Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion. Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules. |
22nd November 2011, 07:45 | #1 | Link |
Registered User
Join Date: Jan 2005
Posts: 23
|
Questions Part 2: Understanding Frame and speed
My full post did not fit all in one posting. So here comes the next part.
I have another Question and it kind of relates to Neuron’s guide in an indirect way about teleciding and Frame order. PS—If I lost you in my calculations google the quick guide where I learned how to make these calculations. And also for my shorthand ( )() means multiplication and ()() / ()() means the product divided by a product. Or even (x/x) / (x/x) which in this case is taking the answer of x/x and then multiplying by the reciprocal because it is division. Not all countries use the same shorthand. And I am terrible at reducing big numbers. Back somewhere on this forum I noticed a posting and some procedures for TFF and BFF. And I know that is not Best Friends Forever or boyfriends forever. LOL. Last year, fields were not even considered in my encoding until literally NOW. And the only reason I WANT to consider them now is because I want to know the Neuron way or the appropriate way of telecine in Avisynth. Ok. So I do the guide’s suggestion. Make a nothing script (source only + AssumeTFF() or AssumeBFF() followed by Separatefields() and open in vdub. Yay. Vdub has been neglected—downloaded but used on rare occasions. I did that in for BOTH TFF and BFF. And it is the same thing in both scripts: left to right as normal but the second field always kind of shifts up slightly or down slightly(not noticeable unless you separated fields). So my question is: Since BOTH AssumeTFF and AssumeBFF does the same thing except shifting up or down would that mean the video is NOT INTERLACED, thus progressive. I know it is progressive since dgavcdec told me but for the sake of knowledge if assumetff and AsumeBFF results in the same left to right direction, can you assume that it is progressive? AND 2.) If it is progessive, should I do “interlace encoding” if burning for DVD? And 3.) can nueron2 add parse DGA to dgavcdec just like dgindex has? I actually encoded a bit of a video, stopped it, put the m2v in dgindex, made a d2v, and loaded the d2v and parsed it to see the order and it said: BFF. But I am not sure if that was done by my encoder QUENC or if that is the way it was when it was just raw h264. That was like a 3 part question. Now, the question about telecine (noun) and teleciding (verb). LOL I do not want to even mention my demuxed audio. I will mention that later. Let us just deal with the video h264. 1.) my question a minute ago was that if something is in progessive, should it be interlaced. So this question is: how do you write the script Telecide for that part? 2.) Now that you answered the question about Telecide what do you recommend as a next part of the script. I ask this because Teleciding usually deals with film to video, not 25 FPS video to 29.97002997002997 ? Again, before 264 video was demuxed it was FILM 23.976, or the long number above, in a mp4 container. So is there a way to speed the 25 h264 video by 4.970029970029967 when using telecide? Once I get the film video at NTSC then I probably can easily speed up the sound to match it. SUMMARY OF ALL QUESTIONS FOR THOSE WHO ARE LOST IN MY WORDINESS: This was from PART ! Questions Resizing (refer back to that posting if need be). 1.) What is the ACTIVE PICTURE AREA in a 640 x 360 matrix since this sampling matrix is not included in the NTSC chart in the article? My guess for active sampling area for 640 x 360 is 646 5/22 x 368. Is this a fair guess? Is it gong to chop off the video? Is it going to add more black stuff? What is your opinion? Why or Why not? Please refer to the procedure as outlined in “A Quick Guide to Digital Video Resolution and Aspect Ratio Conversions. Look at my guesses for PAR as NOT outlined for 640 x 360? Does 640 x 360 actually have an active image area when on an NTSC system? Again, you may ask “Is this a test of our smartness or a guide?” The answer is NO. It is a LONG ESSAY QUESTION that I need help with and requires some calculations so you know where I am coming from. These Questions are from PART 2 Questions: Understanding Frame and Speed 2.) Since BOTH AssumeTFF and AssumeBFF does the same thing except shifting up or down would that mean the video is NOT INTERLACED, thus progressive. I know it is progressive since dgavcdec told me but for the sake of knowledge if assumetff and AsumeBFF results in the same left to right direction, can you assume that it is progressive? AND 2.) If it is progessive, should I do “interlace encoding” ? 3.)can nueron2 add parse DGA to dgavcdec just like dgindex has? I actually encoded a bit of a video, stopped it, put the m2v in dgindex, made a d2v, and loaded the d2v and parsed it to see the order and it said: BFF. But I am not sure if that was done by my encoder QUENC or if that is the way it was when it was just raw h264. 3.) my question a minute ago was that if something is in progressive, should it be interlaced. So this question is: how do you write the script Telecide for that part? 3B.)Now that you answered the question about Telecide what do you recommend as a next part of the script. I ask this because Teleciding usually deals with film to video, not 25 FPS video to 29.97002997002997 ? Again, before 264 video was demuxed it was FILM 23.976, or the long number above, in a mp4 container. So is there a way to speed the 25 h264 video by 4.970029970029967 when using telecide? Once I get the film video at NTSC then I probably can easily speed up the sound to match it. Keep in mind the h264 video when in the MP4 container was film but when demuxed became 25 fps. What method was used for demuxing—Graphedit (gdcl demultiplexor, diavc decoder, wav dest to write as a wav and dump.ax to dump the 264 file and coreaac(I found a freeware version somewhere somehow). For all of those who participate in the reply to this posting I really appreciate it as some people may just read this article, get tired and would rather watch episodes of American Dad or Family Guy or old episodes of The Simpsons than to reply to my posting. For those who do both—you are a megastar supercoder. And again, I will say this one more time: I do not have tons of experience with avisynth. I have picked up bits and pieces through reading things, but really have not bothered to use them such as good old bobby (dgbob). And I used Fdecimate ONCE or twice and that was only because I got a weird video to deal with that was double that rate of film. . Most of the time it has just been source with the HACK and encode and end up with only seeing 90 % of the video on screen. But now maybe those days will change soon. So even though these were several questions to help me with, did you lean anything new? Last edited by Gainnothing; 22nd November 2011 at 08:04. |
22nd November 2011, 17:13 | #2 | Link | |||||
Guest
Join Date: Jan 2002
Posts: 21,901
|
No. That tells you only how it was encoded, and not the nature of the content. For example, you can have progressive content but encode it as fields (interlaced).
Quote:
Quote:
Quote:
Quote:
Quote:
Finally, repeating your questions almost verbatim in the same post is going to turn off a lot of people. If you want your posts to be read, keep them concise. Last edited by Guest; 22nd November 2011 at 17:55. |
|||||
22nd November 2011, 17:25 | #3 | Link | |
Registered User
Join Date: Jan 2006
Posts: 1,867
|
Quote:
|
|
24th November 2011, 08:09 | #5 | Link | |
Registered User
Join Date: Jan 2005
Posts: 23
|
Quote:
But before I revive this thread I just want to ask a second opinion on my statement: If a left to right motion is produced when you drop it into Vdub(making sure the avs script only has source and seperatefileds() and assumetff()....or bff. So if some students on a campus walk like this (left to right):Frame 1 Field one: right motion frame 1 field 2:stationery but shift up Frame 2 field one: motion right Frame 2 Field 2: :stationery but shift up.......and so on and so on..and in BFF,,,,,it does the same thing but shift down(really insignificant)--This is by no doubt about it PROGESSIVE just as Neuron2 stated, right? Just wanted to be sure. I guess the rest of the conversation was a bit confusing but I was only asking how to change 25 FPS to 30 FPS. I thought telecide would do it. Does anyone else know? Again in the mp4 container it was 23.976 but as raw h264 it became 25 FPS, so I have to find a way to get the raw h264 up to 30 or back to 23.976 to match the audio wave file. Is anyone else still confused? Last edited by Gainnothing; 24th November 2011 at 08:13. |
|
24th November 2011, 09:14 | #6 | Link | |
Registered User
Join Date: Aug 2008
Posts: 233
|
Quote:
Yes, sometimes the framerate is lost when demuxing a mp4 file. You should be able to restore the original framerate with: Code:
AssumeFPS(24000, 1001) |
|
24th November 2011, 09:32 | #7 | Link | |
Registered User
Join Date: Jan 2005
Posts: 23
|
Quote:
So could you interpret the script for me. Does it mean assume 24,0000 frames while dropping 1001 frames? Wait a minute. That does not make sense. Oh well. I tried a crack at it? What does it mean? Last edited by Gainnothing; 24th November 2011 at 09:40. |
|
24th November 2011, 10:20 | #9 | Link | |
Registered User
Join Date: Jan 2005
Posts: 23
|
Quote:
I guess maybe Ratios are more accurate than entering in 23.976. 24,000/1001:I should have known that is a ratio. DUH. (off original topic) PS-- people are starting to disagree with what you taught me about the 352 x 360 targeting to DVD. The disagree with your new WIDTH of 516 (typo) and adding borders of 204 on each side. But this is going on here: Not on doom. http://forum.videohelp.com/threads/3...59#post2122759 It seems to me some people can't agree on a finalized formula or something. This is NUTS. Since this is off topic just PM any thoughts. going back to being on topic now. Last edited by Gainnothing; 24th November 2011 at 10:27. |
|
24th November 2011, 19:48 | #10 | Link | |
Registered User
Join Date: Aug 2008
Posts: 233
|
Quote:
If you have 352x360, then it is possible (and I can't say for sure because I have not seen it!) you need to scale it to 704x480 (if the video is 4:3... or it is 16x9 and you want to burn it to widescreen DVD) ... or 704x360 if you want it letterboxed. I was happy to try to help you and I was as patient as could be expected. I am seriously regretting that now. Last edited by vampiredom; 25th November 2011 at 02:27. |
|
25th November 2011, 03:55 | #11 | Link | |
Registered User
Join Date: Jan 2005
Posts: 23
|
Quote:
Sometimes when I post these messages I should read them over several times to look for spelling errors because "the disagree' part does not make sense. That was supposed to be the word "they" referring" to Videohelp. But I am going to look and find my posting that I was trying to refer to. |
|
25th November 2011, 04:01 | #12 | Link |
Registered User
Join Date: Jan 2005
Posts: 23
|
Quote:
352 x 360 = … And just for the sake of fun IF you got this ugly number how would you bump it up to DVD? That would depend on its PAR. If we assume it is 1:1 square pixels, and that 4:3 DVD is 10/11 PAR then it is easy: (480 Target Height / 360 Source Height) = 1.333333333333 Scaling Factor 1.333333333333 * 352 Source Width * 1.1 Reciprocal of PAR) = ~516 So, that would be resized to 516x480 It does not matter that it is not MOD16 because you are going to pad it with AddBorders(102,0,102,0) anyway to make 720x480 ... unless it is not really square pixels to begin with, in which case all bets are off. Last edited by vampiredom; 23rd November 2011 at 04:40. This is the posting I was referring to. Sorry, I don't know how to quote when it is from another post. But I think I might have misquoted you, so maybe there was a communication barrier. But from your statement above, I was trying to get people on videohelp to understand that the source with a wdith of 352 gets resized to 516( and add borders) to get to the 704 x 480.. And they kept on telling me "No." But I stated to the people at videohelp that this is what you told me or this what the people at Doom told me. Maybe I misunderstood . Last edited by Gainnothing; 25th November 2011 at 04:23. |
25th November 2011, 04:12 | #13 | Link |
Registered User
Join Date: Jan 2005
Posts: 23
|
Bye the way, I did your AssumeFPS and it works pretty well. I made a VOB through IFOEDIT(multiplexed) and the audio is synced perfectly. And it is MOSTLY smooth but I noticed maybe a few places where the picture may slightly jump a little ahead. It does not happen often.
I think it is an ENCODER thing. So I tried encoding with Quenc high quality 2 pass instead of a 1 pass--better but some jumps. I am going to try Procoder express to see if it does some jumps. But again, I am very happy that the audio is synced. IFOEDIT multiplexes VERY well. Is there a script in avisynth to try to correct this "jumping" that the video does? Again, maybe procoder express won't jump. I have not tried it yet. Last edited by Gainnothing; 25th November 2011 at 04:14. |
27th November 2011, 08:41 | #15 | Link |
Registered User
Join Date: Jan 2005
Posts: 23
|
I see and the reason why you brought this up this example you showed me earlier was because you did not know what type of pixels my source was using since it was kind of a hypothetical post and I did not have an actual video to show. I got it now. Therefore, you showed me what to do if it they were square pixels.
Ok. So if I did run dgindex or dgavc and and it said that colorimentory was SMPTE 170M would that have meant anything? The actual SMPTE 170m source was a 640 x 360 that I resized to 704 x 360, added some borders, and it looked pretty well. But if there was a 352 x 360 source that was non-square pixels would there be a.)a way to determine that they are indeed non-square pixels using the source ? Omitted second question because I saw the answer above. Also; reverting back to the topic of smooth video. Does avisynth have a plugin that can be used to smooth out the playback of a video that every once in a while may "jerk or skip ahead a tiny bit?" The assumeFPS script you gave me worked pretty well, but I have noticed some slight jumps every once in awhile. Any suggestions? This video that I am referring to now played fine at 23.976 on my standalone player. I found that maybe adding pull down by using software may not be necessary since the dvd player and the xbox played it fine. PS. Think IFOedit did a great job of muxing the streams. The only strange thing was that one of the VOB files that were created said it was 720 x 600, which I thought was pretty odd since that is not a DVD spec. I am not sure what script that came from. I thought it was kind of odd because one of the videos(an AVS scipt I created was only for 704 x 480(with borders)) and the other script was 704 x 360 (with borders). PS Ifoedit saw these as Pan and Scan, letterboxed. The original source was a tv series as a Mp4 640 x 360 before I got processed and demxed, just in case you wanted to know. PS--I found a way to add videos using IFOedit and TITLEWRITER, and they said that it was a convoluted way of doing things. But I just wanted to see if the process I learned from another forum would work, and it indeed worked. I even posted it on videohelp.com. Thanks. Last edited by Gainnothing; 27th November 2011 at 09:17. |
27th November 2011, 10:30 | #16 | Link | ||
Registered User
Join Date: Aug 2008
Posts: 233
|
Quote:
Quote:
|
||
28th November 2011, 09:44 | #17 | Link | |
Registered User
Join Date: Jan 2005
Posts: 23
|
Quote:
When I fed the AVS script to play in MPC, it plays perfectly. I used Neuron's tool DGAVCdec--force filmm hence using a dga source. AFTER ENCODING to a m2v, the video was MOSTLY smooth with some jumps. So when the M2V or VOB file is played in ANY PLAYER, it does the same jumps. It is hard for me to describe the jumps. But I will put it this way:the jump can happen as fast as a person can blink. The video does not stop for the jump. Let us say you see some candles that are camera right and the camera is on a dolly or "train tracks" and the camera moves left on the tracks and the candles slowly disappear and then you see slowly some students and a student is writing on the chalkboard is what becomes the new subject for the camera and the candles are no longer seen. I think you can draw a picture in your head now. Most of all actions happen in real time. But I have noticed the candles shift a little more to the right as if skipping a frame, going slightly faster then the camera on the train tracks is going. Hence, it skips a small action that happens in a blink of an eye. Let me see if I can "paint" you another example. let us say it looks as if a student is about to write on a chalkboard and it looks like his hand is about maybe a foot away from right away from writing on a chalkboard, and then all of a sudden the students hand is already writing on the chalkboard, thus making it look like it jumped the 1 foot--it just happened in a blink of an eye. In other words, this skipping of what seems to be one foot or less from the chalkboard, did not happen in "real time." It skipped. Maybe it skipped a little less then 20 frames or something. By the way, I do NOT mean a FOOT as in a foot of the screen. No no no. I meant, if I were to jump into the scene and become a character or if i was on set, that would be how far the student's hand is from the chalkboard. No--not an actual foot length on my screen. But this so called foot when actually measured on my tv screen may be 0.1 - 0.3 inch or less on a tv screen. I don't know. But this is the best way I can paint you a picture. Again, maybe like 95% of the video happens in a real time structure, but there is like this 2-3% margin where every once in awhile, it may skip. And it did not skip from the original source--it happened in the encoding process. And it probably was not DGAVC because I played in in MPC before encoding it and the AVS script played fine. You know how sometimes while your video is playing, your processor may have lots of processes it is running and memory is not at its peak of potential, and it kind of just stops or jumps a video just for a a blink of an eye. This is what it is like. But less dramatic. I made sure to play the video several times to make sure it was not a processor hang. But I do have a 2.6 GHZ X2 Core 2 duo (~5 ghz) so processor hangs are not that common. One more little bit of information. I encoded the same video and put two different versions on DVD. And both versions do insignificant jumps every once in awhile that may only happen maybe 4 or 5 minutes into the video, so it is not that common. But the jumps happen in different places for both versions of this episode I made a DVD for. That means there were two different AVS scripts and 2 VST sets--one full and one 704 x 360. Both scripts originated from 1 original mp4 source. The audio track is in sync when muxed into a VOB file (m2v and ac3).It is also in sync on my standalone DVD player including my XBOX. Last edited by Gainnothing; 28th November 2011 at 10:24. |
|
|
|