Frage von martinolex:
Hello together.
There are some patents and articles about where it comes from DCT-based video formats to transcode without iDCT / DCT to make. There are the links:
http://www.freepatentsonline.com/EP1155573.pdf
http://cat.inist.fr/?aModele=afficheN&cpsidt=1823092
http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=4039371
It was even of CyberLink PowerVCR II as a
"CyberLink's ultra-fast DV transcoding engine implements. Unfortunately, this only works Transcoder) of tape (IEEE1394. Transcoder is indeed a dll in ActiveX format shows itself not as DirectShow filter.
Maybe someone knows such a transcoder that can transcode the DV AVI file to an MPEG-2 (*. mpg) file without iDCT / DCT? Would be super if it could persist over MPEG-2 only from I-frames and contain exactly the same as the DV video data source.
Thank you very much
Best regards,
martinolex
Antwort von robbie:
as an alternative would be here to offer even downright imx .. it's just 422nd ..
Antwort von martinolex:
as an alternative would be here to offer even downright imx .. it's just 422nd .. Is there such a transcoder for IMX? Combination of hardware or software? What is the thing, and where there's tech. Info? Thank you very much.
IMX, although I can not use next, but then unfortunately the information on such a transcoder interested me anyway.
Greeting
martinolex
Antwort von RickyMartini:
MPEG1/2/4 and H.264/AVC is based in part on the DCT / iDCT!
Why will not you understand?
"The I-frames are compressed using the DCT but PRECISELY! I urgently ask you to follow this link to and you do something about MPEG compression familiar!
WITHOUT MPEG DCT is NOT! http://vsr.informatik.tu-chemnitz.de/~jan/MPEG/HTML/mpeg_tech.html
Antwort von Marco:
Stony me, but what I can gather all these documents, especially this detailed PDF is that the objective of this patent:
- A Videobitstream, which corresponds to a first DCT-based compression algorithm (DV)
to convert
- A Videobitstream, which corresponds to a second DCT-based compression algorithm, which is different from the first compression algorithm (MPEG).
The processes that need to be described this are:
- A Decoding DCT to generate DCT coefficients dequantisierte
- A sub-sampling and an Anti-Alias | Wavefront Mayaing to change after the sizes of the DCT coefficients to
- A DCT coding, to the changed DCT coefficients of the compression algorithm of the output streams to be able to adapt.
Since it is almost the complete description of how the DCT of the DV signal must be converted so that it can rise to a DCT, which can be used for the MPEG signal can be used. By transcoding without DCT can there still be no question.
When I consider that the normal encoding of DV to an I-frame only MPEG-2 with 4:2:2-Farbsampling and 50 Mbps (on my calculator in about half the real So 1:2 - If I were a 4:2:0 I-frame only MPEG-2 with 25 Mbit / s produces, goes even faster the encoding runs), I see no benefit in time savings (because even the mere copying is) not much faster, more In qualitative terms.
Marco
Antwort von RickyMartini:
Jau! This thread describes the absence of sense and expertise on the matter!
In short, about missed!
Antwort von martinolex:
Jau! This thread describes the absence of sense and expertise on the matter!
In short, about missed! No one is forcing you to participate here.
Antwort von RickyMartini:
You should be better versed on the matter do! ;)
It is a very complicated terrain on which you prepare to enter without the knowledge of yourself!
Data compression is to be understood without proper training / knowledge is not exactly easy!
You seem to want to invent from scratch have no clue about. Because you scheiterst!
Antwort von martinolex:
@ RickyMartini
Thank you very much. You have helped me a lot with your advice.
@ Marco
When I consider that the normal encoding of DV to an I-frame only MPEG-2 with 4:2:2-Farbsampling and 50 Mbps (on my calculator in about half the real So 1:2 - If I were a 4:2:0 I-frame only MPEG-2 with 25 Mbit / s produces, goes even faster the encoding runs), I see no benefit in time savings (because even the mere copying is) not much faster, more In qualitative terms. Yes. I had only the interest, whether there exists one transcoder for IMX and how he technically constructed. I realize that my problem is not IMX for einsetztbar.
Idea of the patent, I have understood that a Transfomation is being developed that make DCT boxes of boxes of DCT as DV for example, MPEG-2 directly converts without iDCT / DCT too. Just as in the picture (from EP1155573.pdf) below:
Antwort von Marco:
The most intensive computation is given to place around the DCT transformation. The patents to're around 10 years old. Who knows whether these principles are not now in almost every MPEG encoder felt at home. That will depend on the Manufacturer but usually not s.The great bell.
It indeed must not be synonymous forceably IMX. A reasonably good encoder allows its settings synonymous to "normal" MPEG is adopted so that loss or I-frame only format with 25 Mbit / s as long as that is by default.
Marco
Antwort von martinolex:
Who knows whether these principles are not now in almost every MPEG encoder felt at home. That will depend on the Manufacturer but usually not s.The great bell. Good idea! Sure, if Cyberlink did it why not others.
A reasonably good encoder allows its settings synonymous to "normal" MPEG is adopted so that loss or I-frame only format with 25 Mbit / s as long as that is by default. Can you tell me as a good call MPEG-2 encoder. I have Pinnacle Studio 12 and OEM abbreviated version of Adobe (was here with Sanyo). Both can not be in i-frame set only. Thank you very much.
Greeting
martinolex
Antwort von RickyMartini:
The patent (from 2000) describes how a DCT-based Videobitstream (MJPEG, DV) into another DCT-based Videobitsstream (MPEG) is transcoded.
Here, the DCT is heavily used.
A signal from the DV camcorder can thus be converted to MPEG2, the video can be regarded s.PC liquid.
When Trancoding of DV to MPEG2 should be a 4:1:1 to 4:2:0 color conversion synonymous, so you do not 1:1, the Fields / Frames can take (pictures) (see p. 9).
Antwort von martinolex:
When Trancoding of DV to MPEG2 should be a 4:1:1 to 4:2:0 color conversion synonymous, so you do not 1:1, the Fields / Frames can take (pictures) (see p. 9). 4:1:1 is used only in NTSC DV. PAL DV is 4:2:0 sampled. Patent text is merely of an American LEE Jungwoo :-). As you're saying above:
the absence of sense and expertise Greeting
martinolex
Antwort von RickyMartini:
I-Frame Only brings you no quality improvement over GoP-based MPEG2!
I'm testing now with the MPEG2 encoder Of Canopus Edius.
Result: With the 8Mbit / s, cbr encoded MPEG2 - video is much better than the I-frame based 15Mbit / s, cbr encoded video.
The noise ratio is generally low profile, which significantly improves the image quality.
Antwort von RickyMartini:
When Trancoding of DV to MPEG2 should be a 4:1:1 to 4:2:0 color conversion synonymous, so you do not 1:1, the Fields / Frames can take (pictures) (see p. 9).
4:1:1 is used only in NTSC DV. PAL DV is 4:2:0 sampled. Patent text is merely of an American LEE Jungwoo :-). As you're saying above:
the absence of sense and expertise
Greeting
martinolex I have appointed me to this paper. Thus, my argument is not false. Just the different color compression (PAL / NTSC) does not make a 1:1 implementation possible.
Here DV (PAL / NTSC) is again listed below:
http://de.wikipedia.org/wiki/Digital_Video
Antwort von Marco:
If you consider the fact that none of these patents without massive recalculations DCT manages synonymous and they can be compared to today's standard MPEG encoders do not expect changes with modern time-and quality benefits, the next question is actually, for whatever reason not synonymous simple GOP-based MPEG can be taken.
This goes just as fast, equivalent in quality, without having to make any Grätschen And this still saves you a lot of space and nerve-wracking. So you'll get your entire material may be on a large 1.5 TB disk and have to bend under no established standards.
The I-frame-only rendering I had just been tested with Vegas Pro 9. This, however, several things must be considered so that the data rate of 10 Mbit / s can be set to adjust (Profiles) and other parameters still interact properly.
But - all this is not really worthwhile because it can be so much easier if one simply takes on formats that are quickly encode to offer good quality and have proven themselves for years.
Marco
Antwort von martinolex:
@ Marko
Ok. Persuaded. Thus. MPEG-2 MP @ ML with 15 (or less synonymous) Mbps.
Now I just have to batch-search tool, because it is many files.
Let's say: it's done. How can I check that in the Ergebis no nasty surprises in there are? Watch all the 340 hours?
Is there a tool that compares the videos and any deviation protokoliert numerically with time stamp? Thank you.
Greeting
martinolex
Antwort von Marco:
Tools to test, there's certain, but I know of none. As the pixel count would be reported to words;)
Marco
Antwort von Marco:
I have just synonymous times the standard MPEG encoding (12er GOPs) with 15 Mbit / s tested. Also running faster than real-time and half the quality is subjectively indistinguishable from the original.
Your approximately 350 hours of DV are thus fit on a fast PC in some 100 to 150 hours on two 1.5 TB hard disks and players are of like the WD easily read TV.
Marco
