Is there am advantage of DVCpro vs DV?

[Moppelkotze]

Hello,

I have a movie which I captured from VHS and it saved it once as DV and then also as DVCpro. I uploaded a 1:45min sequence for you, if someone would have a look, please. (about 640MB)
http://www.jottacloud.com/p/eichhoernchen/null

My question, I know, that DV is 4:2:2 and DVCpro is 4:2:0 chroma subsampling, but is there any influence, difference, advantage consequence this will have for the video or me? (I plan to later convert them to h.264 for long time archiving).

Sorry, for the stupid question.

Thanks

[WorBry]

DV uses 4:2:0 chroma sub-sampling for PAL and 4:1:1 for NTSC, whereas DVCPro(25) uses 4:1:1 for both standards. It's DVCPro50 that uses 4:2:2.

https://en.wikipedia.org/wiki/DV#DVCPRO

https://www.adamwilt.com/DV-FAQ-tech.html#colorSampling

So, if I understand your question correctly, you have this VHS movie separately captured in both DV and DVCPro formats and you are just wanting to know which would be the better source for archiving as H264 only....or are you intending to do some processing in between and in/with what system ?

Either way, if it is DVCPro(25), and not DVCPro50, that you captured in, I can't imagine that it would be any better quality than DV - do you see any difference between the two captures ?

My only advice, if you are planning to use AVISynth for processing would be to use Cedocida's DV codec as the decoder for DV, DVCPro(25) or DVCPro50.

...P.S.

Hello,
I uploaded a 1:45min sequence for you, if someone would have a look, please. (about 640MB)
http://www.jottacloud.com/p/eichhoernchen/null


It's a bit much to expect people to download such a large file and not altogether clear what you are wanting people to look at. If it is simply to garner opinion as to whether there is any discernible difference in the quality of the two captures, then short (1-2 sec) frame-matched samples of each would suffice.

[Moppelkotze]

MediaInfo just say DVCpro, so I don't know, if it is 25 or 50.

I uploaded 1:45min. because I thought it is important, because it is not a static image, but in 1-2sec. there was not enough changes (it was the same motive with the same colours). Also usually, I get laughed about, because my Internet bandwidth is so small. I know that I am very slow at moving with time, so I expected other users to have at least DSL6000 or even DSL16000, making it like "in the twinkling of an eye". But I can understand, that it is a bit much to expect someone to watch it, since it is not just like quickly comparing two snapshots or so. Sorry.

Now, back to my question.

I know there is the rule, that you should never insert a transformation between source codec and final codec, because it will worsen the picture, which is quite obvious.
However, there is a certain capture+edit programm that captures/ saves the movie as *DV to HDD, then you can cut it and now, after editing when you export it highest-quality/same-as-source it gets a DVCpro file.
I wonder why this happens, because it being *DV it could just leave the cutted scenes and write the rest to HDD. For my understanding, the great thing about *DV was, that you do not have to re-encode it, because there are only i-frames that can be cut everywhere and glued together at every position.

I know that I can use another programm that will just do that: capture VHS-analog-casette and save as *DV-digital to HDD, cut out trash/scenes and write file new without trash and the result will be a shorter *DV file, not DVcpro. But I still wonder, what advantage it has in that certain other programm to go from DV to DVCpro.

Here is a spectrometry of both http://forum.slashcam.de/files/thumbs/t_camera_matching_189.png (DV vs DVCpro)

5sec https://www.dropbox.com/sh/ybilzubqx0nzrtf/AABshaDBVWN8F5CGIrqrD-n0a?dl=0

MediaInfo data:

thanks in advance, a lot!

[vivan]

Those 2 files are bit-identical, they differ only in the end.

[WorBry]

MediaInfo just say DVCpro, so I don't know, if it is 25 or 50

Examining your uploaded samples.

Clip20.dv is PAL DV:
http://i.imgur.com/GVuxYUx.png

Clip20a.dv is PAL DVCPro(25):
http://i.imgur.com/36wqkuI.png

Note the video bitrate 24.4Mbps. If it was DVCPro50, it would be double that.

And like Vivan said, the clips are bit identical.

But I still wonder, what advantage it has in that certain other programm to go from DV to DVCpro.

I don't know which programs you are referring to, but probably none.

[Moppelkotze]

Thanks for watching at the snippets.

Why did you add the mediaInfo xcreenshot? I uploaded the same as an attachment to this thread. The only thing I did differet was, that I toggled to "text" instead of "simple view", because I thought you need all info.

Since you probably won't believe me that it is mediainfo. I will add the simple view for you, too. So you see the characteristic design of the mediaInfo window.

Edit: the server was down for some minuts so I couldn't complete the post.

[Moppelkotze]

I don't know which programs you are referring to, but probably none.
Thanks for your kindness, I know that I am stupid and that I made a fool out of me, by asking questions, but I am not a liar. That is quite mean of you.

The programm (i.e. application) that captures and you can edit with and then it will write the file afterwards new as DVCpro instead of the intial DV is iMovie versions 3, 5 and 6. (iMovie 8 changed the whole thing and is less usefull then the older versions. In fact it touches your content without asking. iMovie≥08 puts interlacing on every material during export, no matter what you want).

The programm I use just for cutting instead of iMovie, and the keeps the DV file and only removes the parts one cutted out is "Mpegstreamclip".

Now, by saying that I use Mac applications, I seem even more like a fool to you, I know. But I started with Macs before the iPod/iPhone hype.

Also, I know these are no pro applications... but I never said I was a Pro, I am just a stupid dude capturing his VHS cassettes, because the content on it is not available today anymore and asking a Pro to do it for me, the costs would not be reasonable in comparison to what I have on the tapes (including regarding quality).

[WorBry]

Thanks for your kindness, I know that I am stupid and that I made a fool out of me, by asking questions, but I am not a liar. That is quite mean of you.


Whoaaa :eek: I wasn't implying anything of the sort. I was merely answering your question:

... But I still wonder, what advantage it has in that certain other programm to go from DV to DVCpro.


To which my response was "....probably none" i.e. I could not see what particular advantage that would be. That's all.


Why did you add the mediaInfo xcreenshot?

Simply to show you why the second clip was DVPro(25). Nothing more, nothing less. If your MediaInfo attachments had been approved when I posted, I would have just referred to those.

Sorry mate, I don't need this. I was only trying to help you out :(

Maybe someone familiar with iMovie could address the points you raise.

[Moppelkotze]

SORRY!!!

This was a big misunderstandig on my part, because I am used to be treated like that on expert forums.

I just made a wrong connection. You said " I don't know what programms you use, probably none" (I though you were implying that I was using no programms and saying I was making it up).

Really, can you accept my pardon?

The other thing was that your screenshot showed just "DVCpro" same as mine and not "DVCpro(25)". I overlooked, that you said, that if that "DVCpro" meant 50, then the bitrate would be double.

I also didn't know that my attachments were not approved. I was just thinking they don't appear inside the post-box and one (readers, too) would have to click the little red titles of the attachments.

Sorry again! I really mean it. :thanks::stupid::o

[WorBry]

SORRY!!!
Really, can you accept my pardon?


On one condition - you stop putting yourself down ;)

There's nothing wrong with asking questions.

[Ghitulescu]

The only advantage DVCpro has over DV is the width of the tracks and the better quality of the tapes - this would ensure the recordings are better recorded and read back (0 dropouts - at least it is alleged to be so). All other parameters are almost identical.

[Moppelkotze]

On one condition - you stop putting yourself down ;)

There's nothing wrong with asking questions.
OK :)
The only advantage DVCpro has over DV is the width of the tracks and the better quality of the tapes - this would ensure the recordings are better recorded and read back (0 dropouts - at least it is alleged to be so). All other parameters are almost identical.

@all
Does the spectrometry tell a professional something (despite that the two samples are different) and would that tell me?
http://forum.slashcam.de/files/thumbs/t_camera_matching_189.png

[WorBry]

OK :)
@all
Does the spectrometry tell a professional something (despite that the two samples are different) and would that tell me?
http://forum.slashcam.de/files/thumbs/t_camera_matching_189.png

What actually are the videos represented in this composite Vectorscope plot ? And what was the system used to generate this plot?

In your earlier post you indicated:


Here is a spectrometry of both http://forum.slashcam.de/files/thumbs/t_camera_matching_189.png (DV vs DVCpro)


But it can't possibly represent the DV and DVCPro(25) samples that you uploaded. Looks like they are derived from some color test pattern, and the URL links to slashcam.de. I wonder, is this a Vectoscope plot generated from some camera/camcorder comparison using a reference chart?

[Moppelkotze]

Hi,

in the slashcam forum there was a thread, two years ago, where I asked in this forum, if they can see a difference, if I sent them the same file, once saved as *DV file and additionally as a DVCpro file.
I wanted to know, wether the picture gets lowered in quality, when iMovie exports the existing *DV file to DVCpro, when it saves the done cuts and edits - that is exporting out of the project inside iMovie to a seperate file on the HDD, which can then be used by other programms.

One member answered, that one could probably only see a difference, if they used a waveform monitor.

So I started a second thread on slashcam.de, asking, if someone could use a waveform monitor to test this.

Then someone said, this could not be done with a waveform monitor and only vaguely on a view-monitor (don't know if this is the right word), BUT! that he used a vectorscope and appended the picture for me. [I was wrong here, using the word spectroscope in my post above]
The source was two pieces of the same time code scene. One DV one DVCpro.

As you can see there was a difference, though I don't know, what I can make out of it. (well now, that you see that this pircture doesn't make a lot of sense... I guess the comparison didn't have a point anyway.)

Would there be a quality loss exporting DV to DVCpro (same bitrate used)? Because for me exporting from an existing DV file to a DVCpro file, when you are planning to later encode it anyway (example from DV to h.264), doesn't make much sense to me.


--------------------
PS: I gave the person two versions of the same documentary I posted here on doom9. If he used them or maybe didn't understand me and used a DV file and a DVCpro file he had laying arround just to show me that they differ - I don't know.

It might not be the exact same time code that I posted in doom9, but since it is the same files, they should be comparable. If it differed back then it should differ today.

[WorBry]

OK, I'm confused now.

If I understand correctly, according to your first post, the two short clips that you uploaded were samples of your VHS transfer captured in DV and DVPro(25) format. As we have established, those clips are identical. That is to say - comparative metric analysis reveals no difference in the decode outputs of the two clips.

But what you (also) want to know is whether there is any loss incurred in the conversion of DV to DVCPro(25) by iMovie. Correct?

Quite honestly I would forget about this ambiguous vectorscope plot that someone produced way back when. Without seeing the actual images it was derived from and details of how it was produced, it's pointless attempting to derive anything meaningful from it.

Simplest "in-use" test would be to put your DV source into iMovie and, without applying any edit effects, output to DVPro(25). Then upload frame matched samples of each...like you did before with the VHS > DV and DVCPro(25) captures...and I'll analyze them. OK?

[Ghitulescu]

One of the reasons why I do not have a Mac/Apple product is that I do not know and (given their behaviour) never will be sure what they do in the background with my files. I am sure that there are good software to work at pro level with Macs, yet they are very expensive, compared with the offer on windows.

As several times mentioned - if both clips are identical, but some characteristics seem not to be identical (after processing), the culprit MUST be looked where the changes have been carried out, id est where iMovie did a "plain copy". My guess is that iMovie converted the colour information internally (therefore twice, once at import, the other one at export) to fit the storage requirements (probably imposed by quicklime or other internal format iMovie uses). It also may be a misconfiguration or a bad/rapid implementation of the PAL import filter.

Leaving all this philosophically cram aside, get one of the files and go ahead.

The second issue is what to do if iMovies screws the colours? Nothing, if the result is ok, then the losses are only minimal and see-able only on a scope. VHS is anyway not the perfect source in terms of chroma...

[WorBry]

Exactly and a nice summary of the possible factors involved.

Incidentally, I came across this thread posing the same question - so you are not alone:

http://www.kenstone.net/discussions/read.php?3,44206,44217

[WorBry]

Simplest "in-use" test would be to put your DV source into iMovie and, without applying any edit effects, output to DVPro(25). Then upload frame matched samples of each...like you did before with the VHS > DV and DVCPro(25) captures...and I'll analyze them. OK?

No response ??

[Moppelkotze]

I am busy, but I will come back to this thread in the next 1-2 days.

I made a screen-video of what I did today, so you can follow what I was talking about. I will upload it tomorrow.

Twice when I wanted to write back here, the server was not responding.

[Moppelkotze]

OK, I'm confused now.

1. If I understand correctly, according to your first post, the two short clips that you uploaded were samples of your VHS transfer captured in DV and DVPro(25) format. As we have established, those clips are identical. That is to say - comparative metric analysis reveals no difference in the decode outputs of the two clips.

2. But what you (also) want to know is whether there is any loss incurred in the conversion of DV to DVCPro(25) by iMovie. Correct?

3. Quite honestly I would forget about this ambiguous vectorscope plot that someone produced way back when. Without seeing the actual images it was derived from and details of how it was produced, it's pointless attempting to derive anything meaningful from it.

4. Simplest "in-use" test would be to put your DV source into iMovie and, without applying any edit effects, output to DVPro(25). Then upload frame matched samples of each...like you did before with the VHS > DV and DVCPro(25) captures...and I'll analyze them. OK?
1. this was what I said, to simplify the question, to save the discussion from confusion. I usually include to much info and people can't follow me, so I tried to sum it up. I know this was a bit stupid, since a) making two digital results out of an analog source called VHS is different then b) making one digital result (DV) out of a analog source VHS and then make another digital result (DVCpro) out of the previous digital result (DV).

2. Yes, this was actually my first question, that occured, when I once noticed, that iMovie is not just cutting and re-saving the file, but also doing a conversion from DV to DVCpro, which seems unreasonable to me. It could still just re-save/re-write as DV file while omitting the cut out scenes. - Obviously I asked why it does this illogical step and secondly, if there is a loss of quality (since the thumb rule seems to be, always convert to what you last intended format is, if possible).

3. probably better to do as you both say

4. OK, I will do. Is it sufficient enough, if I cut out one frame from the same time code or should I upload the whole file, so you can be sure I selected the same frame?
1. One of the reasons why I do not have a Mac/Apple product is that I do not know and (given their behaviour) never will be sure what they do in the background with my files. I am sure that there are good software to work at pro level with Macs, yet they are very expensive, compared with the offer on windows.

2. As several times mentioned - if both clips are identical, but some characteristics seem not to be identical (after processing), the culprit MUST be looked where the changes have been carried out, id est where iMovie did a "plain copy". My guess is that iMovie converted the colour information internally (therefore twice, once at import, the other one at export) to fit the storage requirements (probably imposed by quicklime or other internal format iMovie uses). It also may be a misconfiguration or a bad/rapid implementation of the PAL import filter.

3. Leaving all this philosophically cram aside, get one of the files and go ahead.

4. The second issue is what to do if iMovies screws the colours? Nothing, if the result is ok, then the losses are only minimal and see-able only on a scope. VHS is anyway not the perfect source in terms of chroma...
1. I agree. But iMovie has always been included for free and for a free App it was better than Win Movie-Maker. Of course that doesn't make away with the point that there is probably better FREE third party software for Win. Pros always had the possibility to buy Final Cut Pro (apple, which included industry standards, until they crippled it along their focus on mobile devices -> http://arstechnica.com/apple/2012/01/video-pros-apple-needs-to-acknowledge-the-pro-industry-and-fast/ ) or they would use Pro-Apps like Pros in the Windows-World do too, like Adobe or Avid (which isn't free for Windows, too).
I don't want to go into too much detail, I am derailing the thread. But Apple even made it worse. While iMovie (though free and only in its first days targetted at and used by Pros, too and later only at the average customer anymore and ultimately since vers 10 or so at the noob who wants to cut movies from his mobile phone) it had a braod feature list and you could do a lot. iMovie until Vers. 6 used DV as internal raw file format until arround 2006. Then with version 8 they introduced ProRes, but a nasty thing, too. Version 8 (2008) started to interlace every incoming material, throwing away half the content, so to say. With vers. 9 they focused their efforts on the iPhone-market and the iMovie included on Mac computers today is crippled even more.
I used version 6, so there should be nothing done to the content on capture.

Here we come to your point 2.
Allthough I use Vers 6, there could be something like you described. Most likely a bad implementation of PAL!

3.+4. Would probably be the best. Just take the file as it is. Since there seems to be no visual difference. But I am a nit picker and it bugs me, if there is something illogical (like the uneccessary conversion to DVCpro), probably just because it COULD be that it has an influence, even, if people say there is something wrong, but just ignore it and "you won't see it".
Exactly and a nice summary of the possible factors involved.

1. Incidentally, I came across this thread posing the same question - so you are not alone:

http://www.kenstone.net/discussions/read.php?3,44206,44217
This is a thread I started back then. I was so unsatisfied, that I thought "ok, there seems to be no solution to the problem, so I might just turn to it another time, when I have more energy and time". So I left the issue be and now I am back at it.

The reason why no other person noticed that in his work with iMovie is probably, that Pros would use other stuff or just ignore and no-pros would not bother to look at it. Most likely though is, that nobody notices, since you really have to look for the problem to know that there is a problem.
The only reason I discovered that iMovie is doing an uneccessary conversion at export is, because I wondered "hm, DV should be something like mpeg2 or jpegs in a row, so you just can cut everywhere and just glue the stuff together, so why does it take that long?". So I looked at the source and the resulting file with mediainfo. (PS: an app like mpegstreamclip for all platforms btw., just needs seconds, because it just does what I expected, it cuts and when it saves the file again, it cuts out what one wnats to be left out.)


NOW breathe in and out once :)

Again, what did I do:
1. put S-VHS tape in VCR and connect composite cable (red, yellow, white) with A/D-box. (! I know YUV Y, R-Y, B-Y i.e. YCbCr would be better - I also know, that VHS is already crappy and composite is not handing over the best possible of the signal, so why do I worry about stuff that iMovie does? You are right. Using composite makes it absurd to look after DV vs DVCpro, since chroma etc. are only handeled in YCbCr. So my question might be pointless anyway, but there is still the question why digitally there is this change by iMovie from DV to DVCpro)
2. put cable into Thompson Canopus ADVC-300 and connect Firewire cable to Mac
3. on Mac open iMovie and select "capzure device: ADVC-300" and hit capture
4. save file
5. the appended video shows what follwed then.

This is what has been done. What is done afterwards is shown in the movie I ulpoaded for you. It shows that the file iMovie uses internally is a DV file (generated during digital import of the analog VHS tape), visible to the user inside the edit-Window are only something like "tags". The user sees what he cuts and what effects he does. He can even delete 50 minutes of a 60min. video and will be shown that the video is now 10min.. - "under the hood" the file is still untouched and 60min. long.
Only, if you export the file, the stuff you see will be applied (back until version 3 or 4 when Disk space was rare, iMovie actually deleted scenes physically and the project file under the hood would get smaller, but still be a DV file).

SORRY for the long text!

Here is the movie of my steps done (4MB): https://www.dropbox.com/s/okyi1t66tyj0m7t/do9%20show.mp4.zip?dl=0

PS: tell me, when you have looked at it, so I can delete the video from the server. / also notice I maximized the window so much, because I didn't want to show my deskop ;)

[Moppelkotze]

PS: link for frame by frame comparison https://www.dropbox.com/sh/6jh59umgo1xuaqj/AACaaCcG24eaM7VTinNUEnvBa?dl=0

this link shows the same frame of the "original" (the file that iMovie wrote to disk while capturing the VHS. This is the file hidden under package symbol, not directly user accessable) and the "after" (when you export the movie/frame out of iMovie).

[WorBry]

Here is the movie of my steps done (4MB): https://www.dropbox.com/s/okyi1t66tyj0m7t/do9%20show.mp4.zip?dl=0


OK, so, to cut a long story short, you've done what was suggested:


Simplest "in-use" test would be to put your DV source into iMovie and, without applying any edit effects, output to DVPro(25)...

...and you now have your input DV file and "pass-through" export DVCPro(25) file. Fine.

Is it sufficient enough, if I cut out one frame from the same time code or should I upload the whole file, so you can be sure I selected the same frame?

No, a single frame is not sufficient. Again, short (5 sec) frame-matched samples from the two files would suffice, but if you are not confident doing that (how did you cut the samples before btw?), well I suppose, yes go ahead and upload the full files and I'll have a look at them. That way there's no question of mismatching. OK?

[WorBry]

PS: link for frame by frame comparison https://www.dropbox.com/sh/6jh59umgo1xuaqj/AACaaCcG24eaM7VTinNUEnvBa?dl=0

this link shows the same frame of the "original" (the file that iMovie wrote to disk while capturing the VHS. This is the file hidden under package symbol, not directly user accessable) and the "after" (when you export the movie/frame out of iMovie).

You bumped my last post. Like I said, upload the full files so we can reach a conclusion on this.

[Moppelkotze]

OK, sorry.

here is 5sec of both (capture and passthrough) https://www.dropbox.com/sh/l21vgmcsctj5p07/AACKrlKELNiHuWNnQijhjdSva?dl=0

I know this is a rethoric question, but how I did it? I used the application mpegstreamclip to cut off the frames and now 5sec of both clips (the one untouched sitting hidden in the project folder and the passthrough being visible on whereever one selects to save it). ;)

If I used iMovie the files would be rewritten and transformed. ;)

[WorBry]

OK, sorry.

here is 5sec of both (capture and passthrough) https://www.dropbox.com/sh/l21vgmcsctj5p07/AACKrlKELNiHuWNnQijhjdSva?dl=0


The "capture" (DV) and "pass-through" ('DVCPro') samples are bit identical. What's more, both have 4:2:0 chroma subsampling, which is interesting, as by definition DVCPro(25) is 4:1:1.

So why MediaInfo reports the export file as being "DVCPro" (Commercial Name), I really not sure. Something to do with "locked audio"... maybe ? Any thoughts Ghitulescu ? You're more knowledgeable about this type of thing than I.

Regardless, I think it's safe to conclude that it is a direct stream copy of the source DV. DV is not a lossless format, and if there was re-encoding going-on, you would expect to pick up some loss in quality on metric analysis. And there is none.

That basic question answered, what might be interesting to determine is whether iMovie truly supports so called "Smart Rendering" of DV, which most Windows-based NLE's do, by default or as an option. By that I mean, if an "effect" is applied to a section of a clip on the timeline, only those frames changed by the effect are re-encoded and the "untouched" frames are passed through (copied). So if only cut edits are applied, there is no re-encoding at all.

[Ghitulescu]

I do not know.
Locked audio is not a problem in non-linear editing, it was developed for real-time ones.

Now that both files are on the HDD, and they are identical, I would use any of them for future processing. The other one, if the space allows, I would keep it as a backup in the case of a complete "ruin" by editing.

[Moppelkotze]

Thank you both very, very, very much!!!!

Now I know, that I can't go wrong letting iMovie do the export.
I started doing some exports and cuts with mpegstreamclips, in the past, after I discovered that they are labeled DVC(pro) bei Mediainfo, if they come from iMovie. Because I thought since the export with mpegstreamclip is still being recognized as DV afterwards, I will be sure to not loose something or alter the file other than cutting.

That basic question answered, what might be interesting to determine is whether iMovie truly supports so called "Smart Rendering" of DV, which most Windows-based NLE's do, by default or as an option. By that I mean, if an "effect" is applied to a section of a clip on the timeline, only those frames changed by the effect are re-encoded and the "untouched" frames are passed through (copied). So if only cut edits are applied, there is no re-encoding at all.
I would have to compare the time it takes to do a passthrough (or writing a cutted file) with iMovie and how long it takes with mpegstreamclip. Maybe I'll do this out of curiosity in the next week and then post back, as this would give a hint.

Hehe, if I were in your country I would sent you a cheap old Mac (they go for 20-40,-EUR and USD) to try iMovie. :D But that would serve nothing more than curiosity and had little practical worth for you of course :)

[WorBry]

Thank you both very, very, very much!!!!

You're welcome.

Hehe, if I were in your country I would sent you a cheap old Mac...

Thanks, but I already have a rain coat ;)
(http://www.yourdictionary.com/mac)

Actually the wife has just bought an iMAC, but I'm not allowed to dabble with it......yet.

[Moppelkotze]

You're welcome.



Thanks, but I already have a rain coat ;)
(http://www.yourdictionary.com/mac)
:D

Actually the wife has just bought an iMAC, but I'm not allowed to dabble with it......yet.
Beware, the included modern iMovie version is very limited anyway... (guessing, that she has a newer model).
Agree which each other to create a second account for yourself ;) You could also connect a HDD/SSD via USB3 and install Mac OS X there, too, then start from the external one, if you want to mess arround with it without anyoing your wife. :)

[WorBry]

Thanks, but I'm really not interested to use it for anything other than maybe Skype - and then purely for convenience when it comes to family video calls. Ironically, I spent an hour last evening getting her set-up with MS Office for Mac.