Mt Fuji - MMC:

David Burg daviburg at
Tue Sep 4 07:57:22 PDT 2007

* From the T10 Reflector (t10 at, posted by:
* David Burg <daviburg at>
Hi Katata-san,
I believe Microsoft has already explained repeatedly the necessity of these
requirements, either on writing like on proposal document, or orally at the
meeting itself. But I will repeat briefly.
4-0 Playback throughput issue is generic
4-1 'Duh' requirement - can't select speed if can't list (e.g. to user) what
speeds are available.
4-2 Need to provide report of device capabilities at any time. Just like you
do already in the manual and promotion material for the hardware. Another
host software company actually confirmed this requirement to us offline.
4-3 The specification already points that the inability to distinguish some
speeds and rotation control is a problem. Further host can't reliably predict
burn time when rotation control is not accurately described.
4-4 'Duh' requirement. Don't provide host with fake data.
4-5 Based on history of problems with speed control, Microsoft intends to
closely verify that new specification is precisely and properly implemented.
4-6 The group had lengthily discussion why devices do adjust speeds.
Furthermore, the group already discussed that host applications are in 3
categories - read only (e.g. video player), write only (e.g. mastering
applications), read-write (e.g. live file systems)
Per the streaming feature introducing description "This Feature identifies
logical units that support reporting and setting of performance parameters."
The main performance parameter is the rotation speed. And not only special
playback software needs to control the rotation speed. Accordingly I believe
it would be improper to limit the designs to special software.
Reporting check condition on read commands as a persistent setting for the
duration of the media been mounted is a heavy responsibility for the host
application doing the playback. Risk of problems with other applications on
the system, e.g. if the special application exits unexpectedly or simply
forgets to restore speed defaults on exit, are very high and likely
Best regards,
David Burg.
-----Original Message-----
From: owner-mtfuji5 at [mailto:owner-mtfuji5 at] On
Behalf Of keiji_katata at
Sent: Monday, September 03, 2007 11:24 PM
To: mtfuji5 at
Cc: T10 Reflector
Subject: RE: Mt Fuji - MMC:
Hi David,
I think you may explain how it is used (usage), why it is necessary
and how it is significant (significance). I think "just for test purpose"
be a reason.
If "adjusted down by device according to media condition" is disabled, drive
will report check condition instead of read data. So the condition control
becomes host responsibility. You may remember that the original Toshiba
talked about only streaming read. So only a special software that can accept
this responsibility will use this control in the case of Toshiba proposal.
Best regards,
Keiji Katata
David Burg <daviburg at> on 2007/09/01
mtfuji5 at avc-pioneer.comに返信してください
送信者:     owner-mtfuji5 at
宛先:  "mtfuji5 at" <mtfuji5 at>
cc:    T10 Reflector <t10 at>
件名:  RE: Mt Fuji - MMC:
Bear with me, it's a long list. I believe "Problem statement" section of our
Read descriptors for Get Performance command last time already listed most of
4. Requirement from Microsoft
 [4-0 Toshiba's requirement, but generic for playback of any type of content
any type of media]
 4-1 Drive must provide read speed descriptors
	4-1-1 The list of read speed must be comprehensive
 4-2 Drive must provide per media-type speed descriptors for all media types
	4-2-1 All media types speed descriptors must be available at all
regardless of current media.
 4-3 Drive must accurately report/describe the different write rotation
[Microsoft assumes no other write rotation control exist currently]
 4-4 Drive must accurately report current read speed
 4-4 Host must be empower to control maximal rotation speed (noise and power
consumption control).
[Alternatively, host must be empower to directly control decibel level and
consumption, but I imagine this is too difficult for the drive vendors.]
 4-5 Drive's implementation correctness must be verifiable by host
[If drive reports a given speed, host must be able to issue I/O and get the
performance of this speed within specified tolerance.]
 4-6 Host must be empower to set read or write speed independently, and rules
temporarily adjust matching write or read speed when necessary must be
explicitly specified.
As non-technical recommendations:
	Unused fields should be market obsolete to simplify the command.
	A model section on speed control should be written.
Also, I thought Pioneer had the requirement that speed may be adjusted down
device according to media condition. Is this no longer a requirement from
Best regards,
David Burg.
-----Original Message-----
From: owner-mtfuji5 at [mailto:owner-mtfuji5 at] On
Behalf Of keiji_katata at
Sent: Tuesday, August 28, 2007 2:33 AM
To: mtfuji5 at
Cc: T10 Reflector
Subject: RE: Mt Fuji - MMC:
Hello David and all,
We need clear requirement for this issue.
1 Requirement from Toshiba
 Drive must keep the higher or equal playback speed for 3X HD content on DVD.
2. Requirement from Pioneer
 2-1 Drive must set its rotation speed for writing when writing occurs.
 2-2 Drive needs to know the end of data writing and beginning of 3X HD
3. Information of 3X HD content
 3X HD content may be recorded in fragments according to its allocation rule.
Could you make clear your requirement? And post it to reflector?
Best regards,
Keiji Katata
David Burg <daviburg at> on 2007/08/28
mtfuji5 at avc-pioneer.comに返信してください
送信者:     owner-mtfuji5 at
宛先:  "mtfuji5 at" <mtfuji5 at>
cc:    T10 Reflector <t10 at>
件名:  RE: Mt Fuji - MMC:
Hi Katata-san,
Thank you for the refreshed information on the Pioneer implementation.
I agree we need to work on the Set Streaming command so that its description
both better match what devices implement and better match what is reasonable
implement - like matching the rotation speed for both read and write when
commands are intermixed, yet leave the possibility of higher read rotation
when there is no writing.
We need to further work on probably both the get performance and the set
streaming command - possibly even the streaming feature itself - to address
HD Video on red laser media case brought to us by Toshiba (and in which
Microsoft has interest as well).
Microsoft also needs the devices to more accurately report their performance
capabilities and well as current state so that proper choices and information
can be presented to the user.
There is too a known need for noise control, and we know that some OEM in
particular is sensitive to the issue.
Finally there is a general push for power saving in the industry, and precise
speed control can help reduce the power consumption of optical devices.
Hey, this is probably one of the most complex and most interesting technical
challenges we had in Mt Fuji - MMC in a while.
David Burg.
-----Original Message-----
From: owner-mtfuji5 at [mailto:owner-mtfuji5 at] On
Behalf Of keiji_katata at
Sent: Monday, August 27, 2007 10:54 PM
To: mtfuji5 at
Cc: T10 Reflector
Subject: Re: Mt Fuji - MMC:
Hi David,
Yes, you are right. Live UDF now does not update Access time of the files.
DirectCD to Live UDF, I had confirmed and requested to stop the action. It is
confirmed that softwares stop it. There are no problem.
We need not discuss the hypothetical operating system but you may understand
that OS/FS spec allows the behavior.
I checked Pioneer implementation of the SetStreaming command. My knowledge
I said in the meeting was old and was not correct.
1. Pioneer drive may ignore the Read Seed field (Read Size/Time) (Notes: on
cases, drive refers the field.)
2. Pioneer drive refers the Write Speed filed (Write Size/Time) and sets the
rotation speed as lower or equal speed if possible. This is same as Toshiba.
3. When writing occurs or when SetStreaming command is received, drive sets
rotation speed according to the write speed. Always reading speed is set to
as writing speed.
4. In the case of CD/DVD-R/RW, BD-R, when a session is closed the reading
is set to maximum.
5. Drive remembers the last SetStreaming command then if next writing occurs,
drive performs the step no.3.
The reason why the reading speed is adjusted to same as writing speed is that
the data writing has the highest priority than any. Drive automatic verify or
host read operation for verify are performed by writing speed. Because all of
setting (spindle servo, pick up servo, equalizer, etc) are adjusted for the
writing speed. When rotation speed is changed unstable condition will
for some time length. This is not acceptable.
And buffered data writing and rotation speed change on a Read command may be
cause of time-out of the host.
I think we have two problems in the current SetSteaming command.
1. Drive may select lower or equal speed if possible on the current
command. This is not the next highest or equal.
2. When writing occurs drive may adjust its rotation speed for the data
not for the data reading.
Best regards,
Keiji Katata
David Burg <daviburg at> on 2007/08/10
mtfuji5 at avc-pioneer.comに返信してください
送信者:     owner-mtfuji5 at
宛先:  <keiji_katata at>
cc:    <mtfuji5 at>, T10 Reflector <t10 at>
件名:  Mt Fuji - MMC:
Dear Katata-san,
My understanding is that you claimed at this week’s Mt Fuji meeting that
Live file system will issue write commands to write-once and rewritable discs
even if the content is only played back, due to update of access time file
properties. Thus you felt the necessity for the Mt Fuji ‐ MMC standardized
devices to workaround a mix of read and writes even for playback
I have discussed with the development lead for Vista Live file system. He
confirmed that we not update the access time on the files. This is disabled
default. If you observe such issue, something else is probably going on,
creating desktop.ini files etc. Ravinder has not seen this for a long time
Vista Live file system seems pretty stable. Please provide particular
and steps taken to reproduce this if you still observe such issue so
team can investigate and fix the problem.
Accordingly, the standard definition shall not be changed and work on future
standard does not need to workaround an hypothetical operating system
implementation problem.
Best regards,
David Burg.
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at

More information about the T10 mailing list