So, %3Ca%20href="http://www.bbc.co.uk/bbchd/">the channel is just over %3Ca%20href="http://www.bbc.co.uk/blogs/bbcinternet/2008/12/happy_birthday_bbc_hd.html">a year old - and what a year it's been! After looking back at some of the %3Ca%20href="http://www.bbc.co.uk/blogs/bbcinternet/hdtv/">blogs and posts I can see how much we've done and worry about how much there's still to do.
Many of you have been asking for a test signal to help line up your own HD TVs, we have been listening but it's taken a while to get it sorted.
From this week the HD promo has two test signals and I want to talk about how to find them and how to use them to line up and check your home systems. I also wanted to share a fascinating mathematical proof that some people (%3Ca%20href="http://www.bbc.co.uk/drama/heroes/characters_cast/hiro.shtml">Heroes style) can change the flow of time!
As many of you have noticed BBC test card has been going out for a couple of weeks, this has now been joined by an Audio/Video Sync test signal. The test card seems to have been given the name "Test Card X" but not by us, it is in fact a modified high definition version of test card W (named because it was widescreen!) and for those interested in the history of test cards, there is an %3Ca%20href="http://en.wikipedia.org/wiki/List_of_BBC_test_cards">interesting "romp" through it here - it even includes the current incarnation!
The HD version uses the very famous picture of Carole (George Hersee's daughter) %3Ca%20href="http://www.barney-wol.net/video/testcardw/testcardw.html">re-scanned in high definition and added to an HD version of Richard Russell's well known widescreen test card.
Now for the purists there's a bit of a disappointment coming. No, not the fact the test card's only there for 90 seconds every two hours! Talking of that, I was with a group of people looking at the promo last week when the test card came up - they all said "does this have to up for so long" and "what's that noise on the sound track" I did attempt to explain how much it was wanted but it just made things worse! I said I had wanted 5 minutes and many of the posts had asked for up to 30 minutes - at that point I felt like I came from another planet and decided to get on with other things! But there is a test card going out and I hope we can all celebrate its reappearance after many years!
The disappointment is a technical one. I am going to admit I have doctored the test card - much to the disgust of many of my Research colleagues. Why? Two reasons actually. A high quality test signal like the HD test card is a very valuable asset and unlike the SD transmission chain the HD one is quite good and quite capable in the being "purloined"! Already some of the posts on digital spy have already gone into great detail with the exact measurements of the card.
This version of the test card can be easily identified as it's the only version with the HD DOG logo at the bottom. Now, I want no DOG posts in this blog, I will ignore them as the DOG debate goes on %3Ca%20href="http://www.bbc.co.uk/blogs/bbcinternet/2008/10/dogs_on_the_blog.html">elsewhere.
What have I done and how useful is this version of the test card? First, white level has been reduced so the peak white box is not 100% (level 235 or 0.7v). The super white spot is now %3Ca%20href="http://www.flickr.com/photos/bbccouk/3113496182/" title="image002 by bbccouk, on Flickr">100% and the linearity of the grey scale is now slightly inaccurate. However no domestic displays have the level of adjustment we expect a broadcast monitor to have, so I this does not affect the usefulness of the test card to help you line up a "normal" TV. Also the colour bars are slightly lower in colour level. My apologies go out to people like Richard Russell and all the others who made these test charts possible - but this does protect the value of the work.
The second reason is to help protect screens from burn in. The full level test card will burn a screen in quite a short period so please heed this warning:
DO NOT leave the test card on screen for more than 2 minutes if your screen is less than three months old or more than 5 minutes on older screens. Make sure you go back to the promo for several minutes before using the test card again.
If you want more detail of the changes there is a%3Ca%20href="http://www.digitalspy.co.uk/forums/showthread.php?t=942385&page=4"> very good post on Digital Spy. If you do have a broadcast style display at home it is quite easy to %3Ca%20href="http://www.digitalspy.co.uk/forums/showthread.php?t=942385&page=4">calculate the offsets to apply to a colorimeter to make sure the readings are correct.
Now for a bit of an explanation about the test card and how to use it to line up your TV, I have done this at home so can say it does work. But before you start to line your set up please take note of the following:
1. Make sure you have the user manual and know where the controls are.
2. Do not do this if you are unsure of any of the controls or there effect on your television picture.
3. It is best to do this in a darkened room, it doesn't need to be completely dark but if it's too bright or there is a lot of light falling on the screen the results will not be good.
4. Many modern flat screen televisions have presets for sound and picture. Write down which one you use so if you get lost you can always go back and start again.
5. If you have a PVR it would be a good idea to record the test card section of the promo. Most of the line up can be carried out on a freeze frame of the test card. If you do this please be mindful of the warning above about screen burn.
6. If your TV has it, change the picture settings mode to "manual" or the equivalent, so any inactive controls become active allowing you to change the settings on the TV.
7. Turn the sharpness setting to off or zero. If there are any picture enhancing options, make sure they are turned off or to zero (if you can). Remember, on some TVs the sharpness control has a centre zero allowing you to soften pictures - please don't do that!
So to start:
BRIGHTNESS AND CONTRAST
%3Ca%20href="http://www.flickr.com/photos/bbccouk/3113496226/" title="image004 by bbccouk, on Flickr">There is a GREY SCALE to the left of the picture on the test card. It's there to show the correct black and white levels of the picture. Broadcast displays have the ability to adjust the grey level independently so there is a linear grey scale between the black and white blocks. I am not going into how to use this here if you are interested %3Ca%20href="http://en.wikipedia.org/wiki/Test_Card_F">start with this.
The top white block has two spots. As I said earlier, usually the white block is peak white with the right spot higher (super white) and the left spot slightly lower. On our test card, the levels are slightly reduced.
The bottom black bock has two spots, the right hand one is below black level (sub black) and the left is slightly above black. The modifications to the test card have not change these levels.
To set the brightness:
1. Turn your brightness control up %3Ca%20href="http://www.flickr.com/photos/bbccouk/3112664121/">until you can see both spots.
2. Turn the brightness down until the sub-black spot disappears but make sure you can still %3Ca%20href="http://www.flickr.com/photos/bbccouk/3113496304/">see the left slightly brighter spot.
When a broadcast monitor is lined-up properly, we use a meter to check the white level however on a domestic television contrast is more a matter of personal choice and will be different on different types of display (LCD, Plasma, Projector etc.)
Adjust the contrast until you like the overall look of the test card while you are doing this, keep an eye on the spots in the top white block to make sure you can still see the left hand one. It doesn't matter if you cannot see the super white spot so don't worry if it's not there.
Again colour level is very much down to personal taste but most TVs have too much of it! Too much colour makes pictures look very odd. It will also make some colours bleed into each other or appear to move so the colour smears over the edge of the object - in other words someone wearing bright colours clothes may have the colour slightly off to one side! The best bit of the test card to use to set colour is the picture of Carole.
%3Ca%20href="http://www.flickr.com/photos/bbccouk/3113496392/" title="image009 by bbccouk, on Flickr">The centre of the test card has all you need to get the colour right. Carole's face should look natural and the primary colours in the picture (red dress and green and blue of the clown) should not be very bright. Colour is a subjective setting so just make sure you like it. Remember, if your colour setting was previously set very high you may not like the correct level until you get used to it!
One of the experts at BBC Research suggested another way to adjust colour level.
Get some Lee Lighting Filters No.181 Congo Blue and place it over the screen. This has the same effect as turning off the Green and Red leaving the Blue component of the picture. Looking at the colour bars around the edge to the test card, adjust the colour control until they all look the same brightness. There are some commercially available line up DVDs that use this method.
When you have adjusted the BRIGHTNESS, CONTRAST and COLOUR have a look at the promo again to see what you think. Watch it for some time so you get used to the new settings and see several different type of programme.
I have the sharpness control on my TV set to zero but some of you may want to add a little bit if the picture looks very soft.
To the right of the picture of Carole is a set of "frequency gratings" The frequencies are:
%3Ca%20href="http://www.flickr.com/photos/bbccouk/3113496466/" title="image011 by bbccouk, on Flickr">1. 5MHz
The BBC HD transmission system will pass frequencies 1 - 4. Most domestic displays will show 1-3 correctly but the 4th might not look quite right. A 50' 1080p display should be able to resolve the 4th grating satisfactorily.
PICTURE SIZE AND POSITION
Not all TVs offer menu setting that allow you to change picture size and position. Even if your TV does allow you to adjust size and position, it's not a good idea not to make anything but small changes unless you know what you are doing. Make a note of the current setting BEFORE you change anything!
Most displays lose a small amount of picture all round. This is called "overscan", it is perfectly normal and programmes have always been made taking this into account.
%3Ca%20href="http://www.flickr.com/photos/bbccouk/3112664391/" title="image013 by bbccouk, on Flickr">Some flat screens do have the option to either turn overscan off or reduce the picture size.
It is perfectly safe to use the "overscan off" option on you TV but you should not use the picture size controls for anything more than small changes.
%3Ca%20href="http://www.flickr.com/photos/bbccouk/3113496672/" title="image015 by bbccouk, on Flickr">The full test card should look like this, with the diamond points just touching the edge of the screen all the way round.
As a mater of interest, the cross on the Noughts and Crosses game is the centre of the picture!
%3Ca%20href="http://flickr.com/photos/bbccouk/3113496824/" title="image017 by bbccouk, on Flickr">
You should now have a picture that looks fairly close to the one we see before transmission. Again watch some of the promo to get used to the new settings. Also if you have turned overscan off, you might want to look at some SD channels to make sure you don't see extra bit of the picture you don't like. You may see some white lines at the top of the screen on some News programmes for example. This happens when signals are brought back that don't fully meet the broadcast standards, but have to go to air too quickly (if not live) so it isn't possible to correct them.
AUDIO VIDEO SYNC
The second test signal is there to help you check and adjust audio/video synchronisation. AV sync has been the bane of my life ever since the test channel started. Remember we have rebuilt the HD Channel infrastructure round a service running 24 hours a day, 7 days a week so occasionally we have had to put new sections into service without being able to fully test them.
A couple of months ago, %3Ca%20href="http://rowan.depomerai.com/about/">Rowan stated %3Ca%20href="http://rowan.depomerai.com/2008/10/welcome-to-bbc-hd/">working with me to try and clear up our surround sound and AV sync problems. It has been a joy watching him dive headlong into the issues and some of you may %3Ca%20href="http://rowan.depomerai.com/category/bbchd-audio/">have visited Rowan's blog to see how his work is progressing.
The last time I saw him, he was waist deep in diagrams of installations and programme signal routes to see if each video process had a suitable audio delay and each Dolby process had an appropriate video delay in circuit. The idea audio processing "takes time" is relatively new to television and we have to remind people to compensate for processing delay in the appropriate place so we can make sure everything is correct it if anything needs to be changed - I hope the message is getting through!
However even when we get it completely correct, some home setups can cause A/V sync problems of their own. The second test signal should help you check and adjust the sync timing of your AV system. This does not work on audio fed through the HDMI cable to the televisions own speakers. Any delays in that situation should be compensated for inside the TV. A/V sync is only adjustable when you use a AV systems connected by the optical/SPDIF output of a set top box or for AV amps that can us the HDMI output and have their own delay controls.
As some systems could have two ways to adjust AV sync - the set top boxes will have an audio delay option in the set up menus and good AV amplifiers may also have audio delay options, you need to start by setting all delays to zero. Again, please make a note of the settings BEFORE you start.
Why is there a need for A/V sync adjustment now? Most flat screen displays introduce a delay while they process the picture before it's displayed. Inside the TV the audio is delayed to match the processing delay but if you connect your set top box to an external audio system, the sound can be one or two frames ahead of the picture. In nature this is not normal and we can detect sound ahead of vision very quickly and it is "just not right"!
Our transmission system can also introduce delay to both audio and video signals. Some of the delay is obvious e.g. if we send the audio through a Dolby E decode/recode process, the sound is delayed by 2 frames so we must add a 2 frame video delay. Other process aren't so easy to check as the delay occurs inside a device that's processing audio and video together so the reason the A/V sync signal was not transmitted two weeks ago was to allow us time to test our whole system to make sure what we send you is actually in sync!
So it's time to introduce the BBC Research sync check signal...
%3Ca%20href="http://flickr.com/photos/bbccouk/3113496946/" title="image019 by bbccouk, on Flickr">
The audio is actually two blocks of wood being banged once a second - nothing to beat the real thing! The video is made up from three components:
1. A travelling bar marked in frames starting 12 frames before the audio clap and going on for 12 more frames after.
2. Three Sync Flash Lines.
3. A sync "plunger" or "clapper bar" (acting like a clapper board)
Before I tell you how to use the signal, you might like to know what we did to make sure the signal itself was synchronised and the transmission system did not put it out of sync. We - or I should say Rowan - have been measuring the signal at every point in the chain to make sure it is as accurate as possible when it arrives at your set top box.
This is the task I set Rowan to about three weeks ago!
%3Ca%20href="http://flickr.com/photos/bbccouk/3112664875/" title="image021 by bbccouk, on Flickr">
Why so complex? I needed to make sure the signal followed the routes many of our programmes do through post production and audio mixing, playout and transmission.
What is sync though? If you think about the speed of light vs. the speed of sound it's fairly obvious that sound arrives a lot later that the image of the "thing" making it. A rough rule is audio takes slightly less than 3ms to travel 1 metre so if you sit 2 and a half metres from your TV the audio takes nearly 7.5ms to reach you - nearly a quarter of a frame.
The effect of AV sync has been measured and tested quite extensively by the international broadcast standards bodies and we usually work to a tolerance of +20ms to -40ms (+half to - 1 frame) for a programme delivered to the BBC. This tolerance has been well tested in SD but there has not been enough work done to see if it's still OK in HD. To make sure there were no major surprises we have tightened the delivery specification to +10ms to -20ms while we do more tests.
In one of my previous blogs I explained that during the trial we found sync varied during a programme, especially live programmes, depending on how hard some of the early equipment was working. Now we are a lot more stable and have had a chance to go through the system from end to end to make sure it's sync. We have just finished testing the chain with an "off air" test of the signal and have a timing error of 0.86ms or 0.0125 of a frame!
Those of you who work or have worked in the business know the phrase "it's alright leaving me". A translation of that is "I'm OK, it's your problem"! To make sure you can use the signal to check your home system, we have gone one step further and made sure the signal is "alright arriving at you", not just "alright leaving me".
%3Ca%20href="http://www.flickr.com/photos/bbccouk/3113497164/" title="image023 by bbccouk, on Flickr">The final measurements of the off-air signal were made by looking at the digital signal from the transmitter - or in this case, the one received from the satellite.
This is what the test signal looks like in the transport stream from the satellite.
Rowan has a more detailed explanation of how and why we did this in his blog.
Part 1: %3Ca%20href="http://rowan.depomerai.com/2008/11/dont-forget-the-kitchen-sync/">Don't forget the kitchen sync
Part 2: %3Ca%20href="http://rowan.depomerai.com/2008/12/testing-the-test/">Testing the test
How do you use the signal to check audio/video (AV) sync?
Remember this only works if you are connected to an external AV system. Check the audio delay setting in the set top box is 0ms and if your AV amplifier has a delay check that is set to zero.
Look at the travelling bar at a point before the centre - look say at 10 on the left of screen. Listen to the clap and see if you think the bar has passed this point before you hear the sound. You might want to mask the right side with a bit of paper or put your finger on the number to help.
If the audio seems to happen after the bar has passed, move on a number and repeat until you think the audio and the point the bar passes your maker coincide. The sync point could be between two numbers but most devices only make corrections in half frame increments you will have to decide if you think it's closer to a number or closer to a half way position.
You should still be on the "video late" side of the zero mark. Read the number (or closest half number) and multiply by 40.
If your number is 3, the audio delay you need is 3 x 40 = 120ms.
If your number is 1 and a half, the audio delay you need is 1.5 x 40 = 60ms
If your sync point is on the right hand side of the zero mark, I'm sorry to say there is nothing we can do to help. Before panicking - check you have no audio delay set then wait for the test signal to come round again.
For those who like a challenge there is an electronic method. The white lines flash for 1 frame at the start of the audio waveform.
I am sure you can think of many ways to use this information to measure AV sync accurately, but here is a simple option for all of you with a dual beam oscilloscope, a photocell and a microphone lying around!
If you place the photocell over the top sync flash line and the microphone on one of your front speakers, connect them to separate input of the scope (with any amplification devices needed to boost or power them) you will get two spikes. Make sure your scope is configured to display the two traces at the same time and measure the difference in ms. Apply this delay to either the set top box or the AV system. If you have a very good AV system you may be able to get this exactly right instead of the nearest 20ms.
If you can't decide between two settings, it is always better to make the audio slightly late than have it in front of the pictures.
Why are there three sets of white flashes? The top line is the reference line, i.e. 0ms A/V offset when measured on an HD CRT. However sometime it is difficult to accurately measure the very top of active picture, possibly because the TV's casing gets in the way of the photocell. The second line is 1ms later (as measured on a CRT) and is usually easier to get an accurate reading from. The third line is in the centre of the active picture so should read 10ms A/V offset on a CRT.
On the various LCD and plasma displays we have tried this on, some show a difference between the three lines and some don't - not much help to you, but I would go for the second line if you can and minimise the delay there!
Please let me know how you get on.
How to record the BBC HD test signals
The HD test card is just over 1 hour into the promo and the AV sync signal is 50 minutes later. To record both signals, check the time the last programme finishes and add 1 hour. So if the last programme ends at 01:30 set your PVR to record from 02:25 to 02:45 for the test card and 03:15 to 03:35 for the AV sync signal.
My last thoughts this time are around phenomena I have discovered that manifests itself around my daughter and what used to be my phone. When she is on the phone I have discovered time slows down!
How do I know this? Simple maths!
Let's take a telephone billing period - call it Tm. A Tm can only have 4 values 28, 29 (every fourth year) 30 or 31. Each Tm is made up from telephone charging units, let's call them Tu.
Each Tu is charged at various rates but I am going to use the maximum UK rate £UK (premium numbers etc are barred).
So my maximum phone bill can only be Tu x £UK. As the Tm can have four values the total bill can vary between each Tm. So why have my bills been consistently two to three times this amount?
My thought are, as the rate £UK is fixed in any Tm (but in general is always rising) and the periods of charging Tm are fixed to one of four values the only thing that can change is the Tu! As this is measured in time, I can only conclude time around my daughter changes as she uses the phone. Oh! I forgot to say, we are on the free weekends and evenings tariff so this time dilation effects occurs in the brief period between the end of school and 18:00!
I think Douglas Adams spotted this change to the laws of maths in the Hitch Hikers Guide - other explanations welcome!
Have a very merry, in sync and well lined-up Christmas and we will speak again in the New Year.
P.S. When she is on the phone to us each Tu is very long period of time of course.
Andy Quested is Principal Technologist, HD, BBC Future Media and Technology.