Audio Authority 9A65 available. Anybody try it yet?

[Darren Wadsworth]

I saw on Digital Connection's site that they have it available now. Has anyone checked out it's capabilities yet?

Darren

[DAP]

I ordered one a couple of days ago, and it arrived today. So far, I'm not a happy camper.

I'm connecting a Samsung SIR-T165 to a Sony FW900 monitor.

Yes, I know the tuner has an RGB output, but due to the DRM stupidity it won't output its YPrPb or Svideo input on the RGB port, and the YPrPb output will only pass those inputs in 720p or 1080i.

The problem I'm having is that when I select 720p or 1080i outputs from the tuner, the picture is extreemly green. 480p outputs the correct color, but then the samsung tuner won't pass the YPrPb or Svideo inputs.

I'm not yet sure if this is a problem with the 9A65, or the tuner. Hard to test without a YPrPb display or another YPrPb HDTV source.

If this turns out to be a DRM caused problem, and it is not possible to get the YPrPb input to the RGB monitor, both products go back, and perhaps the manufacturers will learn they need to fight these idiotic laws.

The whole point of this exercise was to display playstation II games on my computer monitor. (I don't own a TV)

[DAP]

Ok, I have some new data. I went to Fry's and bought a component VGA cable and some BNC to RCA adaptors. I attached the output of my Hipix card to the component inputs of the 9A65, and set my Hipix to output YPrPb.
Everything worked. So there is nothing wrong with my 9A65.

The SIR-T165 is outputing garbage though, and I still have to track that down.

[Graham Johnson]

The colour space settings for 1080i and 480i are different which is maybe why the colours are off.

[Nathan_R]

While I'm new to both transcoders and CRT projectors, I've had my 9a65 for a couple weeks and I'm *mostly* happy. I'm running a Dish 811 into an Electrohome ECP-4100 @ 720p via the AA unit. I'm seeing an intense red push, however I don't know yet if this is due to the Dish unit, the transcoder, or the ECP. :(

[Darren Wadsworth]

Thanks for the reviews. Nathan pls Let me know when you have figured out what is causing the red push.

Did anyone else get one of these?

Darren

[Nathan_R]

Sorry, meant to update my last post last night after I did some playing.

In my case, it was a matter of a bad cable. I replaced the component cable between my Dish 811 and the AA unit, and the red push problem went away. Occasionally, I'm guilty of using el-cheapo factory-boxed cables. This was one of those times.

With that said, the transcoder works well. I wish they had made it easier to toggle the dipswitches, but considering you only need to set them once, this isn't a big deal.

[DAP]

I did some more research this weekend into the problem.

I set both my Hipix card and the Samsung tuner to the same output resolution: 1080i, and tuned to a local PBS station that broadcasts a static immage durring the day.

I found the following differences between the signal:

The Hipix is DC coupled, and the Samsung tuner is AC coupled

The Hipix only outputs synk on the Y channel, the Samsung tuner outputs sync on all three channels, Y, Pr, and Pb

Otherwise the signals are identical.

In the process of my experements, I passed both signals through my Denon AVR-3801 to use it as a quick way of swiching between the signals. A side effect of doing this was that both signals were then AC coupled. This had no effect on either source. The Samsung was still green, and the Hipix was still good.

So the cause of the problem is that the Samsung outputs sync on the Pr and Pb channels.

I looked at the output of the RGB port to see what was going on. The 9A65 is doing a DC restoration since the RGB output is DC coupled. When the Samsung was connected, the R and B signals were slightly DC shifted down, and the G signal slightly shifted up relative to the Hipix device.

So the problem is the 9A65 in that the sync on Pr and Pb disturbs the color balance of the picture.

To verify this, I disconnected the Pr and Pb signals. With this done I got a perfectly good black and white picture from both sources.

[courier72]

I would have thought the progressive component output of the two tuners should have been the same. Have you contacted Audio Authority?

[courier72]

I went ahead and forwarded the thread to Audio Authority. I thinking about picking up a 9A65 myself.

[Darren Wadsworth]

Keep the reviews coming!

Man, do I want to watch football on the PJ! Just don't have the bucks for the "high end" transcoders. Fingers are crossed.

Darren

[DAP]

I got annoyed with this problem enough that I voided my warantee.

I opened the box and started probing around.

The problem is with the DC restoration circuit. There is a FET switch that is used to short the ouput of an AC coupling cap to ground at a time in the signal when all three signals Y, Pr & Pb are supposed to be at ground. The switch is turned on about 300 ns too early, so about half of the time this switch is on, the sync pulse is high, and the switch has enough internal resistance that the circuit does not quite get the DC level back to ground by the time it is turned off.
The result is a slight DC offset to the signal.
Usually there is only a sync pulse on the Y signal. This is the brightness, and it makes the picture slightly darker than it should be, but this is not a problem because most monitors have brighness & contrast adjustments.
When there is a sync pulse on the Pr and Pb signals however, this DC offset results in a color shift.

I fixed the problem with a 74HC123 dual oneshot chip.

The only issue I have with it now is that it generates the sync blanking signal using a PIC processor who's timing is slightly off, it sometimes turns on the blanking just slightly before the end of the video scan line. The result is a slight jagedness to the right side of my underscanned picture. If you have any overscan at all, you will not see the problem.

[courier72]

I received back the following email from Jon Gertz of AUdio Authority today:

Dear Mr. Hamiter,

Thank you for sharing the AV Sciences thread. We always watch AV Sciences
for useful information about ours and others' products.

It appears that the gentleman in San Jose is correct about the timing of the
black clamp. We are working on a fix, but I cannot say how long it will take
for that fix to hit production. It should be noted that only a few sources
(wrongly) put sync on Pb and Pr. With most other sources the 9A65 works
well.

Thanks again for your interest in our products. I suggest you check our
website periodically for updates on Model 9A65.

Best Regards,

Jon Gertz
Chief Engineer

[VideoGrabber]

courier72, thanks for sharing that information. I'm also happy to hear that Audio Authority is working on a fix.

Unfortunately, Mr. Gertz is incorrect when he states:
> It should be noted that only a few sources (wrongly) put sync on Pb and Pr. <

Sync on Pb and Pr is neither incorrect, nor infrequently present. I refer him to SMPTE 240, as described by Extron (emphasis added by me):
> Tri-level sync was introduced with the SMPTE 240 analog HDTV standard. ... HDTV component video has sync present on all three channels: Y, Pb, and Pr. <

Full discussion of the technical details is available here (http://www.extron.com/technology/archive.asp?id=ta062002).

I'm quite impressed that DAP was able to locate and correct this problem so quickly! Perhaps AA will also look into adjusting the timing of the sync blanking, as discovered by DAP, to correct this issue as well.

- Tim

[mp20748]

Originally posted by VideoGrabber


Unfortunately, Mr. Gertz is incorrect when he states:
> It should be noted that only a few sources (wrongly) put sync on Pb and Pr. <

Sync on Pb and Pr is neither incorrect, nor infrequently present. I refer him to SMPTE 240, as described by Extron (emphasis added by me):
> Tri-level sync was introduced with the SMPTE 240 analog HDTV standard. ... HDTV component video has sync present on all three channels: Y, Pb, and Pr. <



I agree. However there are "HDTV" sources that do not have sync on Pr and Pb. The standard for SMPTE 240 is tri-level component sync. But it's not being honored by all manufacturers. For instance, the JVC 3000 only has sync on Y, yet it's a fully compliant 240 signal (with the exception of not having sync on Pr and Pb). The Momitsu 880 produces sync on all three, when in the 1080I mode. but one of the sync's is out of phase with the other two. So it seems that the standard is not clear, or being respected by all manufacturers.

The difficult one to transcode is the TRUE 240 sync (tri-level), which is the one that causes the color shift problems if the process is not timed/clamped properly before the matrix. HDTV signals with sync on 'Y' only is cake (easy).

If the clamping is not accurate before the matrix circuit, the colors could shift, therefore the clamping circuit is very critical for proper color matrix performance (accurate colors). If the color matrix is not seeing properly clamped signals, the matrix circuit 'could' see the signal as being out of phase, and there is when the colors could be slightly off (inaccurate colors but seems right) or could be to the extreme (red or green hue). Also, 1080I is much easier to transcode than 720P HDTV.

DAP is on the right track. In order to get the circuit to perform better, it'll need more ciruitry. And a better performing clamp circuit, is the key to a good transcoder. And when that happens, it should handle most of the sources out there, because it should also be able to correct for the minor flaws of the source device.

[DAP]

I did this without doing any research on the video standards. I was deathly affraid that I might find out that this was some kind of DRM, and that it would be illeagal for me to fix the problem. In my ignorance, I could fix the problem, then it would be illeagal for someone else to tell me that this was a form of DRM since they would be telling me how to defeat it just by identifying it :D . Someone has got to educate the idiots in our government.

Anyway, knowing that this is not a form of DRM I can tell you what I did.
I used the signal out of the REALLY SLOW comparator that detects the sync pulse (seriously, I did not know that comparators with a 300 ns propogation delay existed!) The rising edge of this output triggers the first one shot. This oneshot, 74HCT123D, has its timing set by a 10 k Ohm resistor, and a 100 pF cap. The rising edge of the negative output of this oneshot triggers the second oneshot. The second oneshot (there are two in the chip) has its timing set by a 20 k Ohm resistor and 100 pF. The positive output of this oneshot drives the input of the or gate that drives the FET switches.

Anyway this was the end of about a week and a half of being pissed at buying a Samsung SIR-T165 that I could not use. I bought it primarily to display my playstation games on my computer monitor. Its advertizing copy claims that it accepts 480i and 480p on its YPrPb inputs and will output to RGB. What the idiots DON'T tell you is that when any YPrPb or SVideo input signal is selected, the RGB output is TURNED OFF! Also, it WILL NOT PASS THROUGH YPrPb signals to the YPrPb outputs. If any YPrPb input is selected, it will only output them in 720p or 1080i on the YPrPb outputs, even though the device will do 480i and 480p. Or at least it CLAIMS to output 480i, whenever I select 480i, I get 480p out of the box.

Anyway, after spending $165 for the 9A65, another $100 in cables to make debugging it easier, and about 20 hours on a weekend ON TOP of the $400 I spent on the tuner, I finally have it working!

I'm Still pissed at Samsung for wasting my time & money.

On top of that their tuner hangs on KQED 9.3 after 8PM when they stop broadcasting on that subchannel. I can only recover by disconnecting the antenna, and changing the channel. Then they have the nerve to blame the broadcaster. The broadcast may be broken, but I don't care WHAT is bradcast, the tuner should NOT HANG! I don't mind if it does not display a broken channel, but I should still be able to change the channel.

Did I mention I'm STILL pissed at Samsung?

[VideoGrabber]

mp20748 commented:
> If the clamping is not accurate before the matrix circuit, the colors could shift, therefore the clamping circuit is very critical for proper color matrix performance (accurate colors). If the color matrix is not seeing properly clamped signals, the matrix circuit 'could' see the signal as being out of phase, and there is when the colors could be slightly off (inaccurate colors but seems right) or could be to the extreme (red or green hue). <

Right on the money, Mike. Even when a transcoder isn't broken and producing an obviously wrong color balance (red or green push), that doesn't mean there still aren't more subtle shifts in tint (phase) due to circuit limitations. This is just one area (along with bandwidth and residual noise levels) that separates the men from the boys in the pro vs. consumer-quality transcoder arena.

DAP,
thanks a lot for sharing your trials and tribulations (and solutions). Excellent detective work.

> What the idiots DON'T tell you is that when any YPrPb or SVideo input signal is selected, the RGB output is TURNED OFF! <

I'm getting more than a little tired myself of all the limitations that are being built into the consumer-electronics products, especially when those limitations are not disclosed ("but we never specifically claimed it could do that, though it would be a logical assumption to make"), and make no sense in the first place. (Why is outputting HiDef in the RGB color-space any less secure than in YPbPr?) They're taking good products and crippling them.

- Tim

[mp20748]

Originally posted by VideoGrabber


Right on the money, Mike. Even when a transcoder isn't broken and producing an obviously wrong color balance (red or green push), that doesn't mean there still aren't more subtle shifts in tint (phase) due to circuit limitations. This is just one area (along with bandwidth and residual noise levels) that separates the men from the boys in the pro vs. consumer-quality transcoder arena.

- Tim

Absolutely, and another big oversite is correct or good gamma (this is a hard one). As you know, this is when the image is too dull (not bright enough) or lacking in low end detail. Bandwidth is usually by the book, therefore there's usually not enough video headroom to make for a better image.

Sometime back, Jim Vanderwalker listed the lowest bandwidth for an effective circuit. I remembered reading it and thought that it was too high a figure. Later as I screwed around myself, I found out why he had listed such high bandwidth figure for HDTV. I found that you need high bandwidth to get good gamma and colors, as well to preserve the signals integrity during the transcoding process (it also has to do well below black). He also mentioned the importance of proper phase, He was 200% right on the money with these.

The other nice feature to have is sync to video alignment (horizontal position pot). Here is when you're able to get the image brightness and colors at it's best (one good feature on the KDS unit).

[Chris Satterlee]

I found this thread back in October when I received my 9A65. DAP described exactly the problem I had with my SIR-T150. I sent an e-mail to AA and they replied that they were aware of the problem and working on a fix. Not having heard from them for over two months, I e-mailed them again last week, and got the following response:

-------------------------------------------------------------
Hello Chris,

I apologize for not getting back to you we just decided in the last couple of weeks what we are going to do. Basically we do not have a fix for the 9A65 at some point we might come out with a different unit but at this time we don't feel like the changes we have made on the 9A65 will be enough to work with all the different applications that are out there. The unit seems to work fine at 480I/480P but at the higher resolutions it is hit or miss. With that being said we are prepared to buy back your unit for the price you paid minus shipping. All I would need from you is a copy of your invoice either faxed or emailed to me with your complete shipping info and then I will issue a CR# that you would write on the box and that # brings the unit to my attention.

Thank You

Curtis Lloyd
Customer Svc. Manager
800/322-8346 ext. 160
support@audioauthority.com

---------------------------------------------------------------

So now I have to decide what to do. The reason that I bought this was that the price was right for me. I have a JVC G11 D-ILA digital projector (I know that this is the CRT forum). It has both component and RGBHV inputs. I currently run the SIR-T150 to the component inputs, but it is well-known that the D-ILA projectors are bandwidth-limited on the component inputs. In other words, this is basically a tweak for me. HD looks very good already, but it was worth ~$165 to see how much better it is through the RGBHV inputs. I can't justify spending a lot for a premium transcoder.

DAP - are you still following this thread? If so, I think I'm willing to risk my $165 by trying your workaround if you are willing to share it. A few close-up photos of the rework would be great. Thanks!

- Chris