How to Losslessy Re-enode MPEG2?

[usually_quiet]

If you think that, then you don't know what DVDPatcher does. It is used to correct information stored in video headers when there is a mistake. It doesn't change the video itself.

I know what DVDPatcher does - it can force MPEG-2 to required resolution and it is up to decoder how to handle incorrect data however due of fact that error resilience should work correctly then also 704 marked as 720 should be OK - missing data will be replaced by null or something similar - anyway those 16 pixels are usually behind border..

btw
You can operate at DCT/macroblock levels without decoding video - adding 16 black pixels without transcoding is possible.

Keep pulling this kind of bad advice of your arse if you like. It's good for a laugh.

[pandy]

Keep pulling this kind of bad advice of your arse if you like. It's good for a laugh.

Once again - MPEG-2 (H.262) can be edited without transcoding to spatial domain in DCT domain and at the macroblock level it means that macroblocks can be added to already existing structure to extend video from 704 to 720. Don't need to trust me - ask codecs developers or take H.262 spec in a hand and start learning. And laugh is usually good for health.

[AlanHK]

MPEG-2 (H.262) can be edited without transcoding to spatial domain in DCT domain

In theory.
How do you do that in practice?

[pandy]

MPEG-2 (H.262) can be edited without transcoding to spatial domain in DCT domain

In theory.
How do you do that in practice?

I'm not a programmer however there is at least a few's SW for H.262 that operate on DCT without transcoding back to spatial and back to DCT for example M2VRequantiser or ReQuant.

[AlanHK]

I'm not a programmer however there is at least a few's SW for H.262 that operate on DCT without transcoding back to spatial and back to DCT for example M2VRequantiser or ReQuant.

Nothing that can actually change the image or the real resolution (as opposed to faking it in the header), though.

[pandy]

I'm not a programmer however there is at least a few's SW for H.262 that operate on DCT without transcoding back to spatial and back to DCT for example M2VRequantiser or ReQuant.

Nothing that can actually change the image or the real resolution (as opposed to faking it in the header), though.

I wrote the same when mentioning DVDPatcher.
Btw i verified DVD 704 video patched to 720 and seems it works correctly on HW MPEG-2 decoders made by Broadcom, STM and Intel.

[Guest34343]

I'm not a programmer however there is at least a few's SW for H.262 that operate on DCT without transcoding back to spatial and back to DCT for example M2VRequantiser or ReQuant.

Requantizing does not add or delete macroblocks! I have never heard of the latter being done. Did you just make it up?

For one thing, the VBV model would break if you added or deleted video data.

It may presumably be theoretically possible to make all the required adjustments and avoid decoding/recoding the existing macroblocks. The fact that no-one has tried to do it in preference to simple re-encoding of the stream should tell us something about its practicality.

[pandy]

I'm not a programmer however there is at least a few's SW for H.262 that operate on DCT without transcoding back to spatial and back to DCT for example M2VRequantiser or ReQuant.

Requantizing does not add or delete macroblocks! I have never heard of the latter being done. Did you just make it up?

For one thing, the VBV model would break if you added or deleted video data.

First - i've provided example that various operation can be performed on DCT and macroblocks without transcoding (people think that this can't be done ) so please neuron2 do not add anything over this, kindly asking You - be fair and not try to put not my words in my mouth,

second - i assume that when re-quantization is performed, VBV model will also change and need to be "rebuild".

[Guest34343]

Listen up bud, I'm not putting words in anyone's mouth.

Once again - MPEG-2 (H.262) can be edited without transcoding to spatial domain in DCT domain and at the macroblock level it means that macroblocks can be added to already existing structure to extend video from 704 to 720.

You clearly stated that macroblocks can be added.

I object to you calling me "unfair" when I am interested only in technical truth, and I hope you will have the grace to retract that.

[usually_quiet]

Keep pulling this kind of bad advice of your arse if you like. It's good for a laugh.

Once again - MPEG-2 (H.262) can be edited without transcoding to spatial domain in DCT domain and at the macroblock level it means that macroblocks can be added to already existing structure to extend video from 704 to 720. Don't need to trust me - ask codecs developers or take H.262 spec in a hand and start learning. And laugh is usually good for health.

The bad advice is suggesting that the OP should deliberately place incorrect resolution information in the video header in an attempt to force players to treat the encoded video data for every frame as if it were a different resolution than is actually the case.

That theory is based on a foolish assuption that all players have good "error resilience". Hardware players are much less likely than software players to have that, and it appears that the OP will be authoring a Blu-ray disc to watch with his Blu-ray player. The OP even confirmed your cheat didn't work for him in the past.

The cheat won't necessarily work for software either. Windows Media Player is the software player the average person has and will use, but Windows Media Player didn't play the file correctly when I used your cheat. The picture was full of green blocks. The cheat worked for VLC and MPCHC, although all 16 null pixels were tacked onto the right edge of the frame.

Your theoretical ramblings are of no value to the OP or anyone else who needs help with a real-world problem today. There are not any tools that the OP can actually obtain which change the real resolution of video without re-encoding, so what you are suggesting is in fact not possible in the real world today.

I don't write software for encoding or decoding video, but since neuron2 has some experience in this area, I'll take his word that your theories are wrong too.

[Guest34343]

I didn't say the theory is wrong, just that there is no practical implementation that I know of. That's the point you made as well; theories are not of use to the original poster.

[pandy]

Listen up bud, I'm not putting words in anyone's mouth.

Once again - MPEG-2 (H.262) can be edited without transcoding to spatial domain in DCT domain and at the macroblock level it means that macroblocks can be added to already existing structure to extend video from 704 to 720.

You clearly stated that macroblocks can be added.

And i never mentioned about adding or removing macroblocks for requantization - this what You added - claiming that i make equal sign between requantization and adding macroblocks - i've used requantization as an example of operation that can be performed on H.262 without transcoding from DCT to spatial domain and back - that's all i've write previously.
And macroblocks can be added without transforming from DCT to spatial and back. And yes i agree that probably it is faster to do this in spatial domain and perhaps this is only one way available from average user perspective for today.

second - i assume that when re-quantization is performed, VBV model will also change and need to be "rebuild".

You're just wrong about that. The values change but the amount of data remains the same. Why would you have to assume anything?

If after requantization from for example 8Mbps H.262 stream is reduced to 4Mbps then amount of data remains the same?
So how works requantize or m2vrequantiser if amount of data remains not changed? Or how people downsize DVD9 to DVD5 without changing video data which accordingly to You will change VBV?
I never wiritten any code related to H.262 and i never qeustioned Your knowledge on this area but please explain me how placing macroblock with null (i mean DCT representation of black video) from right side will affect more overall amount video data when compared to requantization from 8 to 4Mbps.

I object to you calling me "unfair" when I am interested only in technical truth, and I hope you will have the grace to retract that.

You searching for a truth OK, fine for me but... i will try write this - You extrapolate some claims even if i never write about this like case where i never write that requantization is same transformations like adding macroblocks except fact that both can be performed in the same DCT space without transcoding to spatial and back to DCT. So they can be performed "losslessly" based on fact that we not changing domain - thats all - OK (and requantization is lossy opearation - i will add this to not start new discussion however not due of transcoding from DCT>spatial>DCT), adding information about VBV that need to be addressed when adding macroblocks it is OK and Thank You, but please do not say that i've said that:

Requantizing does not add or delete macroblocks! I have never heard of the latter being done. Did you just make it up?

.

I never wrote that requantization is made by "adding or removing macroblocks" - never.

So this is a reason where i say that in searching for truth You sometimes are unfair mentioning areas even not signaled by me claiming that i wrote something i never wrote - i've said once: perhaps my English is NOK - sorry for that however i don't like fact that i'm blamed for never written thoughts.

[pandy]

The bad advice is suggesting that the OP should deliberately place incorrect resolution information in the video header in an attempt to force players to treat the encoded video data for every frame as if it were a different resolution than is actually the case.

That theory is based on a foolish assuption that all players have good "error resilience". Hardware players are much less likely than software players to have that, and it appears that the OP will be authoring a Blu-ray disc to watch with his Blu-ray player. The OP even confirmed your cheat didn't work for him in the past.

The cheat won't necessarily work for software either. Windows Media Player is the software player the average person has and will use, but Windows Media Player didn't play the file correctly when I used your cheat. The picture was full of green blocks. The cheat worked for VLC and MPCHC, although all 16 null pixels were tacked onto the right edge of the frame.

Your theoretical ramblings are of no value to the OP or anyone else who needs help with a real-world problem today. There are not any tools that the OP can actually obtain which change the real resolution of video without re-encoding, so what you are suggesting is in fact not possible in the real world today.

I don't write software for encoding or decoding video, but since neuron2 has some experience in this area, I'll take his word that your theories are wrong too.

https://forum.videohelp.com/threads/349131-How-to-Losslessy-Re-enode-MPEG2?p=2186362&vi...=1#post2186362

This what i wrote and seems it was to complicated for You to understand. There is a way to increase resolution without re-encoding. I never write that software already exist and is available for free.
Error resilience is practical part of each decent decoder when data can be corrupted or during transfer or when storage medium is partially damaged.

Your theoretical ramblings are of no value to the OP or anyone else who needs help with a real-world problem today.

If You expecting only 100% working solutions with warranty, perhaps You should consider buying them in a shop or specialized company?

Be well.

[usually_quiet]

Your theoretical ramblings are of no value to the OP or anyone else who needs help with a real-world problem today.

If You expecting only 100% working solutions with warranty, perhaps You should consider buying them in a shop or specialized company?

Be well.

The OP has a problem and needs a solution that actually exists. He started of this thread to get help, not to hear your theories. If all you have to offer is a theory with no actual implementation available in the real world, you should have kept it to yourself.

Not many members here do understand the underpinnings of video encoding. If you want to engage in a rigorous theoretical discussion go to doom9 with your theory. They do understand this stuff although it appears some members will expect rigorous technical accuracy.

[FaeGiN]

To thwe original poster, I would go for the advice to decode to LOSSLESS AVI format and use tools like VdubMod, AviDemux perhaps, also as advised already, demux the WAV audio using either of those tools and then do your editing etc before remuxing and compressing to the desired format.
Good luck!

[pandy]

The OP has a problem and needs a solution that actually exists.

Ask them http://www.sri.com/ - perhaps such software exist perhaps not.

He started of this thread to get help, not to hear your theories. If all you have to offer is a theory with no actual implementation available in the real world, you should have kept it to yourself.

https://forum.videohelp.com/threads/349131-How-to-Losslessy-Re-enode-MPEG2?p=2186362&vi...=1#post2186362

Not reading or reading but without understanding... and seems even to be pride for such skill...
Once again - if You have enough money then i can say that this software exist.

Not many members here do understand the underpinnings of video encoding.

You mean that every one on this forum is like You? I don't think so... Don't be so egocentric.
Be well.

[Guest34343]

@pandy

You started making this personal by calling me unfair and saying I put words in your mouth. I did no such thing. And as you now continue with ad-hominems I'll withdraw. You're the winner.

[usually_quiet]

The OP has a problem and needs a solution that actually exists.

Ask them http://www.sri.com/ - perhaps such software exist perhaps not.

He started of this thread to get help, not to hear your theories. If all you have to offer is a theory with no actual implementation available in the real world, you should have kept it to yourself.

https://forum.videohelp.com/threads/349131-How-to-Losslessy-Re-enode-MPEG2?p=2186362&vi...=1#post2186362

Not reading or reading but without understanding... and seems even to be pride for such skill...
Once again - if You have enough money then i can say that this software exist.

Not many members here do understand the underpinnings of video encoding.

You mean that every one on this forum is like You? I don't think so... Don't be so egocentric.
Be well.

My reading comprehension is actually very good. Your ability to make the point you are trying to express using English, is not. In spite of that, it has always been clear that the solution you proposed is not in any form that any ordinary person can readily obtain and use. So again, of what practical value is it to the OP?

This isn't a good forum for rigorous theoretical discussions on video encoding. I have been here long enough to see that most members here really are like me, or even less knowldgeable in this area. neuron2 is the only member here with the right background to really discuss of your theory who participated in this thread. There are only a handful of others I know of who might have participated but they didn't. They don't appear to come here for that. From what I have seen, they must go to doom9 or somewhere else if they want to have those kinds of discussions.

[pandy]

@pandy

You started making this personal by calling me unfair and saying I put words in your mouth. I did no such thing. And as you now continue with ad-hominems I'll withdraw. You're the winner.

Oh please take responsibility for what You wrote - https://forum.videohelp.com/threads/349131-How-to-Losslessy-Re-enode-MPEG2?p=2186595&vi...=1#post2186595

Requantizing does not add or delete macroblocks! I have never heard of the latter being done. Did you just make it up?

You made this personal not me extrapolating things i never wrote and at the same time accusing me about "make it up" things You've just extrapolated. Or perhaps someone else use Your account ?

I can understand "usually_quiet" but i don't understand person who "search for technical truth". Please stop extrapolating things i've never wrote.

Be well.

My reading comprehension is actually very good. Your ability to make the point you are trying to express using English, is not. In spite of that, it has always been clear that the solution you proposed is not in any form that any ordinary person can readily obtain and use. So again, of what practical value is it to the OP?

I respect Your right to express opinions however i only pointed that making changes in resolution for H.262 is possible based on standard - that's all - suddenly only practical (and in assumption very cheap of free) solutions can be written on this forum, also seems that suddenly You behave like admin deciding what is allowed or not.

This isn't a good forum for rigorous theoretical discussions on video encoding. I have been here long enough to see that most members here really are like me, or even less knowldgeable in this area. neuron2 is the only member here with the right background to really discuss of your theory who participated in this thread. There are only a handful of others I know of who might have participated but they didn't. They don't appear to come here for that. From what I have seen, they must go to doom9 or somewhere else if they want to have those kinds of discussions.

Once again - i only expressed fact that it is possible to operate on H.262 without transcoding - i never claimed anything over this. I also proposed practical workarounds and i expressed my opinion after reading draft for BD spec that 720 is a recommendation not mandatory requirement for H.262 (mandatory requirement is compliance with H.262) and perhaps there is no need to transcode anything.


I see no point in further discussion - You don't like me, i don't care about this - just put ignore on me and be well.

[AlanHK]

There is a way to increase resolution without re-encoding. I never write that software already exist and is available for free.

No, there is not a "way to increase resolution without re-encoding". Such software does not exist at any price.

If you want to spend a few years study and writing the code to do so, you could change that state of affairs, but until then this is just a complete waste of time.

[pandy]

No, there is not a "way to increase resolution without re-encoding". Such software does not exist at any price.

For 25k $ You can have working code - some people coding for fun some for money. If You accepting any=25k$ send me PM.

Be well.

[Guest34343]

For 25k $ You can have working code - some people coding for fun some for money. If You accepting any=25k$ send me PM.

Coming from a guy that is not a programmer (https://forum.videohelp.com/threads/349131-How-to-Losslessy-Re-enode-MPEG2?p=2186586&vi...=1#post2186586), this takes the cake for bloviation.

[AlanHK]

For 25k $ You can have working code - some people coding for fun some for money. If You accepting any=25k$ send me PM.

Be well.

Since no one is offering $25K, then it still does not exist and probably never will.


When someone posts here asking how to solve a specific problem with a video file, they aren't in the least bit helped by suggesting they commission a research program to create new software that you imagine might be possible.

[pandy]

For 25k $ You can have working code - some people coding for fun some for money. If You accepting any=25k$ send me PM.

Coming from a guy that is not a programmer (https://forum.videohelp.com/threads/349131-How-to-Losslessy-Re-enode-MPEG2?p=2186586&vi...=1#post2186586), this takes the cake for bloviation.

I just know salary for a programmers. I know that this sound strange but not only in US people programming.

[usually_quiet]

Just keep digging that hole pandy... Pretty soon everyone will have your name on their ignore list.

[pandy]

[
Since no one is offering $25K, then it still does not exist and probably never will.


When someone posts here asking how to solve a specific problem with a video file, they aren't in the least bit helped by suggesting they commission a research program to create new software that you imagine might be possible.

I've just searching for a "technical truth" - "usually_quiet" wrote that this can't be done without re-encoding - i wrote that it can be done without re-coding. That's all, suddenly i see that "technical truth" is not welcomed unless some VIP decide to participate in discussion... even when VIP says - Ok this can be done but no one do things like this or i don't know anything like this and doing things in that way have no sense for me when we can re-encode... now appear new problem - what is allowed or not, new issue is 25k $ is enough to create a working code...

Be well Guys - it can't be done, 25k $ is not enough for writing such code in country where programmer earns per year 15k $, search for "technical truth" is allowed only for VIP's, "usually_quiet' replaced Baldrick - in 2013 there is end of the world (or maybe in 2012 - i forget but who cares) have a nice time .
from my side EOF
--
I feel ashamed holygamer, please accept my apologies for this OT discussion. (this is honest and not sarcastic).

[AlanHK]

I've just searching for a "technical truth" - "usually_quiet" wrote that this can't be done without re-encoding - i wrote that it can be done without re-coding.

And you are wrong. It can't be done.
At this point in time there is no software that does that, and no prospect of there ever being such.

[FaeGiN]

Did I just make the mistake of trying to help when some kids are just having a schoolboy argument and trying to prove who has the biggest brain or something?
Hehe, oh yes silly me.
Retired from this thread before I make the same mistake twice.

[Cornucopia]

@FaeGiN, your suggestion was fine.

The other stuff is unfortunately one of the accompanying burdens of internet discourse...

Scott