AVR Scaling and Findings

The place for all discussion on gaming hardware
Post Reply
User avatar
Bahn Yuki
Posts: 237
Joined: Thu Oct 15, 2015 11:33 pm
Location: Salem OR
Contact:

AVR Scaling and Findings

Post by Bahn Yuki »

https://www.youtube.com/watch?v=T4VsQU2c2Fs"

Hey guys so I was trying my DENON X4300H with Dolby Vision update and I figured I should try out the scaler built inside. I don't know if there's a topic with comprehensive look at the scalers into these AVRs. I know my old Onkyo 609 was great and used the same chip that's in the Framemeister Mini. This DENON(and my X6300H in my theatre room) does pass through the OSSC in all modes successfully, unlike my Onkyo 636 upstairs.

So I hooked up my PS2 and Xbox. I was able to put a PS1 game(Tales of Phantasia) and unfortunately it processed the 240p as 480i. It looked decent but there was a bit of "fuzziness". Input delay was pretty good in "GAME" mode. Then i proceeded with the test in the video above, using the Xbox with a 480p game(CVS2). As you'll see it digitizes(Component to HDMI) the 480p straight to the LG and it looks ok. I then tried upscaling it to 4K/60hz and the difference was pretty remarkable. I'll definitely be using it in this mode from now on.

So do any of you have these recievers with decent scalers inside? I know the Onkyos recently(818, 636, 656) have been poor with not doing anything over 480p with component input. The Onkyo won't even pass the OSSC signals! Thankfully I've switched over to DENON and very surprised by the results in 4K/60. Perhaps that scaler can be used for future products?
LG 83C1(OLED),LG 42C2(OLED),TCL 75R635(MiniLED),Apple Studio Monitor 21(PCCRT),SONY 34XBR960(HDCRT)
SONY 32XBR250,Samsung UBJ590(LED),Panasonic P50VT20(Plasma),JVC NZ8
User avatar
Xer Xian
Posts: 895
Joined: Sun Feb 06, 2005 3:23 pm
Location: Italy

Re: AVR Scaling and Findings

Post by Xer Xian »

It looks great. Does the X4300H support 4:4:4 at 4K60? (it was set on 4:2:2 on your vid). Anyway, at $800 this looks like a bargain, even for the video processing only (we don't need Lumagen's advanced color management for gaming).

If you have a mic for the Dreamcast you can test the added latency of the Denon with sub-frame precision thanks to the 240p Test Suite audio lag test (you can set it to output both 480i and 480p) *.

And no, as far as I know the video processing capabilities of AVRs are not documented at all around here, as of now.

* Edit: This is assuming that the Denon automatically delays the audio to sync it with post-processed video.
User avatar
Fudoh
Posts: 13044
Joined: Mon Mar 06, 2006 3:29 am
Location: Germany
Contact:

Re: AVR Scaling and Findings

Post by Fudoh »

Sorry that I can't agree with you, but I think the Denon completely mangles the analogue signal. It adds a terrible chroma delay on the secondary color channel (which causes the heavy bleeds on all your saturated reds). This is likely caused due to a poor color down conversion (from 4:4:4 to 4:2:2) right during the A/D stage. Your TV confirms the incoming signal to be 4:2:2.

This would after be identical to what I've experienced on almost any AV receivers I've tried myself.

The problem is that it's extremely easy to really mistreat a solid source signal. This has hardly anything to do with using cheaper or ultra high end processors. With video processors you can have luck with a cheap device that was geared for use with PC signals and you can get bad results with extremely expensive processors that have been built for home cinema use (Lumagen's processors have the very same problem with conversion to 4:2:2).
User avatar
BuckoA51
Posts: 3424
Joined: Sat Oct 02, 2010 10:08 am
Location: Ireland
Contact:

Re: AVR Scaling and Findings

Post by BuckoA51 »

Speaking of the Lumagen, does it do this 4:2:2 conversion to digital sources too (like OSSC?)
OSSC Forums - http://www.videogameperfection.com/forums
Please check the Wiki before posting about Morph, OSSC, XRGB Mini or XRGB3 - http://junkerhq.net/xrgb/index.php/Main_Page
User avatar
Fudoh
Posts: 13044
Joined: Mon Mar 06, 2006 3:29 am
Location: Germany
Contact:

Re: AVR Scaling and Findings

Post by Fudoh »

yes.

BUT if your original source is 240p and you use a 720p or 960/1080p signal into the VP, it hardly matters, because you have enough color resolution left. It mostly hurts on sources running in its native resolution (e.g. a 480p signal from a PS2) because now you actually lose resolution and details get visibly mangled.

Link to my 5-year old posting on the Lumagen boards:

http://www.convergent-av.co.uk/forum/in ... #msg_19070

4:4:4 upscaling on the left, 4:2:2 on the right, but keep in mind that Lumagen is especially bad here. Other 4:2:2 processors smear the color details a little, but they don't have the chroma delay problems. Lumagen's Vision HDQ had 4:4:4 processing. The shown 4:2:2 processing was introduced with the Radiance units.
User avatar
BuckoA51
Posts: 3424
Joined: Sat Oct 02, 2010 10:08 am
Location: Ireland
Contact:

Re: AVR Scaling and Findings

Post by BuckoA51 »

BUT if your original source is 240p and you use a 720p or 960/1080p signal into the VP, it hardly matters, because you have enough color resolution left.
Not sure I understand that, can you explain it in simpler terms (or point me to an article to read or something) ?
OSSC Forums - http://www.videogameperfection.com/forums
Please check the Wiki before posting about Morph, OSSC, XRGB Mini or XRGB3 - http://junkerhq.net/xrgb/index.php/Main_Page
User avatar
Fudoh
Posts: 13044
Joined: Mon Mar 06, 2006 3:29 am
Location: Germany
Contact:

Re: AVR Scaling and Findings

Post by Fudoh »

4:2:2 means full luma resolution, horizontally halved chroma resolution on both color channels. If you take a OSSC signal with 1280x720p and have it converted to 4:2:2 you still have 640x720p on the color channels. Easily enough to keep every detail from a 240p source intact. On a good 4:2:2 processor* the difference might be about the same as the difference between generic sampling and optimized sampling through the OSSC.

If you do the same to a hi-res 480p signal during the A/D stage, you get 360x480p - in other words: you're losing 50% of the color details already prior to the upscaling.

That's what you see on your DVDO. All internal processing is 4:2:2 and I guess you would agree that it's a bit too destructive for 480p sources, but still works nicely on a OSSC's 720p signal.

(* the best one's likely the Framemeister)
Last edited by Fudoh on Sat Oct 07, 2017 9:08 am, edited 2 times in total.
User avatar
FinalBaton
Posts: 4475
Joined: Sun Mar 08, 2015 10:38 pm
Location: Québec City

Re: AVR Scaling and Findings

Post by FinalBaton »

Thanks for setting the record straight Fudoh, it is really appreciated, as always
-FM Synth & Black Metal-
User avatar
BuckoA51
Posts: 3424
Joined: Sat Oct 02, 2010 10:08 am
Location: Ireland
Contact:

Re: AVR Scaling and Findings

Post by BuckoA51 »

^ what he just said, thanks.

I dunno I used to think the Edge + XRGB3 combo was pretty bad ass (until OSSC came along of course :mrgreen: )

Is the VP50 Pro all 4:2:2 processing too?
OSSC Forums - http://www.videogameperfection.com/forums
Please check the Wiki before posting about Morph, OSSC, XRGB Mini or XRGB3 - http://junkerhq.net/xrgb/index.php/Main_Page
User avatar
Fudoh
Posts: 13044
Joined: Mon Mar 06, 2006 3:29 am
Location: Germany
Contact:

Re: AVR Scaling and Findings

Post by Fudoh »

all DVDOs are. Lumagens since the Radiance. The CII is 4:4:4. Some of the Extron/Kramer units are likely 4:4:4 as well, but I don't have an overview. Too many machines with too many different processors inside.
User avatar
Xer Xian
Posts: 895
Joined: Sun Feb 06, 2005 3:23 pm
Location: Italy

Re: AVR Scaling and Findings

Post by Xer Xian »

Fudoh can probably tell 4:4:4 from 4:2:2 processing from a quick glance (while playing Lords of Thunder :) ), but in general it's not easy to know and very few manufacturers declare how the signal is sampled. Moreover, even if a VP supports YCbCr 4:4:4 output, that doesn't guarantee that the internal processing is carried out with full chroma resolution.

I can however add a couple more items to the 4:4:4 list (they're declared as that): the CORIO2 line from TVONE and the Extron DSC HD-HD 4K (I can't even find the price tag of this one).
User avatar
orange808
Posts: 3877
Joined: Sat Aug 20, 2016 5:43 am

Re: AVR Scaling and Findings

Post by orange808 »

Xer Xian wrote:Fudoh can probably tell 4:4:4 from 4:2:2 processing from a quick glance (while playing Lords of Thunder :) ), but in general it's not easy to know and very few manufacturers declare how the signal is sampled. Moreover, even if a VP supports YCbCr 4:4:4 output, that doesn't guarantee that the internal processing is carried out with full chroma resolution.

I can however add a couple more items to the 4:4:4 list (they're declared as that): the CORIO2 line from TVONE and the Extron DSC HD-HD 4K (I can't even find the price tag of this one).
Extron made some 4:4:4 claims in the last generation of processors that were dubious. It seems they have turned over a new leaf. Their new machines are aggressively marketed as 4:4:4 and I believe them.

I'm not sure I believe TvOne about the Corio 2 using 4:4:4 internal processing. Maybe.

Some HQV processor Realta/Reon machines are marketed at 4:4:4. In my experience, they are not.

HDFury still claims their Linker 4k uses 4:4:4. It does not.

Also, watch out for reported latency/lag times. Many manufacturers do not consider building the frame buffer to be "latency". That leaves a significant gap (up to 20ms) between the advertised latency and real world performance.
We apologise for the inconvenience
User avatar
Xyga
Posts: 7181
Joined: Tue Nov 05, 2013 8:22 pm
Location: block

Re: AVR Scaling and Findings

Post by Xyga »

Roland also claim their processors do 4:4:4 10bit internal processing, 4:4:4 8bit output processing.
Strikers1945guy wrote:"Do we....eat chicken balls?!"
Post Reply