Denon DVD-5900 Contrast Bug

[gene]

<font color='#000000'>After our initial first look of the Denon DVD-5900, it quickly became apparent to us, that this player was plagued with a Contrast Bug. &nbsp;We patiently waited for a second player from Denon to verify our objective analysis and visual observations prior to reporting this to our readership. &nbsp;As a result of testing the first and second players, we are now confident that the DVD-5900 has a Contrast Bug. &nbsp; We are hopeful that Denon can resolve this with an easy firmware fix just like they have resolved other operational issues with previous players. &nbsp;We will report their feedback to this issue when we meet their engineers at CES 2004.

Denon DVD-5900 Contrast Bug</font>

 

[Guest]

<font color='#000000'>Gene

Very interesting. My compliment to Audioholics; one of the few places where you consistently find informed, critical, but fair pieces on equipment and how to set it up. &nbsp;I am awaiting Denon's reply.

Keep up the good work.

Christian (from Norway)

</font>

 

[goodman]

<font color='#CCCCCC'>Thanks for the heads-up, Gene.  I just cancelled my 5900 order with Elegant Audio.  First the inactive DVI output delayed shipment, and now it's the contrast bug, but I'd rather wait until they get it right.</font>

 

[Cain]

<font color='#000000'>I know that many people who have tried the 5900 have complained of macroblocking on black and very dark scenes.

I'm guessing that this &quot;contrast bug&quot; has something to do with the macroblocking issues.

Thank you for the review, please keep us posted on any word from Denon.

-- &nbsp;Cain</font>

 

[Rob Babcock]

<font color='#8D38C9'>I guess I'm in the minority, but if I could afford a 5900, I'd take one in a heartbeat. &nbsp;I'd use it for music and use my DVD-2200 for movies. &nbsp;If you can believe the initial reports, it's among the best sounding hi rez universals at any price.

Of course, my DLP FPTV can't really do black anyway, so I probably wouldn't know the diff!

</font>

 

[gene]

<font color='#000000'>Hi guys thanks for the feedback. &nbsp;In actuality, setting the player to 0IRE resolves much of the problem we have noted in the article. &nbsp;However we still feel the DVD-2900 offers a more dynamic picture at least via component video outs with our test displays. &nbsp;We will be gearing up soon to test DVI output as well. &nbsp;Rob you are correct, the DVD-5900 has stunning audio playback in all modes of operation. &nbsp;However configuring between Dlink and playing multi channel SACD can become somewhat confusing and annoying. &nbsp;I will detail this in my review.</font>

 

[Guest]

<font color='#000000'>Gene,
&nbsp; &nbsp; Not to nitpick, but the graph that you show is not a luminance graph, but the RGB report from Colorfacts. &nbsp;What was your measured luminance from the 5900? &nbsp;Also, your display is not calibrated to D65K with your blue being at 140% (probably closer to D85K). &nbsp;Finally, Red Push is a Color Decoder issue, not a grayscale issue. &nbsp;A properly calibrated display should have equal parts of blue, red and green at all measured IRE's from 20-100 IRE. &nbsp;Some displays do this better than others. &nbsp;In my case, I have a Hitachi 57SWX20B. &nbsp;There is a blue hump in the middle IRE's of about 10% which causes grayscale to jump from D65 at 20IRE to elevate to D72K in the middle IRE's. &nbsp;The only way to correct this is with an external scaler that has gamma control on it. &nbsp;I tried using the 5900 gamma controls, but to no avail (through DVI and component). &nbsp;

One last thing, the DVI port on the 5900 elevates the blue by an additional 10% in the middle IRE's (at least with my hitachi). &nbsp;I don't have another DVI TV to see if this is inherent to my TV or with the DVI port in general.

Dave</font>

 

[Guest]

<font color='#000000'>Gene,
&nbsp; &nbsp; One last thing. &nbsp;Did you use AVIA to do your grayscale setting or VE? &nbsp;If you used AVIA, then your grayscale is really messed up because there is chroma added to the grayscale windows in AVIA. &nbsp;AVIA is only good for color decoding, not grayscale.

Dave</font>

 

[steve]

<font color='#000000'>Dave,

We did the test with the blue set down lower on another display.  Let me point this out that we have discussed this subject with many ISF and non-ISF technicians.  The blue is typically pushed to a higher level to compensate for our eye's difficiency in seeing this color.  When we turned up and down the blue driver (BDRV) and blue cut-off (BCUT) on the display, the only thing it did was add or subtract a blue hue, it did not alter the overall brightness of the black 0-IRE window.

As for your comment regarding the AVIA DVD, there is indeed rumor that this disk is not accurate for grayscale.  We discussed this subject with both Milori and Ovation Software, the creator of this DVD.  Both parties indicated that this is indeed a rumor.  They did state that there is a slight problem with the grayscale, but not enough to be measured or observed.

Another thing I'd like to point out is that our own observations showed that the contrast improved with the DVD-2900 on two different monitors.  This was in instant observations.  Keeping the calibrations the same for the TV, and changing the DVD player only, should provide a fair comparison of the black window, somewhat regardless of the settings.  We altered many settings before concluding that the contrast ratio of the 5900 is less than that of other players and the DVD-2900.</font>

 

[Guest]

<font color='#000000'>Gary Kuo has directly stated that AVIA is flawed for Grayscale. &nbsp;This is not a rumor. &nbsp;Also, every DVD player is different in its output. &nbsp;When you change the DVD player, brightness, contrast and grayscale needs to be redone to ensure proper calibration. &nbsp;

As far as the blue being at 140%, it is dead wrong. &nbsp;Your display is not calibrated to D65, plain and simple. &nbsp;Also, can you please post the luminance graph so I can see what it looks like. &nbsp;The only true way to test for contrast is with an O-scope, not through a display. &nbsp;There are too many other things going on that can affect this.

Dave</font>

 

[steve]

<font color='#000000'>What is D65 and where can I get a copy of it?  As for your comment on blue, it's not dead wrong.  It's true.  It's done on MANY RPTV's for the very reason stated.  Also, ISF supports this method.

The only way I'm aware of to change the blue is by using the BCUT (cut-off) and BDRV (Driver) in the service menus.  When doing so, it alters the blue by bringing it down some, but it does not become flat at 100%.  The only other way to do this is to open the TV and manually adjust the voltage to the blue CRT, and I'm not interested in doing this as it seems inherently problematic and difficult.  I've checked and measured two different Sony displays and some Hitachi displays (RPTV's) and they all have a push with blue.  In fact, the new Sony model used in this test does not have the red push problem that previously plagued RPTV's, but it still pushes the blue to 140%.  I agree that anything over this bleads through enough where you can see the blue push in the IRE windows, but at 140%, it's not noticable to the eye.  I've gone through it with the AVIA disk and a Sencore Video Pro image generator.  The Sencore image generator compared to the AVIA DVD showed very similar results regarding the grayscale measurements.

In the very near future, we will turn the blue setting down on the RPTV and re-measure the contrast ratio's.  We will also do our best to improve the calibration between the DVD-5900 and the RPTV to verify that our results are accurate.  Should we find anything different, we will update our article and make note of everyone's comments and help.  At this point, I am still certain that the player does not have the contrast of the DVD-2900 and therefore does not produce as good a picture.

As for my permenant setting for this and other TV's, I will continue to push the blue to compensate for my eye.  I've done this on enough monitors and have talked with enough people to recognize the validity of this setting.  I've also seen that over 140% does indeed show through in the grayscale, but at or below 140% does not.  Instead, it improves the perceivable color of the RPTV.  

I've also noted in articles that not only is blue pushed this high in MANY RPTV's and other TV's, but it's also partially out of focus on most RPTV's.  The only way to really change this is by lens flopping and adjusting the focus from within the TV, which I have no intention on getting into.</font>

 

[Guest]

<font color='#000000'>D65 is 6500 Kelvin which is what a TV is supposed to be calibrated to. &nbsp;You are correct in that blue is defocused, but this has nothing to do with it being at 140%. &nbsp;If all are at 100%, then you adjust the Color DECODER to make sure that red, blue and green are correct for the color of the TV. &nbsp;Adjusting the color decoder is different from grayscale (apples and oranges). &nbsp;With Avia, you can adjust the color decoder without any test equipment. &nbsp;To adjust grayscale, you need to have test equipment and an optical comparitor to get it correct. &nbsp;If you look at the Avia 10IRE window in a dark room, you can see the red coming through on the picture if your display is calibrated properly. &nbsp;

All I am saying is that you should try and adjust the blue down to 100% to get your grayscale at 6500 Kelvin. &nbsp;With the blue being up so high, it can effect brightness and contrast of the TV (although it is the same for both the 5900 and 2900). &nbsp;I have used both DVD players and preferred the 5900 and 2900 in side to side comparisons. &nbsp;I had the both calibrated on different inputs on my TV to as close to D65K as possible on my Hitachi. &nbsp;The picture from both were excellent, but the sharpness of the 5900 was much better. &nbsp;That is why I kept it and sold my 2900.

Have a nice day.

Dave</font>

 

[steve]

<font color='#000000'>At 100%, two of the CRT's are running close to 6500 kelvins. &nbsp;The blue CRT is not. &nbsp;Again, as on most TV's I've encountered, the blue is slightly out of focus, and running at 140%. &nbsp;When we re-do the test next week, we'll turn down the blue on the CRT, but I don't think it will change my comparative results.

Thanks for your input.</font>

 

[Guest]

<font color='#000000'>Steve,
&nbsp; &nbsp; Thanks for the response. &nbsp;I don't think there is a problem with the 5900 though. &nbsp;As long as it displays 0 IRE and 100 IRE properly is all that matters. &nbsp;The 2900 could be displaying the 100 IRE window a little &quot;hot&quot; (say 106 IRE) which would make the 5900 seem like it has a bug, when in reality it doesn't. &nbsp;To conform to spec, it needs to have 0 IRE (below black) and 100 IRE (white). &nbsp; By using the VE test, it is passing below black information. &nbsp;The only way to see if it is passing 100 IRE properly is with an O-scope or a wavemeter (or so I have read).

Dave

PS. &nbsp;Can you post a picture of the luminance scale from Colorfacts. &nbsp;I am curious as to what the luminance is across the IRE field from your calibration.</font>

 

[steve]

<font color='#000000'>Greetings everyone,

The following response was taken from another forum discussion, and included here for informational purposes.

Once again, we thank you all for your comments and concerns about our article. Since this weekend, I have contacted Milori and had a fairly lengthy discussion with them. While the key person is traveling, I was able to catch him on his cell phone. From his initial comments, I have the following information to report.

1) Validity of Test - First of all, Mark was intrigued by our testing procedure and indicated that if the only variable is the DVD player, then this seems to be a valid method of comparison. As we indicated in our review, this is not an absolute measurement, but comparative. He seemed to agree that this is indeed valid.

2) Trichromat Sensor - Mark told me that it is very true that this sensor (as with many others including the Sencore) does not measure color within 2% at or below 10-IRE. I asked him the following. If the sensor can not measure color within 2% at 10-IRE, then how can it measure black and white and provide an accurate contrast ratio? Mark indicated that this was a great question, and offered the following response. The Trichromat Sensor is made up of four sensors potted into a housing. Three sensors are for color and the fourth gell mounted sensor is for measuring light (luminous) intensity. While the sensor does go &quot;color blind&quot; below 10-IRE and can not measure color below this, it is still able to measure light power and therefore, fully capable of providing contrast ratios, as indicated in their ColorFacts software.

3) Mark is still coming back from traveling and on a busy schedule, but he did assure me that he will review our article and our test procedure. From what I've described to him, he still seems to feel that while we are not providing absolute measurements, we are providing a &quot;real world&quot; application and showing a valid comparison.

Lastly, I did spend some time updating the article based on everyones feedback herein. For starters, I toned down the accusation of the &quot;Contrast Bug&quot; until we really know for sure. The other thing I added was the fact that our measurements were based on the DVD player being on the 7.5-IRE setting. To be honest with you all, the manual for this player stinks, and we honestly didn't know the player had this setting. We were aware of the chroma settings, and brightness and contrast settings, and did our best to adjust them, but we didn't know it had another entire setting for clipping the player at 7.5-IRE. I have mailed the sensor to Gene and he will be doing the test over at the 0-IRE setting later this week. But from what we've seen, it seems to not only tone down the black, but the white as well when switching this setting.

Anyway, we really do appreciate everyone's professional, proactive comments.</font>

 

[Guest]

<font color='#000000'>Steve,
&nbsp; &nbsp; If set to 0 IRE, then set the brightness to AVIA. &nbsp;Next, set contrast per Colorfacts to achieve a luminance curve of 2.2 (gamma graph). &nbsp;This should yeild better results. &nbsp;The 7.5 IRE setting on the 5900 does make the picture washed out and flat. &nbsp;I have noticed this as well. &nbsp;The 0IRE setting is so much better for overall picture.

Thanks for the hard work.

Dave</font>

 

[Guest]

<font color='#000000'>Thanks for the report of the chroma bug. Now I notice that Denon is reporting &quot;chroma bug free&quot; in their description of the 5900.</font>

 

[Guest]

<font color='#000000'>THERE IS NO CHROMA BUG ON THE 5900!!!! &nbsp;Where did you read THAT?

</font>

 

[pam]

<table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>

Guest : <font color='#000000'>THERE IS NO CHROMA BUG ON THE 5900!!!!  Where did you read THAT?

</font>

<font color='#728FCE'>These guys seems to agree with you:
http://www.hometheaterhifi.com/cgi-bin....VD-5900</font>

 

[Guest]

<font color='#000000'>I had macro blocking in at least 6-8 movies in dark scenes even whith the ire set to 0.
I also had transport grinding/lockup of the player that many have reported at avs forum.

This player is not anywhere near the best universals in sound quality as someone posted either.
The people who think the 5900 sounds great are correct....but not compared to a 4k universal or 1500.00 cd only player or a 2k sa-cd/cd player or even a 500.00 modded player for that matter.

This player has a LOT of bugs and it is nice to see someone pointing some of them out as other testing that was done whithout any actual veiwing involved....is pointless imo.

I am also sick and tired of seeing others try to justify the 5900 video bugs as just a case of the player being more resolving

</font>