Video capture with I-O DATA GV-MDVD2 and GV-MDVD3 on Win7 x64

 The

I-O DATA GV-MDVD

Is a product line of video capture boxes from Japan , they aren't often seen outside of Japan because they use the NTSC-J variant of the NTSC color signaling standard which has a wider sampling range than in North America with the Pedestal Set-up. 

They are "capture" boxes, and so are not detected as Broadcast Tuner capable devices and Windows does not support their use with the Windows Media Center products.

They produce a rather unique MPEG-2 stream which contains both audio and video in their output pin, which can be directed to a file, or decoded and displayed within programs that understand the MPEG-2 stream. VLC is the most often used. VLC is also capable of saving the output to files in various formats.

The GV-MDVD2 has no 64 bit windows device driver, however it does work with the later GV-MDVD3 64 bit device driver. In fact it is detected as a GV-MDVD3 device. The hardware MPEG-2 encoder chip is the CX2316-12 and the video decoder chip used to process incoming video signals is reported to be the PHILIPS SAA7115HL. It is my guess the difference between the GV-MDVD2 and 3 is only with the driver disc provided and the Windows versions marketed towards. They appear to be the same hardware with different labeling.

One difficulty is routine access to the hardware switch within the device called a video switching Crossbar. This can be accessed either using the 32 bit version of GraphEdit with a GRAPH.GRF file. Or with the widely available "Crossbar Thingy" application. Either of these will load the Windows DirectShow component installed along with the device driver which allows a dialog window that allows selecting the Input source, to switch from Composite to S-Video.

Fair warning, the GRAPH.GRF file must be "opened" or "dropped" on an x86 (32 bit version) of GraphEdit or it will fail to open. Some components are not compatible with the 64 bit only version of the GraphEdit program, but work just fine with the 32 bit version of GraphEdit running on Windows 7 x64 bit natively.. and "not" in Windows XP mode.

I have not tested, but its suspected versions of MyDVD, Intervideo WinDVR and ULEAD capture products will "probably" work, since those cater more towards the MPEG-2 streams.

Windows Movie Maker, being for Raw YUV 422 or UVC products does not work with the GV-MDVD products.

The GV-MDVD (1) product is not like the GV-MDVD2 or GB-MDVD3 products, being from an earlier time it is based on the popular WisGO007 pre-compression hardware chip.. which produced a YUV output or a soft compression GO stream which could be software compression "finished" into MPEG1/2/4 versions. This sounds more versatile, but it is not by today's conventions.. because it relies on many software libraries that are only available on the Windows XP platform and were not brought forward. The Monsoon Snappysoft software from Luminate is known to work with it on the Windows XP 32 bit platform only.

I-O Data seems to have had a very long import line into the Japanese market with the GV-MDVD product line which seems to have favored importing the YUAN video capture products.. but details are scarce.

Using VirtualDub 1.9.11 to capture Video from ADVC-55 on WinXP

VirtualDub is a pretty old, familiar program to most people.

It starts in Edit mode, but you can go into Capture AVI mode from the File menu.

However with the ADVC-55 the initial window is Blank or Black with no sound.

To begin preview, the AVC Tape controls must be used to "Start" the Preview.

This is rather un-intuitive.

Normally you:

1. Go to Device  "to select" > Microsoft DV Camera and VCR (DirectShow)

2. Go to Video "to select"  > Video source (not necessary with ADVC as it auto selects Input)

3. Go to Audio "to select" >  Enable audio playback

 

But the display will remain Blank / Black

 

For a Firewire or ADVC device you also need to:

4. Go to Video " to select" > Capture filter ...

A Dialog called "Digital VCR Control" appears, simply Press the green Play button [ > ] and then Press [OK] to close the dialog and return to the running Preview.

This is rather different from other programs like Windows Movie Maker, in that in those you have to configure the program to allow "Manual" rather than "Automatic" capture control. -- In fact its the opposite procedure conceptually.

It makes apparent that most capture programs will attempt to "Start" the stream by itself automatically when the program starts, rather than waiting for the User to queue up video manually on a playback device, or to start the stream. This can be a problem for things like DV codecs since they cannot start until a video signal is actually coming into them.. they will produce an error code. With a real DV camcorder or video camera capable of reacting to the Start command, the video signal starts playing before the codec is engaged and there is no error.. but not so for a standalone DV codec which expects a video signal to be present and will fail and shutdown if one is not present before starting the encoder.

Its just not a step that most people consider, they assume "something" will send a blank signal which the encoder can use to start producing blank frames before the video signal is actually queued up.. but such a device is generally very expensive in the video world and is called a "frame synchronizer" and would have driven the cost of a device from those years up.

Instead, the user starts the Video to generate frames which the encoder then starts to process to produce a stream, and then the VirtualDub program assumes the encoder is sending frames that it can preview.

For raw video capture its similar, but frames are always there from the moment video is presented.

Its the option to "automatically" command the DV camcorder that makes all of this "different" from a stream that is always running. The pros and cons of having complete control.

How to block Phantom Floppy device / drivers from setup and loading in Windows XP

On a recent new laptop setup that had no Floppy drive. 

I kept getting a Phantom Floppy Drive in the Device Manager.

It would load the fdc floppy drive controller service and the floppy device driver and fail to start, producing the familiar Error 10 "fail to start" and a Yellow cone Warning sign opened in the Device Manager tree.

This was annoying, while from Vista forward you could use a Group Policy Object security policy to "Block" loading device drivers for specific Hardware IDs. You can't do that on Windows XP. And if your using Home Editions you often don't have Group Policy Objects or a GP Editor.

I found you could "Fix" this by first uninstalling the Phantom drive from Device Manager then going to C:\Windows\Inf and temporarily removing the fdc and floppy (.inf) files from that directory (both the .inf and precompiled .pnf versions) (say move them to the Desktop) then perform a Manual Scan for Hardware Changes in Device Manager.

It will fail to find the fdc and floppy device drivers, and setup failed "stubs" in the Device Tree.

Manually selecting and installed the "Modem/Unknown Modem" device driver for the Phantom will not set up a "service" and will collapse the Device Manager tree as "satisfied" Permanently... with no Yellow Cone Warnings.

It is a "trade" from a Phantom Floppy Drive to a Phantom Unknown Modem device, but much less alarming.. and typically the Unknown Modem is used as a proxy for the example of a "Null Device Driver" that doesn't install or register any start up service for a device in literature.. so its kind of a "lazy mans" way of writing a Null device driver for a Hardware ID that is either wrong or fake.. that you would rather did not produce a false positive error in the logs or on the Device Manager dashboard.

 After doing this replacing the  fdc and floppy .inf /.pnf files into C:\Windows\inf does not return to trying to dislodge the manual assertion between the device and the Unknown Modem "device driver". Future reboots and manual hardware scans continue to report the device is installed and associated with the Unknown Modem "device driver" and do nothing annoying. This is good in case you do install a really floppy device via USB or some other port means.

Using Firewire and Thunderbolt 1.0 with Windows XP Professional

So.. the other shoe drops.

Windows XP works perfectly fine with a laptop that has a Thunderbolt 1.0 port (the mini-display port kind).

And at least when the HP Thunderbolt Firmware modification too is used to disable the "Only Certified for Windows" Thunderbolt port block. 

Windows XP works fine with the Apple Thunderbolt to Firewire adapter.

I am not talking about using Apple hardware with Bootcamp. I am talking about using HP hardware running Windows XP Professional (and its probably not specific to "Professional").

The one thing different between Windows 7 x86 32 bit edition and Windows XP x86 32 bit edition when setting up a Thunderbolt connection is that after its hot plugged into the Windows XP system, it requires a reboot to start the service setup for the PCI bridge that is detected, until then it will have a Yellow Warning cone in Device Manager for it.. after it reboots it will finish enumerating and finding the OHCI port on the other side of the adapter.. this kind of makes perfect sense for the way that Windows XP does things.

On the laptop I have the ACPI HAL works for Windows 7 x86 32 bit, but does not for Windows XP x86 32 bit.. I had to use nLite to shoehorn the AHCI device drivers into the installer, make a .iso image.. and used an IODD external CD/DVD drive emulator to boot from the .iso image to start the installation.

When it got to the F6 additional device driver step, I used the undocumented F5 to select the specific HAL. This had the side effect that ACPI can never be used to finish shutdown in Windows XP.. and I did try all the other HALs.. but only Standard PC (legacy support) worked on this particular laptop. 

Until I selected a specific HAL it would always shutoff after the F6 step and not complete XP installation.

Once I bypassed the autodetect of the HAL and continued the installation process.

I connected a Canopus ADVC-55 to the Apple Firewire adapter and it was dutifully setup and the Autoruns service of Win XP offered to start Windows Movie Maker to capture video.. which worked as expected.

In any event, the Movie Maker captured video also captured sound and imported it into the Movie Maker application for editing.

This particular example of using Thunderbolt 1.0 with XP on an HP laptop is rather unique in that at the time the laptop would have been over $1500 in 2012 dollars, so not many people may have tried it.

What stands out as very interesting is that while Thunderbolt has a firmware layer, it is independent of the general operating system. Once configured with tools for other operating systems it can be used with any operating system. To all operating systems it just appears as a hot plug PCI bridge device, any scanning of the other side of the bridge is at the discretion of the operating system once the port is in use.

Using Firewire and Thunderbolt 1.0 with Windows 7 x86 Professional

 Along about 2010 Firewire was beginning to be abandoned in favor of USB and faster alternatives in laptops. However there was still a good use for Firewire or IEEE1384 ports for connecting legacy hard drives, cameras and video capture equipment.

For a time laptops could still be found with Expresscard ports and used with Expresscard to Firewire / IEEE1384 ports but eventually those went away as well.

Sony Vaios, Lenovos, Dell and HP products initally had some laptops with "mini" IEEE1384 ports but they too disappeared.

Apple Macbook Pros eased the transition by including a Firewire 800 port and a Display port, and then replaced the Display port with the identical port on Macbook Retinas with a Thunderbolt 1.0 port.

It wasn't called Thunderbolt 1.0 but retroactively can now be referred to as Thunderbolt 1.0 since there were Thunderbolt 2.0 and Thunderbolt 3.0 versions later.

Thunderbolt 1.0 is basically a Display port connector with additional lines to support  PCIexpress connections in the same connector as the original mini Display port connector.

Thunderbolt 2.0 is essentially the same as Thunderbolt 1.0 using the same mini Display port connector with a logical speed boost from 10 Gbps to 20 Gbps.

Thunderbolt 3.0 uses a different connector (USB-C) and also supports a USB 3.0 mode as well as getting a speed boost to 40 Gbps.

Besides Apple, a few manufacturers also made early laptops with Thunderbolt 1.0 connectors. HP being one with the Spectre XT 15 Ultrabook line. These had one mini Display port Thunderbolt 1.0 connector. 

Designed for Windows 8 (and later upgradeable to 8.1 and 10) out of the box, they would only support Thunderbolt 1.0 devices certified for the Windows OS.. and would not support The Apple Firewire to Thunderbolt adapter.

However HP retroactively released a Thunderbolt 1.0 port configuration tool for their Thunderbolt firmware that could enable support "Any" Thunderbolt connected device whether it was certified with Windows or not. This enables Windows 8/8.1/10 to support the Apple Firewire to Thunderbolt 1.0/2.0 adapter.

Basically once this tool is run the Thunderbolt port is set to enable all Thunderbolt devices connected to its port. This is not the default because of the Intel specification barring support for certified devices for specific operating systems. Its rather nonsensical and unintuitive. Without doing this inserting the Apple Firewire to Thunderbolt adapter only adds six PCI to PCI bridge devices to the device manager tree and stops.. it does not enumerate the 1394 OCHI device on the other end of the adapter.

The HP Spectre XT 15 is based on the Intel IVY chip architecture which means it fully supports Windows 7 and has a Legacy CSM BIOS mode for booting from MBR partitions.. which means not only is Windows 7 x64 (64 bit mode) supported, but Windows 7 x86 (32 bit mode) is also supported.

Once the Thunderbolt port is configured "open" to support uncertified Thunderbolt connected devices.. this carries over if Windows 7 x64 is installed, even for Windows 7 x86.

What happens when the Apple Firewire to Thunderbolt adapter is inserted (hot swapped) into Windows 7 x86 (32 bit) is six PCI to PCI bridge devices are created attached to one of the PCI express root ports, off of one of those a 1394 OCHI device is enumerated, and then an AVC and AVC Tape device is created.. if you have a firewire video capture device like an ADVC Canopus video capture codec attached.

The default 1394 OHCI device driver is not best and should be swapped for the (Legacy) version for regular work.. but it works.. in 32 bit.

This is a pretty slick method of using a Thunderbolt port on a Windows PC to connect up a native Firewire device for video capture and feed that into many 32 Bit Non-Linear Editor workflows from the past.. like Sony Vegas, Adobe Premiere and so on.

This also worked for later HP Zbook and Folio Ultrabooks.. as long as you manually downloaded the HP tool for re-configuring the Thunderbolt port and opened it up.. for Thunderbolt 2.0 and Thunderbolt 3.0 later chip architectures however make it less compatible with older operating systems.

The HP Spectre XT 15 is a little unique in having the IVY chipset since in theory it can also support Windows XP .. a rare thing today.. if you need it. And it will be interesting to see if Thunderbolt 1.0 can be supported under XP in a native manner.. stay tuned.

What if we're Pinball Machines not Memory Machines?

 Thinking about thought.

We tend to categorize by like is like, associations.

So we think of our brain and our thoughts like a computer programming running on a central processor.

Its something we sort of intuitively understand, and is sort of like the machines we hold in our hands. Turn crank, procedural kinetic motion produces output.

Memory is a little more like referring to a pen and paper and referring to a book for past mechanical motions.. the results.. become memory.

And we think our memory is like a book, the end result of our past mechanical actions.

But what if its more like a result machine, not that it methodically copies and encodes inputs like writing on a strip of paper.. but more like a spikey soccer ball with lots and lots of spikes.. that the envionment twists and turns and tumbles and leaves it laying in a pseudorandom state.

Not in a methodical easy to copy and recompose manner.. but the end result of a wad of paper crumpled and torn, and shaped by endless unpredictable blows and interactions with its environment.

More like a 'pinball' in a pinball machine.. its not as much a copier of the inputs from the environment.. as it is a reactor machine that seeks merely survive to experience another day.. an endless purgatory machine endlessly suffering.. the best outcome is not to cease functioning.

To extract memories.. or something more portable to easy to download and upload then.

Might be much harder.

We might rather need to  subject this tumbling soccer ball to a virtual environment to observe how its current state 'reacts' to all of its past experiences and currently interprets that which we erroneously perceive as the 'past' at the present time.

In such a scenario then, the 'past' or 'memory' becomes an endless sense of partially true, but also partially self described recollections of what its been through.. its not so much what it remembers.. as it is how it dictates or understands as its current view of the past. So.. in order to 'recall' the past.. it must forma virtual environment.. a virtual machine inside its head.. and use its imaginary narrative pointer and point it at this virtual machine in order to play out scenarios.. sort of 'imagine its 10 years ago and I'm walking through my house.. what do I see?' Then the current brain or mind 'state' begins reacting to the virtual environment and feeding senses of sight, hearing, smell into its current inputs via virtual imaginary sensors and playing out the Theater as if it were a play.. and holding those experiences  in something akin to a temporary memory buffer in our heads.. much more akin to a linear computer memory buffer.. so we can use those experiences as a near term second set of senses that can act as a mini-brain in a virtual environment to model a scenario we are trying to solve for.. to predict a near term desirable outcome.. sort of a Monte Carlo situation where we try try again in our tiny version of minecraft to come up with a solution we think will be worth betting on an the real world with our bodies.

If true.. if you want to tap into that.

It might be to extract or copy a memory.. or make new ones by uploading.. you need to tap into the virtual machine used for near term temporary memory.. and the senses it uses for input and output.. be they real or imaginary.

The real senses.. using head gear and tactile feedback like mice our keyboards.. we kind of have a grasp upon.

The temporary virtual machine senses of near term memory, we kind of don't have a good grasp of.

Magnetic Induction, neuralink.. are pokes in that direction.. but until we have a more solid connection to the temporary memory virtual machine and can spawn a set of neuralink virtual senses which the brain will accept as 'like' those it creates on the fly for visualizing scenarios for solving problems.. it could be a tough road.

Video games and game console controllers are close, they can immerse a person in a world without physical contact.. even trigger an imaginary experience where we use our perceived memory of reaction based on our past experiences. In such a situation the first person shooter is literally building their own bridge to the outside world and reacting based on their memories.. downloading  in that scenario is akin to interrogating a player for what they recall.. in real time.. since their arms and legs.. can only move in real time.

But we know we can think much faster than physical action.. the effective bandwidth can be increased in a dream state.. abbreviating or hopscotching across a dreamscape to lessor or more important elements of the story.

And in fact we know we can Upload memories very easily.. by watching a movie. Again the brain finds a way to bridge the gap.

Given the way the brain uses glucose and oxygen.. it may be that a person can only upload and download at faster than real time in a subdued, physically inactive dream state.. in which they are paralyzed to optimize nutrients and oxygen and blood supply for the brain.

Startech USB2TVTuner is based on EMPia Audio and EMPia Video chips Not WIS GO007

 

So it turned out the Startech USB2TVTuner is not a hardware compression based device. Hailing from the year 2002 and sold up thru 2007 it was basically a simple YUV digitizer with USB bridge to get the raw 4:2:2 from the capture chips to the software on a PC.

I found this out after examining a unit that no longer worked.. the hardware was busted in some manner. 

And confirmed looking into a device driver .INF file which only referenced EMPia capture chip technology.

This leaves the ADS DX2 Express and the Pinnacle dazzle AVC-130 and AVC-170 as the most common WIS GO007 hardware capture devices. But only the ADS DX2 had a not ready for prime time 64 bit device driver that was never signed and roughly works under Windows 7 x64 with difficulty.

It appears Micronas and then TDK acquired the WIS chip rights and it might have been taken off the market or folded into another portfolio.. but WIS based hardware compression capture chips disappeared after XP support basically ended for things like the Plextor M402 and related series.

ATI had a go at continuing MPEG2 hardware compression until the end of that company being acquired by AMD. ATI had the 550, 650 and later 750 chips with mostly Windows Media Center support under XP and then Vista.

Lumanate would produce the excellent Dell Angel MPEG USB series, which targeted the Windows Media Centers worked with Monsoon SnappySoft capture software.. and AMCap (after a fashion).

AverMedia of Taiwan and Hauppauge of New Jersey/China (?) are still offering products and had a long line of offerings both with recognized and individually proprietary and their supported versions of capture software.