compatibility modes etc. - for hardware? (EyePal under Windows 7)

[J. P. Gilliver (John)]

Crossposted to another 'group, hence the addition to the subject - I
should have thought of adding that 'group in the first place. Anyone
from there who has experience of Eye-Pal under 7, please share!
(Tomorrow below is now today, but still some hours away.)

In message <jhndko$31n$1...@dont-email.me>, Paul <nos...@needed.com>
writes:
>J. P. Gilliver (John) wrote:
>> I'm going tomorrow to try to help out a blind friend who is trying to
>>use a scanner, with OCR software, that works fine under XP. Apparently
>>she's tried to load the software and drivers under 7, and not had too
>>much success.
>> The scanner is basically a downward-facing camera; there's slightly
>>more hardware interfacing involved, though, as it has lights in it
>>which can be controlled by the software. (If anyone's interested, it
>>is the 1.2 version of Eyepal rather than the 4.12 or later Zoomex,
>>which is the replacement, which I presume _does_ work problem-free under 7.)
[]
>The software is a 190MB download. Eye-Pal 1.2. The download page also
>has their version 4 stuff as well. No idea how any license keys fit into
>this. I haven't tried installing it, just looked at the files.
>
>http://www.abisee.com/downloads/EyePal_Setup_1_1_2.zip

Thanks for that. I should explain: my friend has been sold this by her
and my other friend, also blind, who has upgraded (?) to the Zoomex, but
is still on XP. I will find when I get there later today what discs etc.
have been sent with it! Though the recipient has broadband so the
download isn't a problem if I find we need it, unless the server is slow
or down.
>
>The abisee vision products are $1000-$2000 range, so kinda pricey.

My friend has sold it to her friend for somewhat less than that (though
it is as I've said not the latest model).

>But what you'd expect when developing for a niche market (developers
>gotta eat).

Oh, understood, though I feel some parts of the vish. market are worse
than others (either inefficiently-run or just taking advantage). It's
tempting to say that this particular combination is just a webcam, OCR,
and some speech output (and a rather cleverly-designed collapsible stand
that places the camera at the right height and aligns the target), but I
can I suppose see the cost may be justified - niche market, as you say.
However, I feel the prices in the market give at least me more
expectation of support for older products than is generally available.
>
>The voice synthesis is via a project called NVDA. This is one part of
>Eye-Pal 1.2.
>
>"NonVisual Desktop Access (NVDA) is an open-source screen reader
[]
> to do most tasks they would need to. It is not as stable or as bug-free
> as some of the commercial screen readers. But since December 2006 the
> creator of NVDA has been able to use NVDA full-time as his primary screen
> reader, finally giving up his original commercial product.
>"
I hadn't realised the EyePal uses NVDA; both of my friends have one of
the commercial readers anyway (I think they both have JAWS), so don't
actually need the speech output, though it's probably better fitted to
the hardware and software involved. (Incidentally, I've noticed that
equipment in the vish. market tends to be marketed as if it is a package
- in this case, the camera [and stand], camera control package, and OCR;
I suppose this is in many ways a sensible policy, but I sometimes feel
the fact that you _can_ use different software is unnecessarily hidden.)
>
>http://www.nvda-project.org/wiki/About
>
>The application itself could be written in Python (the thing that
>appears as the application, glues things together). I see some files
>with .py in the name.

Interesting!
>
>I see scanman.exe , and that could be ABBYY Fine Reader. That might be
>the part that does the OCR for the camera.

Ah, I know ABBYY; a reasonable OCR that is often supplied with scanners.
(Eastern European in origin, IIRR, but I could be wrong about that.)
>
>The user manual claims WinXP or Vista, so that implies Vista/Windows7.
>This is some text extracted from the three small PDF files in the ZIP.
>
> "Minimum System Requirements Windows XP (recommended) or Vista"
>
> "Note: Install Eye-Pal software prior to connecting
> the Eye-Pal device to your computer"
>
>Note that Vista is not "recommended" :-(

Hmm. Yes, it would be nice to know why (-:!
>
>The USB driver for the camera has a VID and PID and file name,
>which I cannot find in Google. A complete fail. So the camera
>driver may be the weak link.

Yes, I fear so. I know my first friend found that she had to have room
lights on to use it initially; it was only after she first involved me
that we found how to turn on the lights in the camera. (I can't
remember, now, whether that was a matter of updating the driver or
something like that, or was just an option in the software she hadn't
found.) After we'd got that going, IIRR, the lights came on when you
told it to do a scan.
[]
>The "CyMvEyep.sys" file is small. The description on the file
>says "Cypress Generic USB Device Driver". The file original name
>is CyUsb.sys. I didn't think Cypress made camera controller chips,
>but maybe they do ?

Pass - though it could be something to do with the USB side: when I've
been saying "the camera", I really mean the camera head, which is a
small (about 2 inches square) PCB, which contains the camera, the light
(white LEDs, three IIRR), and possibly other chips. (Needless to say,
the whole thing connects to and is powered by USB.)
>
>In any case, I'd say you should be in for some fun. On the one

I'm expecting it! (They are going to feed me! Though I've told them that
isn't necessary; I like helping out in this manner.)

>hand, the manual mentions WinXP/Vista for 1.2. But on the other
>hand, the 190MB download is a bunch of pieces of a puzzle that
>fit together. Increasing the odds something could break along the
>way.

Indeed! But if there is text about installing under Vista, I'll have a
look at that; it's possible the recipient hasn't followed that, if it's
hidden somewhere and she hasn't found it. (I just hope that, if I have
to uninstall things she has installed, they don't prove too stubborn!
I'm not that familiar with 7 and its foibles.)
>
>*******
>
>If you're going to be doing "Compatibility Mode", be aware that
>the 190MB package has a number of pieces. So there may be more
>than one executable that needs attention.
>
> Paul

Good point, thanks.
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G.5AL-IS-P--Ch++(p)Ar@T0H+Sh0!:`)DNAf

Old soldiers never die - only young ones

[J. P. Gilliver (John)]

In message <ilKI05zp...@soft255.demon.co.uk>, "J. P. Gilliver
(John)" <G6...@soft255.demon.co.uk> writes:
[]

>In message <jhndko$31n$1...@dont-email.me>, Paul <nos...@needed.com>
>writes:

[]
>>The software is a 190MB download. Eye-Pal 1.2. The download page also
>>has their version 4 stuff as well. No idea how any license keys fit into
>>this. I haven't tried installing it, just looked at the files.
>>
>>http://www.abisee.com/downloads/EyePal_Setup_1_1_2.zip

Thanks - she'd already downloaded it, as it happens. (And my other
friend had sent it on a CDR as well as the official 1.1.1 CD, but the
CDR wouldn't read.) I don't think any licence keys are involved - it
being part of a package, you have to have bought the hardware for it to
be that useful. I suppose the OCR software could thus be used free with
images from other sources, but that wasn't our intention.
[]

>I hadn't realised the EyePal uses NVDA; both of my friends have one of
>the commercial readers anyway (I think they both have JAWS), so don't
>actually need the speech output, though it's probably better fitted to
>the hardware and software involved. (Incidentally, I've noticed that

[]
Interestingly, it spoke with different voices on the different systems -
I'm talking about the Eye-Pal software, not the already-present speech
software, though there may well have been some interaction between them.
On the Windows 7 system (which has JAWS as its basic speech), Eye-Pal
spoke in an excellent voice; on the XP system (which had some cheaper
speech - I forget the name, but it's one that goes on-line to work),
Eye-Pal spoke with a more robotic voice.
[]

>>The user manual claims WinXP or Vista, so that implies Vista/Windows7.
>>This is some text extracted from the three small PDF files in the ZIP.
>>
>> "Minimum System Requirements Windows XP (recommended) or Vista"
>>
>> "Note: Install Eye-Pal software prior to connecting
>> the Eye-Pal device to your computer"
>>
>>Note that Vista is not "recommended" :-(

[]

>>The USB driver for the camera has a VID and PID and file name,
>>which I cannot find in Google. A complete fail. So the camera
>>driver may be the weak link.

Yes, seems it was - I think.
[]

>>In any case, I'd say you should be in for some fun. On the one

>>hand, the manual mentions WinXP/Vista for 1.2. But on the other
>>hand, the 190MB download is a bunch of pieces of a puzzle that
>>fit together. Increasing the odds something could break along the
>>way.

[]

>>If you're going to be doing "Compatibility Mode", be aware that
>>the 190MB package has a number of pieces. So there may be more
>>than one executable that needs attention.

[]
I ran out of time (when I visit them, I also fix other things - last
night this included establishing that her flatbed scanner was not
working because (a) its power lead had come out (b) its USB lead was
broken - fortunately she had another. And charging her van battery. And
trying to find why a Samsung 310 laptop was dead [not successful: its
PSU was fine, but it showed no signs of life other than a
slowly-flashing LED, orange if battery present green otherwise.
Apparently it had done, but had been running on battery, until that ran
down.]).

Anyway, here's what we found - four systems (three really).

Firstly, an XP tower system. We thought we'd try this first as we know
it works on XP. The software loaded OK; we plugged in the EyePal; the
driver loaded OK. (I should say we tried with 1.1.1, from the original
disc that had come with the scanner, and 1.1.2 from the downloaded zip -
no difference.) Run the programme: it says either "The software detects
that no device is connected" but you can still use earlier results, if
the EyePal isn't plugged in, or "camera started" if it is. When the
software starts, a sighted person sees on screen what the camera is
seeing. All appreas to be working, except that the screen shows lots of
diagonal black and white lines, like a malfunctioning TV. (No, not video
with lost lock - there was no sign of what was under the camera, though
that _could_ have been because it has some sort of self-adjustment and
it might have saturated.) The surrounding part of the display - the
menu, close buttons, and so on - were as they should be, only the part
where the camera image should be was stripy instead (or showed plain
text if the camera wasn't connected). We put some text under it anyway,
but as soon as we pressed the appropriate keys, it said "no text". I
_think_ the toggle for the light in the camera didn't work either.

Secondly, this netbook; I already had the software 1.1.2 on here, from
when my friends had shown me the device some months ago. This was really
just to check that the Eye-Pal was still working. All worked fine.

Thirdly, the Windows 7 tower. The software loaded (and runs) fine; the
driver (tried before and after plugging in the device) no luck. My
unfamiliarity with Windows 7 means I didn't find where to tell W7
manually where to look for a driver, which the (sparse) documentation
says you should do rather than let it do its automatic thing.

I looked at the FAQ on their website. One of the FAQs was do AbiSee
products work under Windows 7, to which the answer was the single word
Yes; however, I have a nagging feeling that this may not apply to the
earlier EyePal (this is a 3 megapixel one not the later 5 - I'd
established that earlier, while establishing, for the previous owner,
that it could not be used with the version 4.x software). Another FAQ
was something like "what does it mean when the software says no device
found", to which the (fairly obvious) answer was something like "this
means that the driver is not loaded properly. See loading drivers."
Unfortunately, "loading drivers" or whatever was _not_ a link, which
wasn't helpful.

The website did mention a support number (free 1-800-681-5909 for US
users); we tried it via Skype, but unfortunately there was no-one there
(we got through, just recorded announcement giving opening hours -
though it didn't say which timezone or which days of the week). My
friend may well call them again during the week: I'm assuming AbiSee's
Erica is used to dealing with blind users over the 'phone, so that
shouldn't be a great problem.

Finally, we tried a Vista system - a rather neat little Philips laptop
(11.1"), which my friend doesn't use much as it's rather slow. (She's
right - it is; not particularly slow in general, but seems to stop
responding, to do I know not what, rather a lot.) The software and
driver loaded fine (though taking ages), and the whole thing ran fine -
and, once the software had started, ran very quickly (started to speak
the text as soon as the E key was pressed). Even the light toggle
(Ctrl-L) works. (We found turning the light _off_ for glossy material -
like the back of a shiny book - made it work better, provided there was
room light; the EyePal's light is in the camera head itself, so causes a
flare in the centre of what it sees if the material is glossy. For matt
material, the light is generally better on.)

Just thought I'd report back in case anyone's interested! My friend's
off to Germany for ten days from tomorrow so nothing will happen for a
while, but I'll report further after that.

Oh, I wondered if the XP system - which seemed to think it was working
but showed stripes instead of the camera image - had a video driver
problem. I looked at what video card it thought it had, and it seemed to
think it had two, one with an extra three letters; when I looked I could
only see one card, though it does have both a VGA and a phono output
(and I think an S-video one). If I remember rightly, it was something
like Radeon 9200 SE (with its other half as something like Radeon 9200
SE DEC). I could well be remembering the number, and especially the two
and three letter groupings, wrongly - this was just a final thought
before I left. If anyone knows that this causes problems of this sort,
please share (-: [especially if you know the cure!].
--
J. P. Gilliver

War is God's way of teaching Americans geography. -Ambrose Bierce, writer
(1842-1914)

[Paul]

J. P. Gilliver (John) wrote:

When displaying a video rectangle on the screen, there are a few options.
At one time, this was called the "overlay plane", and was a hardware feature,
limited as to how many video items you could display at a time. In programs like
Windows Media Player, you might have a preference, that selects "overlay
plane", or some other standard, such as VMR7 or VMR9. I don't know how
other programs would handle this, or whether Windows is clever enough to
have backward compatibility. Sometimes, when this fails, you get a single-color
box, instead of your video content. This also causes issues, if you're capturing
the screen, as 2D content, video rectangle, and 3D content, may require separate
procedures to be captured (such as with FRAPs capture).

http://en.wikipedia.org/wiki/Directshow

"DirectShow 6.0, released as part of DirectX Media introduced the
Overlay Mixer renderer designed for DVD playback and broadcast video streams"

"Windows XP introduced a new filter called the Video Mixing Renderer 7 (VMR-7
or sometimes just referred to as VMR)."

"DirectX 9 included VMR-9. This version uses Direct3D 9 instead of DirectDraw,
allowing developers to transform video images using the Direct3D pixel shaders."

"Windows Vista and Windows 7 ship with... Enhanced Video Renderer (EVR)"

With regard to plumbing the driver side of things, I'd probably start with the
"working" system, open Device Manager, and see what item(s) are listed in the
drivers for the device. Just to see whether there is something there other
than that Cypress generic file.

I ended up with the Cypress generic file on my system, when restoring my
USB IDE enclosure to health. The USB chip (Cypress branded) lost correct
config space info, and during the procedure to write info back to flash, a
driver was needed so that the repair procedure could "see" the USB chip. Normally,
the OS uses config space info (like "class" info), to figure out what kind of
driver to use. Devices not following a standard, have a code that says "I'm custom",
so that a custom driver can be used. I don't see exactly how the Cypress generic
fits in there exactly. In any case, I'd be probing Device Manager for an answer.

You might want to contact the support at Abisee, and see if there is an issue with
Overlay Plane for the camera video.

Cameras can be complex to deal with. I understand that some of them, the OS driver,
actually downloads firmware into the camera controller chip. So the camera won't
do things as you'd expect, unless the driver loads the firmware. Firmware allows
"enhancements" in behavior, after the product has been released.

Paul

[J. P. Gilliver (John)]

In message <jhrgk7$2l8$1...@dont-email.me>, Paul <nos...@needed.com>
writes:
[]

>When displaying a video rectangle on the screen, there are a few options.

I suspect it is a problem with something like that.
[]

> "Windows Vista and Windows 7 ship with... Enhanced Video Renderer (EVR)"

The problem with the Windows 7 system is that the correct drivers for
the camera were not loaded - probably because my unfamiliarity with '7
meant I did not know how to tell it to go off looking for them but
instead to look locally.

>
>With regard to plumbing the driver side of things, I'd probably start with the
>"working" system, open Device Manager, and see what item(s) are listed in the
>drivers for the device. Just to see whether there is something there other
>than that Cypress generic file.

Good thinking. I will do that next time I visit my friend. (There's no
sign of it at all in DM at the moment, presumably because it's not
plugged in.)
[]

>driver to use. Devices not following a standard, have a code that says
>"I'm custom",
>so that a custom driver can be used. I don't see exactly how the

I suspect it is, since it's a camera and a light, which webcams etc.
don't usually have AFAIK, at least not under software control. (When
it's working properly, Ctrl-L toggles the light [normally better with it
on, but since the light is in the camera head itself, better off when
reading something glossy].)
[]

>You might want to contact the support at Abisee, and see if there is an
>issue with
>Overlay Plane for the camera video.

I've just emailed them, describing the symptoms. I await an answer with
interest.

>
>Cameras can be complex to deal with. I understand that some of them,
>the OS driver,
>actually downloads firmware into the camera controller chip. So the
>camera won't
>do things as you'd expect, unless the driver loads the firmware.
>Firmware allows
>"enhancements" in behavior, after the product has been released.

[]
I see. I don't know if there's some adjustment in the camera or if it's
all in software in the computer: when it's working, when you start it,
the brightness of the image on screen seems to adjust itself until it's
about right.

--
J. P. Gilliver. UMRA: 1960/<1985 MB++G.5AL-IS-P--Ch++(p)Ar@T0H+Sh0!:`)DNAf

What would be unusual would be if there weren't any coincidences at all for
several days in a row. Andy Roberts (UMRAt), 23rd. October 1998.

[Paul]

J. P. Gilliver (John) wrote:

I've seen webcams with infrared LEDs around the periphery of the
lens, so the configuration is not that obscure. (Its used for
surveillance applications, when the manufacturer wants to pretend
the camera has a "zero light" capability.) What I don't know,
is if LEDs and their control, are part of UVC standard or not. If they weren't
included, the camera manufacturer could create a composite USB device,
and the LED control would be a separate block in the design, with
its own endpoints.

My webcam is a composite device, and the video capture and audio
capture from a built-in microphone, are separate devices behind
the composite device. The LEDs could be handled in a similar way.

In the past, I would have suggested using UVCView, but when I worked
through an example with another poster, we didn't really learn anything
useful. Since these camera designs can have multiple VID/PID identities,
it's easy for the real manufacturer of the camera and imager chips to be
hidden, thereby making it impossible to find another driver for the
thing (if it needs a driver). This is unlike some other kinds of
computer hardware, where a one to one mapping is maintained, and there
is less "monkey business".

Paul

[J. P. Gilliver (John)]

In message <jhs6ka$2kh$1...@dont-email.me>, Paul <nos...@needed.com>
writes:
[]

>I've seen webcams with infrared LEDs around the periphery of the
>lens, so the configuration is not that obscure. (Its used for
>surveillance applications, when the manufacturer wants to pretend
>the camera has a "zero light" capability.) What I don't know,
>is if LEDs and their control, are part of UVC standard or not. If they weren't
>included, the camera manufacturer could create a composite USB device,
>and the LED control would be a separate block in the design, with
>its own endpoints.

I haven't played with the web version of such things, but I have with
such cameras that produce an analogue composite video output: they
switch of their own accord, with no interaction from the person at the
monitor end, from colour and IRLEDs off to monochrome and IRLEDs on - I
presume it's based on average picture level output (presumably with some
hysteresis). I suspect the webcam variants are the same - i. e. no
control, from the software end, of which mode the camera/lights
combination is operating in.
[]

>In the past, I would have suggested using UVCView, but when I worked
>through an example with another poster, we didn't really learn anything
>useful. Since these camera designs can have multiple VID/PID identities,
>it's easy for the real manufacturer of the camera and imager chips to be
>hidden, thereby making it impossible to find another driver for the
>thing (if it needs a driver). This is unlike some other kinds of
>computer hardware, where a one to one mapping is maintained, and there
>is less "monkey business".
>
> Paul

I'm sincerely hoping not to have to get into that sort of detail just to
get the EyePal working!

--
J. P. Gilliver. UMRA: 1960/<1985 MB++G.5AL-IS-P--Ch++(p)Ar@T0H+Sh0!:`)DNAf

I multitask... I read in the bathroom.