What is the truth about EDGE ENHANCEMENT? (1 Viewer)

[Mark Zimmer]

Seems to me that the key in dealing with industry types is to not refer to "enhancement" or "edge enhancement" since those seem to be concepts alien to them, but to simply ask, "Why is there all this ringing? Make it stop." That way there's no need to get into semantics of what's enhancement and what's not and whether it's intentional.

After all, it's the end result on the screen that we're upset about, not what specific technique caused it in the first place.

 

[Scott_MacD]

so the change from an uncompressed 480 video signal into a MPEG2 format compressed file happens in essentially one step.

Some commercial encoders have a variety of options for preprocessing the input in order to get a superior compression bitrate, this can involve lowpass filtering of the image, thus filtering out detail. Since there is less variation in the video, this lowers the bitrate, and hence the size of the output MPEG2 file. This process usually causes the resultant image to look soft and undetailed. The compressionist notices this on the monitor and adds a degree of sharpness to the video. (Whether this sharpness filter is applied by default is unknown, which is why I was asking about the Sony encoder, since for different ratios the compressionist seems to apply different amounts of sharpness.)

ahem. sharpness = EE.

 

[Artur Meinild]

Vince, you raise a very interesting point, and of course you support it with extensive research as always.
I don't know what is the cause of the "ringing" or "EE" seen on some DVD's, all I know is that some DVD's have it more than others, and therefore it should be possible to avoid.
It all comes down to quality control. Studios simply have to be sure that the product they press and distribute is of good quality. That is not the case with TPM. I don't care whether the EE was a result of telecine enhancement, MPEG compression or some other means, all I know is that it shouldn't be there, as it's an undesired image artifact.
I do however hope that some people in the industry will have a clue or even know for sure what causes these artifacts, else it might be time to find out now before new DVD's are pressed with more of those "mysterious and ghostly" edge artifacts...

 

[Sean Oneil]

And if it was a direct digital, that seems to support my point even more. The ide being that the endge enhancement artifact we're seeing is not the same enhancement function of the telcine process-- rather an artifact introduced later in the process

Yes, it is simply some sort of filtering that is being applied to the video files once they have been 'captured' by telecine, or after rendering down to 720x480 for Direct Digital.

All I was saying is that this "EE" is not an artifact of the MPEG encoding process.

 

[Dave H]

All I was saying is that this "EE" is not an artifact of the MPEG encoding process.

I agree. If this were the case, wouldn't we see EE on every DVD - we know this isn't the case.

 

[Vince Maskeeper]

No, in fact that wouldn't be the case. Compression is not a single function or process- the degree and types of compression applied to every single disc is different. The application of various degrees and formulas of compression are different from shot to shot on many discs!

You could see horrible artifacts on one disc, while a second was completely clean-- yet both were caused by MPEG compression. It's a matter of how these controls are used.

The argument that it can't be MPEG compression since it's not that way on every disc is like arguing that salt can't be used to ruin a meal because otherwise all meals with salt would be ruined. It's not the intrinsic use of salt that is the problem, but rather how the salt is used...

The point was that the edge filtering effect could be a specific artifact of specific settings/functions of the modern compression process. Just like the stereotypical artifactiing of MPEG2 can be avoided with careful compression settings (and again, this is not just one simple slider to determine compression, rather hundreds of intricate controls)-- I was theorizing that the increase in the edging effect could have resulted as an new artifact of poor settings in newer compression tools.

Just like improper use of traditional compression tools resulted in horrible digital artifacts... I wondered if newer compression schemes were creating artificial edge filtering as part of their control settings.

Again- I'm just hoping to get more of a dialog with the video professionals who create these masters so we can understand why these things happen in order to avoid them. I think assumptions have been made in the past, incorrectly, about the source of these artifacts... and in order to get someone to take a serious look at the problem we need to take a step back and open up the options of what the cause might be.

Again, by referring to it as "edge enhancement"- I think this is how THX can claim that none was applied to TPM. I would guess what they know as "edge enhancement" or specifically "enhancement" of the telcine process was NOT applied to TPM--- instead filtering later in the authoring process (during conversion, compression, encoding, authoring, whatever) caused what we see. But by incorrectly using the terminology we've used, we may never get to the bottom of the "why"- which means we may never really understand how it can be fixed.

 

[Scott_MacD]

Vince: If you find out any solid information about the compression step of the film Dev/Null, including Aperture Enhancement, sharpness or any other synonym for EE, please let us know. I mean, you're close to the project.
Or if someone here could perform some testing involving professional compression equipment.
On another note, did David Prior (valued member of the HTF) work on the Die Hard with a Vengeance SE disc? If so, it'd be interesting to hear his knowledge on this. No blame, no fueling fire, just I think we'd like to know what happened with the quality control on the resultant DVD.
I also understand THX were running a demonstration to "prove" that there was no filters applied to the transfer of TPM, taking us from every step. Has this happened? Or did it not. :frowning:

 

[Jeff Kleist]

The reason why you don't see EE added to your rinky dink encoder is because every single one I've seen softens and sacrifices detail rather than do edge enhancement

 

[Derek Miner]

The compressionist notices this on the monitor and adds a degree of sharpness to the video. (Whether this sharpness filter is applied by default is unknown, which is why I was asking about the Sony encoder, since for different ratios the compressionist seems to apply different amounts of sharpness.)

Perhaps I'm oversimplifying here, but all of us here could recreate similar effects by using too much compression on a JPEG file. I'm not choosing to add some sort of enhancement to compensate, but I still see the same result. Therefore, I'd speculate that a compressionist who makes no adjustments other than the default rates of compression can produce product with these symptoms.

 

[Nick_Scott]

Perhaps I'm oversimplifying here, but all of us

here could recreate similar effects by using too

much compression on a JPEG file. I'm not choosing

to add some sort of enhancement to compensate,

but I still see the same result. Therefore, I'd

speculate that a compressionist who makes no

adjustments other than the default rates of

compression can produce product with these symptoms.

__________________________________________________ ____

I don't think the problem is too much compression, but too much filtering.

I work alot with large JPEGs that I transfer over the net, that are often too big to transfer.

I can add extra compression, but compression artifacts look terrible, and are usually very obvious.

Instead, I lowpass filter (or BLUR), then sharpen (or EE). Usually this drastically cuts the size of the file with minimal artifacts.

 

[DaViD Boulet]

Vince,
This thread points to the likely problem with various MPEG encoders but this idea isn't new.
Bjoern and others have already concluded that the most likely of culprits are:
1. sometimes ringing is added "intentionally" to sharpen the image.
2. sometimes the MPEG encoder being used possibly may add ringing/EE to compensate for the filtering it performs to aid in compression (this is *not* a compression artifact, but an pre-compression artifact of some encoders).
So yes, we're all ok here not assuming that Fox *intentionally* added ringing. That's never been the point. Many many times it's been put forth that perhaps this is something implicit that is happening at the MPEG encoder level prior to compression.
Either way, the big-boys over and Fox and THX and other studios should be smart enough to figure it out and try to make their future "statement" DVDs look right. We should never, EVER have to put up with discs like Sound of Music and Episode I. On the big-screen they're downright distracting :frowning:
And why is it the job of the consumer to try to teach THX what a bad DVD image looks like vs a good DVD image? I think that's where a lot of the frustration comes in too. It's they're job not only to KNOW what a good DVD image looks like, but to indicate to me by their logo that a particular DVD on the shelf is going to deliver it. THAT'S THEIR JOB. We consumers shouldn't have to play patti-cake with them to try and warm them up using euphonic semantics or carefully crafted arguments to get them to see what's painfully obvious on any large-screen system by any casual viewer.
These studio guys should be pro-active enough not just to cop-out with a "we didn't add any" line. If they know they didn't *mean* to add any EE, but there's some sort of EE there all the same, then obviously they DID add some by accident somewhere along the line. If the MPEG encoder is at fault, they need to get a new encoder.
Don't these guys give the discs a look before they stamp a million copies? We publish a book at the company where I work. If I give the document to the publisher which looks perfect, but then our "test" copy comes back from the publisher and the color is wrong...it doesn't matter whether or not my original digital file (the telecine transfer here) is to blame...we make sure the problem is corrected before we print our run.
If I can do it, so can THX.
-dave

 

[Todd Hochard]

I am convinced that ringing artifacts are not inherent to MPEG compression (which Bjoern has capably shown has its own set of artifacts such as "mosquito noise" and the like), but I am also pretty sure that they are an artifact inherent to certain MPEG encoders.

I absolutely agree with this.

I've had a similar experience to Vince's with a friend's short film. It was done on Mini-DV (Sony TRV-900), and edited on his dual-PIII Canopus DV-RexRT setup. The finished product on Mini-DV (with its proprietary low compression MJPEG style) is completely free of artifacts (at least on my 53"). Run it through MPEG2, and BAM! there it is! EE/ringing. Bitrates, when set too low, have something to do with it, but it seems the different encoders that we can get our hands on have the bigger effect.

I think that, ultimately, it may fall back to the same kind of deal as the Chroma Bug thing. Only one (or a few) of the companies making the encoders are getting it right.

I also bet that we continue to see the same sort of ringing on DVHS, but to a lesser extent, for two reasons:

1. At 1920x1080, the ringing will be "compressed" up against the edge in fewer lines, so some displays that aren't particularly good at resolving 1920 will miss it (like my Elite 510 RPTV, for instance)

2. The bitrate is run much higher, at 28.2Mbps.

I wonder if wavelet/MPEG4 object-oriented encoding does the same sort of thing?? Haven't played with it enough to say.

Just my amateur opinion.

Todd

 

[Bjoern Roy]

I still haven't found the time to read all the posts here and post something useful, but...
Todd,

Quote:

So Steve (and basically every other compressionist and telecine operator i talked with) does have the same opinion on THX as David and I.
More later
Bjoern