Marantz DVD Driver

[jrlewis]

Theoretically this driver should support the DV4001, DV6001, DV7001, and VC6001. If it doesn't go up in a puff of smoke I'll be amazed. Sending commands to the device should be pretty solid, but the parsing of status replies may not be quite up to snuff. From the protocol it is unclear which commands should return an ACK. Get it up and running and then just step through each field and see if it responds ok or you get timeouts. Make a list of the fields that timeout on field writes and post them here. For other unexplained behavior just put the verbosity into high and grab the appropriate log sections.

The driver makes the assumption that only a power command is valid when the power is off. I didn't actually look at the operator's manual so this may not be the case. You would have to confirm with hyperterminal or the likes. If the user's can sus this out and get me the details it won't be hard to get the driver up to speed in no time. I have posted both a driver pack and manifest and driver file in plain text if it needs to be built on your system. Its pretty easy to do if you need to. No html document for now.

Posted Version 0.81
0.5 Added MenuCmds,Cursor,NumericKey,Increased poll time
0.6 Added MaxTimeouts, fixed bonehead mistake in polling
0.7 Changed several fields to not wait for response after field write,Fixed PlayMode string command
0.8 Fixed missing EndCase in Poll()
0.81 Added logging

[IVB]

Thanks dude, i'll run a test tonight. I believe the power assumption is correct, as when I did a @PWR:? it returned power-off.

[jrlewis]

For the power stuff what I meant is in the driver code it will reject any field write except for a power command when the power is off. This isn't always necessary as some fields can be changed/queried when a device is off. It just depends on the device. When I get some time I will take a look at the operator's manual.

I just saw you had posted about this in the Driver's Request. I haven't yet implement the menu/remote type commands. The framework is in place so it is pretty easy to add stuff like this as you go. Let me know how it goes. I know for sure some of the boolean fields are going to give you value rejected errors even though they actually work on the device.

[jrlewis]

Now that I think about it if the driver won't connect when you load it and the power is off turn the power on and it should connect. That would be an indication that it doesn't allow queries when the power is off. I forgot to worry about that as it was kind of a rush job.

[IVB]

It's alternating between WaitForConnect (most of the time), and WaitForCommResource(occasionally). I tried putting it into standby then un-standby to no avail.

I confirmed it's alive by connecting via HyperTerminal and issuing a @PWR:? command and getting a response.

Here's the logfile

05/27 21:21:33-CORE2, CQCServer, CQCDrv_MarantzDV6001_Thread18
{
CQCKit, CQCKit_DriverBase.cpp.2513, Status/App Status
Driver 'MarantzDV6001' has its comm resource
}
05/27 21:21:33-CORE2, CQCServer, CQCDrv_MarantzDV6001_Thread18
{
CQCKit, CQCKit_DriverBase.cpp.2531, Status/App Status
Driver 'MarantzDV6001' is trying to connect to its device
}
05/27 21:21:33-CORE2, CQCServer, CQCDrv_MarantzDV6001_Thread18
{
CQCKit, MEng.System.CQC.Runtime.CQCLogger.623, Status/App Status
Exception_MarantzDV6001.GetMsg.1078: Communication timed out
}
05/27 21:21:33-CORE2, CQCServer, CQCDrv_MarantzDV6001_Thread18
{
CQCKit, MEng.System.CQC.Runtime.CQCLogger.623, Status/App Status
Exception_MarantzDV6001.Connect.1078: Communication timed out
}
05/27 21:21:33-CORE2, CQCServer, CQCDrv_MarantzDV6001_Thread18
{
CQCKit, CQCKit_DriverBase.cpp.2586, Status/App Status
Driver 'MarantzDV6001' has lost its communcations resource
}
05/27 21:21:33-CORE2, CQCServer, CQCDrv_MarantzDV6001_Thread18
{
CQCKit, CQCKit_DriverBase.cpp.2474, Status/App Status
Driver 'MarantzDV6001' is trying to get its comm resource
}
05/27 21:21:33-CORE2, CQCServer, CQCDrv_MarantzDV6001_Thread18
{
CQCKit, CQCKit_DriverBase.cpp.2513, Status/App Status
Driver 'MarantzDV6001' has its comm resource
}
05/27 21:21:33-CORE2, CQCServer, CQCDrv_MarantzDV6001_Thread18
{
CQCKit, CQCKit_DriverBase.cpp.2531, Status/App Status
Driver 'MarantzDV6001' is trying to connect to its device
}
05/27 21:21:34-CORE2, CQCServer, CQCDrv_MarantzDV6001_Thread18
{
CQCKit, MEng.System.CQC.Runtime.CQCLogger.623, Status/App Status
Exception_MarantzDV6001.GetMsg.1078: Communication timed out
}

[jrlewis]

Is the power on or off?

[jrlewis]

I posted a version 0.2 that has some logging added to the method that does the initial querying of the device so it should let us know which field it's choking on.

[IVB]

Tried both in power-on-but-standby and power-on.


05/27 22:20:51-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread19
{
CQCKit, MEng.System.CQC.Runtime.CQCLogger.624, Status/App Status
Exception_MarantzDVD.GetMsg.1079: Communication timed out
}
05/27 22:20:51-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread19
{
CQCKit, MEng.System.CQC.Runtime.CQCLogger.624, Status/App Status
Msg_MarantzDVD.QueryAll timed out for AEX:?.
}
05/27 22:20:51-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread19
{
CQCKit, MEng.System.CQC.Runtime.CQCLogger.624, Status/App Status
Exception_MarantzDVD.Connect.1079: Communication timed out
}
05/27 22:20:51-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread19
{
CQCKit, CQCKit_DriverBase.cpp.2586, Status/App Status
Driver 'MarantzDVD' has lost its communcations resource
}
05/27 22:20:51-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread19
{
CQCKit, CQCKit_DriverBase.cpp.2474, Status/App Status
Driver 'MarantzDVD' is trying to get its comm resource
}
05/27 22:20:51-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread19
{
CQCKit, CQCKit_DriverBase.cpp.2513, Status/App Status
Driver 'MarantzDVD' has its comm resource
}
05/27 22:20:51-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread19
{
CQCKit, CQCKit_DriverBase.cpp.2531, Status/App Status
Driver 'MarantzDVD' is trying to connect to its device
}
05/27 22:20:51-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread19
{
CQCKit, MEng.System.CQC.Runtime.CQCLogger.624, Status/App Status
Exception_MarantzDVD.GetMsg.1079: Communication timed out
}
05/27 22:20:51-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread19
{
CQCKit, MEng.System.CQC.Runtime.CQCLogger.624, Status/App Status
Msg_MarantzDVD.QueryAll timed out for AEX:?.
}

Looks like the AEX command

[jrlewis]

That is the first one. Can you send an AEX:? to the unit with hyperterminal to see what reposne you get.

[IVB]

I send @AEX:? and I get a @AEX:1 back.

Did you send the @ sign in the beginning of each command? I missed that in the doc too, it's really written very badly.

[jrlewis]

I do prepend an @ to every command and append a carriage return 0x0D. Let me take a look and see if there is something obvious I"m missing.

[jrlewis]

About how long is it taking it to respond?

[IVB]

it's pretty quick, def <1s, dunno if it's really <500ms as the protocol says.

[jrlewis]

I posted a new version. I changed the way the driver was reading the incoming buffer. It should work just fine the way it was, but let's see if this works.

[IVB]

Same deal. Here's the logfile. BTW, if you want, i can open up the cqc remote port server so you can muck with it directly.


05/27 23:08:49-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread20
{
CQCKit, MEng.System.CQC.Runtime.CQCLogger.625, Status/App Status
Exception_MarantzDVD.GetMsg.1082: Communication timed out
}
05/27 23:08:49-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread20
{
CQCKit, MEng.System.CQC.Runtime.CQCLogger.625, Status/App Status
Msg_MarantzDVD.QueryAll timed out for AEX:?.
}
05/27 23:08:49-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread20
{
CQCKit, MEng.System.CQC.Runtime.CQCLogger.625, Status/App Status
Exception_MarantzDVD.Connect.1082: Communication timed out
}
05/27 23:08:49-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread20
{
CQCKit, CQCKit_DriverBase.cpp.2586, Status/App Status
Driver 'MarantzDVD' has lost its communcations resource
}
05/27 23:08:49-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread20
{
CQCKit, CQCKit_DriverBase.cpp.2474, Status/App Status
Driver 'MarantzDVD' is trying to get its comm resource
}
05/27 23:08:49-CORE2, CQCServer, CQCDrv_MarantzDVD_Thread20
{
CQCKit, CQCKit_DriverBase.cpp.2513, Status/App Status
Driver 'MarantzDVD' has its comm resource
}

[jrlewis]

That would be great. I looked to see if I was getting something wrong in the port configuration or something, but it doesn't look like it. I know I'm just missing something simple. PM me with the details for the remote port server...

[jrlewis]

The reason for the timeouts was a dumb mistake...I was using the whole query value as a check for a reply (e.g.,PWR:?) instead of just the prefix. That and I had a few other glitches here and there. It still needs a little work. Without being able to interact with a menu on a DVD or an audio disc the PlayMode and what I'm guessing are fields dependent on the PlayMode need to be explored. I was seeing some very large response times (>5 secs) for some changes so there is a TimeOut field that defaults to and maxes out at 10000 ms .You can use this for now to dial in a reasonable timeout. I'm not sure if it is a network lag issue or an actual device latency issue. For the PlayMode I don't wait for a response and just pick it up in the poll since it can take a long time to respond.

One real oddity is that if you shut the power off via the driver it quits responding and just goes away. You can't query it for its power status and it just dies. I played around with adding a powerup in the connect after a case where it dies and it behaves very sluggish as it is coming back online with several timeouts even with a 10s timeout. Funny thing is that it didn't matter if it would try to start up right away or I let it sit for a minute to settle down. Either way it isn't happy when it wakes back up. I have also seen several messages that aren't listed in the protocol that I assume are level 3 & 4 type stuff. I downloaded the crestron example files from the marantz website and in there you can see the extra commands, several I haven't seen, and they don't list one I have seen which is TRK which I assume to be track. This one is a funny egg. Let me know how it goes...

[IVB]

duuuude, bada bing! It works! I've only tried PlayMode And GoTo Track(Next/prev) so far, but it works great.

Thanks!

[jrlewis]

I know you just want the basics, but it would help for driver development if you give this a thorough thrashing. Try different types of discs and determine what fields work with what and I can put interlocks on those based on the disc type. I would rather reject a field write straight up rather than just sending the command and waiting for the timeout. Also play with the timeout field. Set it to something ridiculously low like 100 and then crank it up until you no longer experience timeouts on a field you know shouldn't.

On the side when you get a chance try using hyperterminal to check on some undocumented commands. You will have to query them and see what comes up and do a little sleuthing to see if they are what I think they are.

TRK:? - I think it is the current track number
PGR:? - May be progressive scan. Try turning off and on PGR:1 or PGR:2 probably.
REZ:? - Resolution I believe. Probably REZ:1 - REZ:6 480p,576p,720p,1080i,1080p,Auto REZ:0 probably toggles.
TMN:? - Time?
VRL:? - ???

[IVB]

sure, no problem. I'll need a few days.

[Dean Roddey]

You should just set up the remote port server so that he can work directly against the box. He could get it done very quickly and be sure it's really working right.

[jrlewis]

I'm pretty sure all the commands and what not are now correct, but there are some things he just needs to test directly. The PlayMode means different things depending on the type of disc, menu structure, where in the menu you are, etc. Me pushing stuff on this end doesn't do much good as it can appear things aren't working properly when they are actually working just as they should. There appears to be a lot of other functions that are disc-type dependent as well that I can't test remotely. Same goes for the undocumented stuff. He needs to physically be looking at the device or ouput from it on a display to determine what those may represent.

[IVB]

Can you add support for the following commands? Just started to bang on it, realized I can't actually move through the DVD menus without some of these:

Write Commands:
Enter
Cursor (up/down/left/right)
Return
Menu
Top Menu

[jrlewis]

Sure thing. Probably the best grouping would be to have Cursor movement as its own field and then have Enter,Return,Menu, and Top Menu in another field. I will also add the numeric keys. I just need to uncomment a couple of lines for those. Give me 20 minutes or so.

Edit: Done.

[IVB]

Thanks for that lightning fast update. One glitch now: The driver is frequently losing it's connection. In verbose mode, the log says:

Exception_MarantzDVD.GetMsg.1159: Communication timed out.
Exception_MarantzDVD.Poll.1159: Communication timed out.

I got the "menu" part of the menu command & cursor command tested, but am having difficulty with "enter" due to that. I'll keep plugging away to see if it's my side or the driver.

[jrlewis]

The only thing being polled is the power and the timeout is set way high so likely it is being polled faster than it likes. I didn't bump the version, but download the one posted. I set the poll time to 1 second.

Edit: Having time to think it was already polling at approximately 1 second. The way it works is it looks for asynch messages and if it sees any it clears a counter, otherwise it bumps it. Once the counter hits a predetermined value (in this case 5 at 200ms polling = 1 second) it will poll the power just to make sure it is still there. It may just not like to respond at certain times and the messages you were sending provide no reply so we may need to go the route of being a little forgiving on timeouts and set a limit on the number of successive timeouts that can be seen before the driver times out. I will put out a new version this evening.

[jrlewis]

How frequently is it losing the connection? How much time do you see between pairs of timeout messages?

[IVB]

used the latest version, loses it's connection every 14-30seconds.

[jrlewis]

At the current setting that is every 3-6 polling cycles of the power. Let me add a little forgiveness on the timeouts.

Edit: When it loses its connection and reconnects do you see any other timeouts being thrown?

[IVB]

nope, no other timeouts. Is it just polling the power? For testing purposes, what if we drop that to once per minute?

[jrlewis]

I'm almost embarrased to let y'all know what the problem was. You can't expect a reply if you don't tell it you want one. I wasn't sending a query message to the device so of course it wasn't going to respond...ahhhhhh. I know it was there originally, but I am going to go back and look at the early versions just to make sure. I still added a MaxTimeouts during polling for now. The only place you may see it timeout is when it takes too long to responsd from a field write. If you run into those situations let me know.

[IVB]

well, it's been loaded for 5mins with no timeouts. I'll try using it later tonight when the kids aren't whining or fighting or bickering or complaining. (ie, after they go to sleep)

[jrlewis]

In between 0.4 and 0.5 I managed to whack the power query command. Don't know how, but I did. You shouldn't see any timeouts from polling, only possibly from field writes where the device is slow to respond.

[klindy]

well, it's been loaded for 5mins with no timeouts. I'll try using it later tonight when the kids aren't whining or fighting or bickering or complaining. (ie, after they go to sleep)

They didn't do any of that when I was there!! :-)

[IVB]

we were in the basement...

[IVB]

some minor testing so far. There have been 8 timeouts, but zero lost connections. And MY GOD is this unit fast to respond, but i'll post about that and some other thoughts in the "HA progress" thread later tonight.

[jrlewis]

Are the timeouts on field writes and they are giving you a rejected value? If so crank up the verbosity and it will log which fields are timing out. If it is giving a rejected value is the action still happening on the device even though the reply from the device times out?

[IVB]

well, the # hasn't changed since I posted. I just put into high verbosity, i'll tell you if anything happens.

[bph]

some minor testing so far.
Does that involve testing the "bickering children"?

[IVB]

it's odd; some commands seem to work fine, but get "value rejected by driver". I'll show which. In looking at the pattern, it looks like the device can't handle messages in rapid sequence, but i'll let the techie folks provide guidance there.

Cursor works fine
GoToTrack works fine
MenuCmd.Menu works fine but throws the "value rejected" error
SubTitleToggle works fine, but if I go too quickly it will throw "value rejected"
Zoom works fine, but if I issue multiples too quickly it will throw "value rejected"

Fwd & Rewind commands are hinky:
Playmode.FastFWD is currently Fast Rewind
Playmode.FastRev does nothing
ResumeStop throws an error claiming "value rejected by driver"
SlowFwd is currently slow rewind
SlowRev is currently fast forward

Aspect works, but most of the time it throws an erorr claiming "value rejected by driver"

More later, i'll have to read up on what the other commands actually do.

Does that involve testing the "bickering children"?

oh, i was clearly the one being tested there.

[jrlewis]

I noticed when I was just pushing buttons on my end some commands take a long time to respond at times. That being the case I have just set those fields to not wait for a response to confirm a field write and the response should get picked up in the polling. I'm surprised the SubtitleToggle was rejected as that is already one of those fields that just sends a command and moves on. All the toggle fields only accept a True value and they are set to AlwaysWrite.

For the PlayMode I was looking at the responses as opposed to the commands in the protocol. I'm hoping that the protocol isn't correct with respect to send/reply messages and the same command you send is returned when its in that state. If the commands you send work fine, but the reply puts the PlayMode in a different state I will need to make a modification to the driver to handle that. The ResumeStop is just a return value so you can't set it. If you try it will reject it. I guess I could just let it send it as I don't wait for a response for the PlayMode so it shouldn't hurt anything. Version 0.7 is up for you to play with.

[IVB]

See http://www.charmedquark.com/vb_forum/showthread.php?p=86573#post86573 . it didn't work after the upgrade to 2.3.28, so i uploaded 0.7 and it still sits there on "Wait for Connect".

i'll assume this is a CQC-side thing since it worked fine before the upgrade. If you're not on .28 yet, can you hold off upgrading? I can open up the remote port, you can confirm you connect, then I can check via my system.

[jrlewis]

Open it up and I'll take a look...

[IVB]

Open it up and I'll take a look...
dsl has been down for almost an hour. (this is via my cellphone). I will ping here when its back up.

[jrlewis]

In the meantime put the verbosity on high and see if you can see anything in the logs. Look for Exception_Moniker.Method.#: Description. If you see errors in the GetCommResource method with a description "Could not open /Local/Com11" something has the com.

[IVB]

internet is back up, i just opened up rportsrv, i'll PM you the WAN IP.

It's definitely CQC or the driver as I could hyperterminal in just fine. It says "Wait for COMM resource", but that ain't true due to the HT working fine. HT is now down, and the driver is gone, so you should be set.

[jrlewis]

I can connect to the device on my end. For some reason the query for the Tray was timing out, but other than that it looks like business as usual. Do you use any type of messenger or yahoo chat? If so pm me the details so we can talk while I try a couple of things.

[ismewor]

Hi Jr,
Does your driver work with DV7600?

Let me know
Thanks,
A

[jrlewis]

The protocol is very similar, but there are several changes/additions for that model that make it incompatible with the current driver. I can't do it now, but give me a couple of weeks and I will add support for it.

[ismewor]

Just want to check out would this work with DV7600 as well?

Thanks
ismewor

[jrlewis]

The DV7600 would not work with this driver. This driver would need to be updated to accomodate it or probably the better option would be to build a separate driver. The protocol for some of the device fields for the DV7600 are handled much differently than the devices that this driver currently supports. It wouldn't be too hard to mod this one to work for the 7600. Do you have one?

[ismewor]

Thanks for the response, yes, i had 2 of this 7600 currently. it would be nice to be able to control. if you need me to test out or try out feel free.
Thanks in advance

Thanks,
ismewor

[IVB]

i just upgraded to v2.4.14, this sits at "wait for initialization". Was working fine before the upgrade.

I know dean made some changes, dunno if any of them would have impacted this.

[Dean Roddey]

It was missing an EndCase which isn't allowed now. I've fixed it. The new one is attached.

[IVB]

It was missing an EndCase which isn't allowed now. I've fixed it. The new one is attached.
ah, so it has. Thanks!

[jrlewis]

Sure enough...in the Poll(). Other than that have you had any troubles with the driver? Its been awfully quiet so I assume all is well. And actually I changed the logic at some point to only query the power since it supplies async's for everything else so a switch really isn't even needed there. I will edit it and load the new one up at the start of the thread.

[IVB]

It's been working great. Then again, all I ever do is stop/pause/play/up/down/left/right/next chapter/prev chapter. Hell, I don't even like doing a FFW/REW, be it on live DVDs or rips :-)

If it's not in the official list, it should be. (did you ever remove that hardcoded COM11? I never tried testing as i'm out of IP Serial Server ports, so I need it to be on COM11)

[jrlewis]

The last version posted isn't hard coded. I just did that to get past the bump we had with one of the betas. I will get the new one up and get some documentation for this done and then once you have a chance to bang on it some more it should be ready to go. Eventually I need to create a companion to this for ismeor. If it doesn't fit I may need to change the Model for this so I it reflects the devices it supports. It is fairly easy to do a Model change on a driver. Just create a template pack for the templates that use the driver, import and install the new driver, and then reimport the templates and since it has the same moniker/fldnames it should pick it up if memory serves me correctly.

[IVB]

Is there any reason this wouldn't work on 3.1? I haven't used this device in a year, just noticed the driver wasn't hooked up, but now i'd like to integrate it to CQC again. But it cycles between 'Wait for comm resource' and 'wait for connect'.

This could easily be user error as i'm going via an IP Serial Server on a port I haven't used in years, i'm going to try and hunt down a normal serial cable and test via laptop, but that's a lot easier said than found.

[jrlewis]

Surprised it compiles at all if using 0.7. It had the inadvertent issue with a missing EndCase statement in the Poll routine. Try the version 0.8 posted.

[IVB]

Surprised it compiles at all if using 0.7. It had the inadvertent issue with a missing EndCase statement in the Poll routine. Try the version 0.8 posted.

yeah, tried that version, same issue.

[jrlewis]

You know the drill then. Crank up the verbosity and see what comes up in the log. With that type of cycling it isn't making it through the connect.

[IVB]

The salient bit is below. Not terribly revealing. I'm still trying to find a spare regular serial cable so I can hook it up to my laptop & see if it works without all the crap in the middle.

http://www.myhometheaterpc.com/temp/marantz_dvd_error.jpg

[LesAuber]

IVB, Cheap shot. If you haven't used for a while check and see if power save (or similar) is set to on in the set up menu for the player. Denon added this little feature but don't know if Marantz followed.

Power save turns off the serial port leaving the thing dead to the world when put into standby. It's pretty well buried in my 3800. On is the default condition if the player was reset for any reason.

[IVB]

I used a regular serial cable to no avail, then realized it's definitely the driver. If I don't even hook up a serial cable to the com port, it does the same thing (cycle between those two). IE, in theory either of you could try adding a Marantz DVD-DV6001 driver and you'd see the same situation.

[jrlewis]

It doing the same thing without hooking anything up doesn't tell you anything. With a serial connection with no flow control it will open it up and write to it all day long and it doesn't care if anything is on the other end or not. I posted 0.81 with some logging that should tell you how far it is getting along in connecting to the device. I think you have an issue somewhere in the configuration of the player since nothing in the driver or cqc would have changed to cause it to no longer connect.

[IVB]

Thanks. I just tried that via the IP serial server, here's what it says. I'll try with the laptop & serial cable in a little bit (making dinner for the kids).

http://www.myhometheaterpc.com/temp/marantz_dvd_error.jpg

[IVB]

Oh Good Lord!

Not sure how/why, but it appears that turning it on (it was in "sleep/standby" mode) makes this connect. I just put it back to sleep, then was able to use the "Power" command to bring this out of sleep mode.

Eh, sorry about that, thanks for creating that 0.81 version (and for the driver itself).

[jrlewis]

There must be a hard "off" as opposed to sleep mode so it was dead to rights. It should still respond to a power query when in standby, but that is the only thing it will respond to. When connecting the driver queries power state of the device and then turns it on if it is off to get all the associated field values instead of creating a bunch of "Unknown" values for quite a few string fields.

[IVB]

Well, it was in a standby mode, i just pressed the wakeup button on the front of the unit.

Regardless, it's back in standby, and still connected, so if anyone else reads this with one of these in the future, make sure to try that simple test first.

[jrlewis]

It must have been in a bad mood. Hammer on this if you would to see if everything works or you can get it to break. If it looks ok I can give it a little polish as I haven't looked at it for quite some time and we can get it into the build. I have a large back log of drivers that will be bubbling up here finally so want to clear as much stuff out as possible.