Picture Quality *To The Average Person*

Eric Bilodeau · September 24th, 2003, 08:51 AM

A thing to know about edge enhancement:

First thing, it is mostly a "capture thing", it is a bad interpretation of the border between a bright area and a dark area witch results in an undesirable dark green-blue edge line. It is a direct result of an overload of the CCD because of direct light entering through the lenses directly to the CCD. Again, edge enhancement is not present when shooting through a mini35 adapter because the light does not blast directly on the CCD. Edge enhancement can be worsened by a higher contrast on the TV but not created. Chroma noise will be more visible in a high contrast monitor too.

Charles Henrich · September 24th, 2003, 12:46 PM

I believe in the case of the JVC this has nothing to do with the CCD and instead is
in-camera (software) filtering effects. Find a high-contrast transition and accentuate
it by darkening the one side and lightening the other to make it really pop. Much
like SVM. It generally sucks, but the threshold seems to be fairly okay on the HD1
from the clips I've seen (I have one clip my friend put together where its not terribly
apparent except on a sharp transition between building/sky). Still, being a purest
I was forced to shell out for the HD10. Even though im sure at somepoint someone
will figure out a way to change the incamera settings on the HD1 and make the
two identicle :)

Eric Bilodeau · September 24th, 2003, 01:32 PM

I would say that it is a combination of both but it has to do with the CCD, not only the software. I tested the mini35 with a pd150/vx2000 adapter on the JVC HD10, though it is not made for it it worked (not perfectly) but it worked. There was NO visible edge enhancement when shooting with the device on scenes witch had some without it (mosly on shiny metal pipes). By that test it was clear that it was not a software only problem because if so there would still be edge enhancement. On the other hand, the fact that the HD1 and HD10 use the same CCD means that there is a software or hardware thing going on that differenciates the two. Maybe there is a limiter device or filter on the HD10 that diminishes the contrast ratio but it has to do with light entering directly to the CCD. This edge enhancement thing has been happening since the beginning of video. I have a hard time figuring how a technician at JVC or elsewhere would prefer that silly line over the edge over a more realistic contour and decide to boost it "because it is more pleasing".

My thought on this is that they probably needed a filter (physical) to attenuate the overload of the CCD. That meant getting the price of the device up by a few hundred bucks so they decided to create two versions of the camera knowing that it might be a deal breaker for consumer not interrested in a more professional image (to most people 3499$ sounds closer to 3000$ than 4000$ but 3999$ sound like 4000$ and goes over that price with taxes). If they don't care about the edge enhancement why bother, but they knew pros would pretty much care about it and they could not make it a consumer only toy, it's too expensive. I mean why create two models with a 500$ difference at the same time? to attract two type of users? If so the differences would have been more noticeable (better handling, manual controls, etc...) and the price difference would have been greater. I would not be surprised if someone found out that there is a piece or two that is on the HD10 and not on the HD1 (internally of course). If it could have been taken care of in the consumer model (without boosting the price) there would probably not have been a prosumer model released, there would have been one model that would have pleased both the consumers and the pros.

Charles Henrich · September 24th, 2003, 01:44 PM

While I cant say for sure, I would bet it has nothing to do with hardware. They did it for the same reason that virtually every television on the planet has scanning velocity modulation heavily enabled. Because the average consumer wants to see a "sharp" image. Heavy edge transitions please the average viewer. What your doing with the filters is reducing apparent contrast, which is causing the software filter to not trigger. Edge enhancement algorithms look for sharp edges and stretch those even further. If you reduce them before it gets to the software you "solve" the problem. Of course you've softened your entire image to avoid the filter. Better to have the software features disabled in the first place :)

Eric Bilodeau · September 24th, 2003, 02:05 PM

Yes, people want to see a sharp image but again there is a difference between a sharp edge and an edge enhancement, if you sharpen the image from a DVD of a movie that was shot in 35mm for example, you will not create the edge enhancement, even when cranked to the max. Of course an image already polluted will be worsened...

Anyway, it's only my 2 cents, but I think if it was software there would not have been two models.

Andre Jesmanowicz · September 24th, 2003, 04:24 PM

The most important question. Is enhancement done on the component output or is it recorded to the tape? I have GR-HD1 and when filming white airplane with shadows from some parts at sunny day, the enhancement is visible on a computer. After editing and dumping to 40000 it is visible on a HDTV also.

Can the enhancement be changed by a slight modification using a special remote?

I would be even willing to change a single board in the camera to avoid annoying artifacts. Have experience. I replaced an optical block in my camera ($300), but after gluing the optical stabilizer lens in place. Now I don't have vibrations in freefall when skydiving. Before when I hit the sun in freefall I had a green vertical line looking like a snake. Now it is straight like suppose to be! :-) The optical stabilizer is supported by a large spring that transforms vibration from a tripod to the stabilizing lens with all negative effects on image stability even when stabilizer is off. A very poor construction that is vulnerable to gravitational and inertial forces.
Andre

Eric Bilodeau · September 24th, 2003, 06:01 PM

Unfortunately it is recorded onto tape. it is very difficult to efficiently (even impossible) to get rid of those enhancements without sacrificing the image in the process. Even admitting one could do that, it would certainly not be perfect.

John Eriksson · September 24th, 2003, 06:49 PM

I want tol learn more / is the P+S Technik going to make an jy-hd10u/e adapter for the mini35 anytime soon?

Could you please expand on this subject please:
Eric Bilodeau:
"I tested the mini35 with a pd150/vx2000 adapter on the JVC HD10, though it is not made for it it worked (not perfectly) but it worked."

..explain more how did you get it to work??

Eric Bilodeau · September 24th, 2003, 10:28 PM

The adapter of the PD150 unlike the adapter of the XL1 is a converter made to fit in front of the lenses (the one for the XL1 is made to fit in front of the CCD without lenses). The size of the cameras (PD-150, VX2000 and HD10) are close so I figured I might check out how it could work. I heard of plans from PS+technik for the JVC (unoficial sources) but the process is long (it is not even finished for the DVX100) and I encountered a slight problem in the process: As the adapter is not made for the HD10, the precise distance and focal lenght is not right so it has to be adjusted manually. Adjustment for this system is quite tricky so as the HD10 is of very high def compared to the miniDVs, the screen was visible when the iris of the 35mm prime lens was under f2 under normal light conditions, the only concluding tests where with a lot of light (outside) where the screen was not visible in most f stops. Also, the screw base (or camera stand?) of the mini35 is a little too low and to the right so an adapter had to be "homemade" to fit it. Other than that, it could work.

I did not had the pPD150 adapter for very long though, we where testing the mini35 gizmo on a XL1 for the DOP (of a production I am technical consultant on) and I had this "hunch" that a PD150 adapter might just fit with the HD10. The PAs made a few calls and we had an adapter coming to the office on the next day but as we had one day left of tests, I pretty much made it over an afternoon. So tests are preliminary. Actually, as we have access to a varicam, tests for the HD10 have been suspended so I should not push it further for the moment. The only remaining scheduled test is for 35mm blowup of some early test footage.

If anybody has access to a mini35 with a PD150 adapter and a HD10, it might be interresting to see how it ends up...

Dave Largent · September 25th, 2003, 05:50 PM

Steve, to answer your question as to what the TV contrast was set at: 75%.
Like I was saying before, I had the sharpening set at max because I had assumed that "a sharp picture is best". And like
I said, I started messing around with the sharpness --0%, 50%,
100% -- trying to dial in the best-looking picture. Seeing as I
was also playing with the sharpness to aide in the decision regarding a large purchase, I kept "experimanting" for a while,
flipping through different stations. Well, the wife came in and sat down to watch some TV, as I kept fiddling. After a while she said to me "Enough already, would you just leave it at 0%". I asked her why, and she said "Cuz it looks a lot better that way". I would have to agree with her. Damn, but I *want* to like the cheaper cam better.

Charles Henrich · September 25th, 2003, 05:56 PM

Dave, you need to go get yourself a copy of the DVD "Digital Video Essentials". It has tons of test patterns and the like for assisting in the proper set up and calibration of a TV. Its a must for anyone who does anything with video (or really its a must for everyone who watches TV, but I digress).

Alex Raskin · September 25th, 2003, 08:39 PM

According to JVC tech support, both cameras process images similarly.

This is not so in my experience (and no, this isn't a monitor issue - the edge enhancement shows throughout the whole process from tape to editing to SD downconversion on DVD. See a partial fix below).

I owned both cameras (and kept HD10). Bottom line:

HD1: notoriuos edge enhancement to the extent that the image looks unnatural, all lines around the objects - car moldings, box edges, power line wires, you name it - are so busy looking, they "dance". Workaround: in post, blur the Green channel. In AE 5.5 I found that setting 2 provides acceptable compromise between the overall picture resolution and getting rid of the ugly oversharpened edges.

HD10: no edge enhancement, period. Overall picture looks a tad softer, but very natural.

Another issue, though: signal drop-outs! Please see a new thread about it. Thank you.

Eric Bilodeau · September 25th, 2003, 10:21 PM

Good to see that you confirmed what I experienced Alex (I did not play a lot with the HD1 so it is good to have opinions of a guy who had both).

Gints Klimanis · October 8th, 2003, 03:09 PM

Here's an interesting page on edge enhancement, with 8x10 color glossy photographs with pictures, arrows and a paragraph on back of each one, describing what each one was.

http://www.videophile.info/Guide_EE/Page_01.htm

Edge enhancement seems to be a 1D or 2D high shelf filter, which boosts high frequencies above a certain frequency. Amplifying high frequency information boosts rapid changes, so edges look sharper and high frequency noise is more prominent.

Brian Mitchell Warshawsky · October 8th, 2003, 09:37 PM

Alex wrote:

>>>Workaround: in post, blur the Green channel. In AE 5.5 I found that setting 2 provides acceptable compromise between the overall picture resolution and getting rid of the ugly oversharpened edges.>>>>

Question: If you shoot in B&W mode (4:4:4:4?) or separate out the colors in post, does this negate Edge Enhancement as a concern?

If so, does this mean that for B&W, there is little difference between the HD1 and HD10u?

Brian

September 24th, 2003, 08:51 AM	#16
Eric Bilodeau Major Player Join Date: Jun 2003 Location: Chicoutimi, Canada Posts: 334	A thing to know about edge enhancement: First thing, it is mostly a "capture thing", it is a bad interpretation of the border between a bright area and a dark area witch results in an undesirable dark green-blue edge line. It is a direct result of an overload of the CCD because of direct light entering through the lenses directly to the CCD. Again, edge enhancement is not present when shooting through a mini35 adapter because the light does not blast directly on the CCD. Edge enhancement can be worsened by a higher contrast on the TV but not created. Chroma noise will be more visible in a high contrast monitor too. __________________ Eric Bilodeau video SFX,DOP ___________________ http://www.fictis.net info@fictis.net

September 24th, 2003, 01:32 PM	#18
Eric Bilodeau Major Player Join Date: Jun 2003 Location: Chicoutimi, Canada Posts: 334	I would say that it is a combination of both but it has to do with the CCD, not only the software. I tested the mini35 with a pd150/vx2000 adapter on the JVC HD10, though it is not made for it it worked (not perfectly) but it worked. There was NO visible edge enhancement when shooting with the device on scenes witch had some without it (mosly on shiny metal pipes). By that test it was clear that it was not a software only problem because if so there would still be edge enhancement. On the other hand, the fact that the HD1 and HD10 use the same CCD means that there is a software or hardware thing going on that differenciates the two. Maybe there is a limiter device or filter on the HD10 that diminishes the contrast ratio but it has to do with light entering directly to the CCD. This edge enhancement thing has been happening since the beginning of video. I have a hard time figuring how a technician at JVC or elsewhere would prefer that silly line over the edge over a more realistic contour and decide to boost it "because it is more pleasing". My thought on this is that they probably needed a filter (physical) to attenuate the overload of the CCD. That meant getting the price of the device up by a few hundred bucks so they decided to create two versions of the camera knowing that it might be a deal breaker for consumer not interrested in a more professional image (to most people 3499$ sounds closer to 3000$ than 4000$ but 3999$ sound like 4000$ and goes over that price with taxes). If they don't care about the edge enhancement why bother, but they knew pros would pretty much care about it and they could not make it a consumer only toy, it's too expensive. I mean why create two models with a 500$ difference at the same time? to attract two type of users? If so the differences would have been more noticeable (better handling, manual controls, etc...) and the price difference would have been greater. I would not be surprised if someone found out that there is a piece or two that is on the HD10 and not on the HD1 (internally of course). If it could have been taken care of in the consumer model (without boosting the price) there would probably not have been a prosumer model released, there would have been one model that would have pleased both the consumers and the pros. __________________ Eric Bilodeau video SFX,DOP ___________________ http://www.fictis.net info@fictis.net

September 24th, 2003, 02:05 PM	#20
Eric Bilodeau Major Player Join Date: Jun 2003 Location: Chicoutimi, Canada Posts: 334	Yes, people want to see a sharp image but again there is a difference between a sharp edge and an edge enhancement, if you sharpen the image from a DVD of a movie that was shot in 35mm for example, you will not create the edge enhancement, even when cranked to the max. Of course an image already polluted will be worsened... Anyway, it's only my 2 cents, but I think if it was software there would not have been two models. __________________ Eric Bilodeau video SFX,DOP ___________________ http://www.fictis.net info@fictis.net

September 24th, 2003, 04:24 PM	#21
Andre Jesmanowicz Regular Crew Join Date: Jun 2003 Location: Milwaukee WI Posts: 36	The most important question. Is enhancement done on the component output or is it recorded to the tape? I have GR-HD1 and when filming white airplane with shadows from some parts at sunny day, the enhancement is visible on a computer. After editing and dumping to 40000 it is visible on a HDTV also. Can the enhancement be changed by a slight modification using a special remote? I would be even willing to change a single board in the camera to avoid annoying artifacts. Have experience. I replaced an optical block in my camera ($300), but after gluing the optical stabilizer lens in place. Now I don't have vibrations in freefall when skydiving. Before when I hit the sun in freefall I had a green vertical line looking like a snake. Now it is straight like suppose to be! :-) The optical stabilizer is supported by a large spring that transforms vibration from a tripod to the stabilizing lens with all negative effects on image stability even when stabilizer is off. A very poor construction that is vulnerable to gravitational and inertial forces. Andre __________________ Andre

September 24th, 2003, 06:01 PM	#22
Eric Bilodeau Major Player Join Date: Jun 2003 Location: Chicoutimi, Canada Posts: 334	Unfortunately it is recorded onto tape. it is very difficult to efficiently (even impossible) to get rid of those enhancements without sacrificing the image in the process. Even admitting one could do that, it would certainly not be perfect. __________________ Eric Bilodeau video SFX,DOP ___________________ http://www.fictis.net info@fictis.net

September 24th, 2003, 12:46 PM	#17
Charles Henrich New Boot Join Date: Aug 2003 Location: LA, CA Posts: 14	I believe in the case of the JVC this has nothing to do with the CCD and instead is in-camera (software) filtering effects. Find a high-contrast transition and accentuate it by darkening the one side and lightening the other to make it really pop. Much like SVM. It generally sucks, but the threshold seems to be fairly okay on the HD1 from the clips I've seen (I have one clip my friend put together where its not terribly apparent except on a sharp transition between building/sky). Still, being a purest I was forced to shell out for the HD10. Even though im sure at somepoint someone will figure out a way to change the incamera settings on the HD1 and make the two identicle :)

September 24th, 2003, 01:44 PM	#19
Charles Henrich New Boot Join Date: Aug 2003 Location: LA, CA Posts: 14	While I cant say for sure, I would bet it has nothing to do with hardware. They did it for the same reason that virtually every television on the planet has scanning velocity modulation heavily enabled. Because the average consumer wants to see a "sharp" image. Heavy edge transitions please the average viewer. What your doing with the filters is reducing apparent contrast, which is causing the software filter to not trigger. Edge enhancement algorithms look for sharp edges and stretch those even further. If you reduce them before it gets to the software you "solve" the problem. Of course you've softened your entire image to avoid the filter. Better to have the software features disabled in the first place :)

September 24th, 2003, 06:49 PM	#23
John Eriksson Regular Crew Join Date: Aug 2003 Posts: 57	Hi! I want tol learn more / is the P+S Technik going to make an jy-hd10u/e adapter for the mini35 anytime soon? Could you please expand on this subject please: Eric Bilodeau: "I tested the mini35 with a pd150/vx2000 adapter on the JVC HD10, though it is not made for it it worked (not perfectly) but it worked." ..explain more how did you get it to work??

September 24th, 2003, 10:28 PM	#24
Eric Bilodeau Major Player Join Date: Jun 2003 Location: Chicoutimi, Canada Posts: 334	The adapter of the PD150 unlike the adapter of the XL1 is a converter made to fit in front of the lenses (the one for the XL1 is made to fit in front of the CCD without lenses). The size of the cameras (PD-150, VX2000 and HD10) are close so I figured I might check out how it could work. I heard of plans from PS+technik for the JVC (unoficial sources) but the process is long (it is not even finished for the DVX100) and I encountered a slight problem in the process: As the adapter is not made for the HD10, the precise distance and focal lenght is not right so it has to be adjusted manually. Adjustment for this system is quite tricky so as the HD10 is of very high def compared to the miniDVs, the screen was visible when the iris of the 35mm prime lens was under f2 under normal light conditions, the only concluding tests where with a lot of light (outside) where the screen was not visible in most f stops. Also, the screw base (or camera stand?) of the mini35 is a little too low and to the right so an adapter had to be "homemade" to fit it. Other than that, it could work. I did not had the pPD150 adapter for very long though, we where testing the mini35 gizmo on a XL1 for the DOP (of a production I am technical consultant on) and I had this "hunch" that a PD150 adapter might just fit with the HD10. The PAs made a few calls and we had an adapter coming to the office on the next day but as we had one day left of tests, I pretty much made it over an afternoon. So tests are preliminary. Actually, as we have access to a varicam, tests for the HD10 have been suspended so I should not push it further for the moment. The only remaining scheduled test is for 35mm blowup of some early test footage. If anybody has access to a mini35 with a PD150 adapter and a HD10, it might be interresting to see how it ends up... __________________ Eric Bilodeau video SFX,DOP ___________________ http://www.fictis.net info@fictis.net

September 25th, 2003, 05:50 PM	#25
Dave Largent Trustee Join Date: Aug 2003 Posts: 1,483	Steve, to answer your question as to what the TV contrast was set at: 75%. Like I was saying before, I had the sharpening set at max because I had assumed that "a sharp picture is best". And like I said, I started messing around with the sharpness --0%, 50%, 100% -- trying to dial in the best-looking picture. Seeing as I was also playing with the sharpness to aide in the decision regarding a large purchase, I kept "experimanting" for a while, flipping through different stations. Well, the wife came in and sat down to watch some TV, as I kept fiddling. After a while she said to me "Enough already, would you just leave it at 0%". I asked her why, and she said "Cuz it looks a lot better that way". I would have to agree with her. Damn, but I want to like the cheaper cam better.

September 25th, 2003, 05:56 PM	#26
Charles Henrich New Boot Join Date: Aug 2003 Location: LA, CA Posts: 14	Dave, you need to go get yourself a copy of the DVD "Digital Video Essentials". It has tons of test patterns and the like for assisting in the proper set up and calibration of a TV. Its a must for anyone who does anything with video (or really its a must for everyone who watches TV, but I digress).

September 25th, 2003, 08:39 PM	#27
Alex Raskin Trustee Join Date: Sep 2003 Posts: 1,435	GR-HD1 vs JY-HD10: edge enhancement According to JVC tech support, both cameras process images similarly. This is not so in my experience (and no, this isn't a monitor issue - the edge enhancement shows throughout the whole process from tape to editing to SD downconversion on DVD. See a partial fix below). I owned both cameras (and kept HD10). Bottom line: HD1: notoriuos edge enhancement to the extent that the image looks unnatural, all lines around the objects - car moldings, box edges, power line wires, you name it - are so busy looking, they "dance". Workaround: in post, blur the Green channel. In AE 5.5 I found that setting 2 provides acceptable compromise between the overall picture resolution and getting rid of the ugly oversharpened edges. HD10: no edge enhancement, period. Overall picture looks a tad softer, but very natural. Another issue, though: signal drop-outs! Please see a new thread about it. Thank you.

September 25th, 2003, 10:21 PM	#28
Eric Bilodeau Major Player Join Date: Jun 2003 Location: Chicoutimi, Canada Posts: 334	Good to see that you confirmed what I experienced Alex (I did not play a lot with the HD1 so it is good to have opinions of a guy who had both). __________________ Eric Bilodeau video SFX,DOP ___________________ http://www.fictis.net info@fictis.net

October 8th, 2003, 03:09 PM	#29
Gints Klimanis Inner Circle Join Date: Jun 2003 Location: San Jose, CA Posts: 2,222	Here's an interesting page on edge enhancement, with 8x10 color glossy photographs with pictures, arrows and a paragraph on back of each one, describing what each one was. http://www.videophile.info/Guide_EE/Page_01.htm Edge enhancement seems to be a 1D or 2D high shelf filter, which boosts high frequencies above a certain frequency. Amplifying high frequency information boosts rapid changes, so edges look sharper and high frequency noise is more prominent.

October 8th, 2003, 09:37 PM	#30
Brian Mitchell Warshawsky Major Player Join Date: Aug 2003 Location: Bay Area Posts: 216	Alex wrote: >>>Workaround: in post, blur the Green channel. In AE 5.5 I found that setting 2 provides acceptable compromise between the overall picture resolution and getting rid of the ugly oversharpened edges.>>>> Question: If you shoot in B&W mode (4:4:4:4?) or separate out the colors in post, does this negate Edge Enhancement as a concern? If so, does this mean that for B&W, there is little difference between the HD1 and HD10u? Brian