "Locked vs unlocked audio". See if that helps.
http://www.adamwilt.com/DV-FAQ-tech.html

[JohnnyMalaria]

If the unlocked audio conforms to DV spec, it cannot drift.

When you play a DV tape in the camcorder, have you ever heard of it drifting?

Locked audio exists for editing purposes. It imposes strict requirements - every frame must have a prescribed number of audio samples. With PAL, this is trivial - each frame has the same number. For NTSC, it is a pain (as usual). For the locked audio to be maintained, simple cuts can only be made at specific frames (a bit like the I-frame requirement in MPEG) otherwise the audio has to be regenerated. Unlocked has a distinct advantage here.

As far as the captured files are concerned, what happens if you seek the file to near the end and play it? Is it in sync? If not and the files are Type-2, it suggests that the second copy of the audio in a separate AVI stream is being read at the wrong location w.r.t. the video. If Type-1 files don't have the problem, that adds further credence to the argument.

Another possibilty is that the capture software is using an incorrect value in the AVI header for the frame rate. With PAL, it is trivial because it is always a whole number (25). NTSC isn't but the AVI header can't store floating point numbers. So a denominator and a numerator are used, e.g., 2997/100. But that is an incorrect value since the frame rate isn't exactly 29.97. More appropriate is 30000/1001. Adobe Premiere was notorious for using the former value and users experienced audio drift on long timelines whether or not the format was DV. So, if the capture software is using the wrong value, the audio will drift. A lot of software uses DirectShow for capture and it takes care of everything and gets it right. Video for Windows-based software has to create the AVI header itself. Any software that can view the AVI header will reveal the numbers used. If they are wrong, they can be changed.

[jlorelle]

@JohnnyMalaria

Fascinating stuff. I will try to capture to the Type 1 and see if the audio drifts on the longer videos. I guess I don't know what I am talking about in reference to GSpot. I was wrong - all .DV files showed the same IEC number regardless if they came from your capture program or the Mac program (Live Capture Plus) - sorry.