MP4 files collection (Test CD) for testing "NeroDigital" players + results

[Videostation]

Hi!

Here are my results with the Siemssen and firmware from 05.02. They are are a little bit different than from telemachus. The differences are red marked.

audio folder

audio/mpeg-1 sub-folder
- video ok, no audio

audio/mpeg1/mp3 sub-folder
- video ok, no audio

audio/mpeg-4 folder
- video ok, but audio without SBR infos

audio/mpeg-4/multichannel sub-folder
- plays very well, but it shows only 2 channels in the infos
- I can’t test the mapping, because I haven’t a 5.1 system

------------------------------------------

special folder

special/pictures folder
- shows: This profile is not supported

special/private folder
alac_itunes.m4a – player freeze
theora_mp4box.mp4 – shows: This profile is not supported
vorbis_mp4box.mp4 – jumps to the main menu

special/subtitles folder
timedtext_mp4box.mp4 – video ok, no subs
vobsub_sp-3ivx_3ivx.mp4 – video + subs ok, showing also the correct language via infos
vobsub_sp-3ivx_mp4ui.mp4 – video + subs ok, showing also the correct language via infos

special/systems folder
kangaroo.mp4 - shows: This profile is not supported

special/xcd folder
sp_3ivx_mp4box.mp4.dat – black picture

------------------------------------------

video folder

asp-adapt_nero_mp4box.mp4– black picture
asp-custom_xvid_mp4box.mp4 - shows: This profile is not supported, if I use a custom matrix in an AVI it works perfectly
asp-gmc-1wp_divx_mp4box.mp4 - OK
asp-gmc-3wp_xvid_mp4box.mp4 – it works, but artifacts are hidden because of the high compression, with my own material I have clearly visible artifacts in the picture
asp-mpeg_xvid_mp4box.mp4 - OK
asp-qpel_divx_mp4box.mp4 - OK
sp_3ivx_mp4box.mp4 - OK

video/anamorphic sub-folder
- is shown, but consider no PAR

video/avc sub-folder
avc-base_nero.mp4 – doesn’t play

video/b-frames sub-folder
- plays all very well, except „asp-1b-pb_divx_mp4ui.mp4” -> jerky playback

video/chapters sub-folder
- no problems with navigation

video/mpeg1-2 sub-folder
- player freeze

CU Videostation

[bond]

cool, thx a lot :)

[SeeMoreDigital]

Yes, many thanks Videostation,

I wonder why Mpeg4 encodes with anamorphic signalling don't work correctly?

Maybe some more tests files could be generated with DAR and PAR signalling, in both .AVI and .MP4 containers?


Cheers

[Videostation]

I've already tested AVIs (XviD) with different PARs, but the result is always the same. The player doesn't consider the PAR. But in the next days I will test a little bit more.

CU Vidostation

[bond]

Originally posted by Videostation
asp-custom_xvid_mp4box.mp4 - shows: This profile is not supported, if I use a custom matrix in an AVI it works perfectlyto my knowledge the matrix used in this test was hvs-best
did you try other matrices in .mp4, eg encoded not with xvid, but with nero for example?

[Videostation]

Yes, I've tested also other matrices in .mp4, but neither Soulhunter's v6 nor Sharktooths EQM V3LR works.
Didn't try Nero yet, maybe in the next days. But afaik it should work.

CU Videostation

[bond]

Originally posted by Videostation
Didn't try Nero yet, maybe in the next days.would be great

But afaik it should worki would be very surprised if it would work

[Videostation]

i would be very surprised if it would work Be surprised. :D
I can confirm that custom matrices work within .mp4, if I use Recode. The files playing perfectly on my Siemssen. I used Sharktooth EQM V3LR and the hvs-best matrix for my tests.

CU Videostation

[bond]

can you try to demux these nero streams to raw and than reimport to .mp4 with both mp4box and mp4creator and try if they work similar to what is discussed here (http://forum.doom9.org/showthread.php?s=&threadid=88568&perpage=20&pagenumber=6)

that way we should be able to check what really causes the problem (xvid, mp4box, nero, siemmsen...)

[SeeMoreDigital]

Originally posted by bond
...that way we should be able to check what really causes the problem (xvid, mp4box, nero, siemmsen...) Good idea!

Can you start with using MP4BoxGUI v0.14 please... Which is available here (http://forum.doom9.org/showthread.php?s=&postid=617798#post617798) - but marked as v0.13!


EDIT: You will need Micro$oft .NET and the latest version of MP4Box.exe installed.


Cheers

[Videostation]

Ok, I'm going to do it and post my results later.

[bond]

if possible dont use mp4boxgui or mp4muxer to avoid problems introduced by the gui (dunno if there are some, but who knows)

[SeeMoreDigital]

Originally posted by bond
if possible dont use mp4boxgui or mp4muxer to avoid problems introduced by the gui (dunno if there are some, but who knows) In that case it may be prudent to try muxing, with and without a GUI.

One things for sure, .MP4 still needs some accurate muxing tools.

The sooner we establish that muxing tools such MP4BoxGUI are able to generate encodes that work with the Siemssen (and Rjtech RJ-1500DVX II) the better.... So you might want to try Mpeg4 streams generated without MPEG matricies as well. And maybe different PAR/DAR values too!

Just a thought...


Cheers

[bond]

mp4boxgui uses mp4box.exe directly, so it should produce exactly the same output
its just that you might never know whether there is a bug in mp4boxgui which changes something, so why risk anything

One things for sure, .MP4 still needs some accurate muxing tools. whats inaccurate with mp4box and mp4creator?
till now we dont know whether the custom matrix problem is caused by them or siemmssen

[SeeMoreDigital]

Originally posted by bond
whats inaccurate with mp4box and mp4creator?Nothing as far as I know.... It's just that not everybody likes doing everything via command line!

Originally posted by bond
till now we dont know whether the custom matrix problem is caused by them or siemmssen It's a curious one for sure... especially as the custom matrix streams appear to work in .AVI

I've tried many custom matrixed .MP4 files (muxed using mp4UI). Most of which worked with my Xcard.

I wonder whether somebody can provide some small sample files for me to try?


Cheers

[Videostation]

I've tried mp4creator (1.2.12) as well as mp4box (2.4.0) with the demuxed nero-avi and it almost worked. But the problem is a jerky playback.

mp4creator:
-----------
avi2raw -> mp4creator -create=nero.avi -H -rate=25 nero.mp4 -> mp4creator -create=audio.aac -H -I -rate=25 -optimize nero.mp4

mp4box:
-------
mp4box -convert nero.avi nero.mp4 -> mp4box -add audio.mp4#trackID=1 nero.mp4

First i demuxed the recode mp4-file with graphedit and 3ivx splitter.
I hope the steps are right?

CU Videostation

[SeeMoreDigital]

Hi Videostation,

Can you confirm whether the de-muxed .AVI streams play okay?

By-the-way... do any of them contain B-VOP's, GMC or Qpel?


Cheers

[Videostation]

Can you confirm whether the de-muxed .AVI streams play okay? Yes, I can. No problems if I mux the nero .AVI with MP3.
By-the-way... do any of them contain B-VOP's, GMC or Qpel? I used B-VOP's and QPel, but no GMC.

[SeeMoreDigital]

Originally posted by Videostation
Yes, I can. No problems if I mux the nero .AVI with MP3.
I used B-VOP's and QPel, but no GMC. Did you remember to remove/disable packed bit-stream?

Have you considered generating some "custom" tests without B-VOP, Qpel and GMC?


Cheers

[bond]

mp4box handles packed bitstream .avi files right
mp4creator doesnt (you have to do a .avi -> raw -> .mp4 conversion if you want to use mp4creator with packed bitstreams)

but this issue (packed bitstream) is irrelevant here as a nero stream extracted from .mp4 is surely not packed

still its very strange that both muxers show exactly the same result (choppy playback), i can only imagine that its siemmssens fault, cause that would be a too big coincidence that two independant muxers are borked the same way

you might want to try to demux the nero stream from .mp4 to raw and than reimport the raw stream (and not the .avi) to .mp4 with mp4creator and mp4box just to be sure
and also dont add the audio stream in your tests, so you can be sure that the choppyness isnt somehow caused by an audio issue

[SeeMoreDigital]

Sorry... I scanned passed the bit about the custom MPEG encodes being Nero... so no, they wont contain packed bit-stream!

And yes, it's a good idea not to mux the audio stream into the .MP4 container at this stage.


Cheers

[Videostation]

There is another problem at the momemt. If I load the demuxed nero .AVI with MPEG4Modifier the quant type is either MPEG or h.263. But never MPEG custom, how it should be normally. I don't know why?
Does another program exist, that is able to read out the matrices?

CU Videostation

[SeeMoreDigital]

I've never come across that with any of the NeroDigital encodes I've generated.

What are your settings? Maybe I can replicate it myself with a small source and provide a sample for you to try!


Cheers

[Videostation]

settings: I've used the profile maximum definition, but with QPel, B-VOP's, 4 vectors per macroblock and hvs-best-picture matrix.

[bond]

btw the vol created by mp4box on custom matrix files (the place where the matrix info is stored btw) looks pretty normally:

Video Object Layer
random_accessible_vol - 0
video object type - 17
is_object_layer_identifier - 1
verid - 2
vol_priority - 4
aspect_ratio_info 1
vol_control_parameters - 1
chroma format - 1
low delay - 0
vbv_parameters - 0
video object layer shape - 0
vop_time_increment_resolution - 25
fixed_vop_rate - 1
fixed_vop_time_increment - bits 5 1
vol width - 640
vol height - 256
interlaced - 0
obmc_disable - 1
sprite - 2
no_of_sprite_warping_points - 3
sprite_warping_accuracy - 3
sprite brightness change - 0
not_8_bit - 0
quant_type - 1
load_intra_quant_mat - 1
8 16 16 16 16 16 16 16 16 16 17 16 17 16 17 18
17 18 18 17 18 21 19 20 21 20 19 21 24 22 22 24
24 22 22 24 25 25 27 30 27 25 25 29 31 35 35 31
29 36 41 44 41 36 47 54 54 47 65 70 65 88 88 115

load_nonintra_quant_mat - 1
18 18 18 18 18 18 18 18 18 18 19 18 19 18 19 21
19 20 20 19 21 23 21 22 24 22 21 23 27 24 24 27
27 24 24 27 29 28 30 33 30 28 29 32 35 39 39 35
32 40 46 50 46 40 53 61 61 53 73 79 73 98 98 129

quarter sample - 1
complexity estimation - 1
resync_marker_disable - 1
data partitioned - 0
newpred_enable - 0
reduced_resolution_vop_enable - 0
scalability - 0

[bond]

Originally posted by Videostation
There is another problem at the momemt. If I load the demuxed nero .AVI with MPEG4Modifier the quant type is either MPEG or h.263. But never MPEG custom, how it should be normally. I don't know why?ok i demuxed a .mp4 file containing a custom matrix xvid stream with mp4box to .avi and fed this .avi to mpeg4modifier
and mpeg4modifier detects the custom matrix perfectly!

maybe the nero streams have a bug and the siemmssen only works together with these buggy nero streams correctly?
still that would also be very strange

[Videostation]

Yes, if I use a xvid stream there is no problem. The matrix is showed correctly. Only the nero stream makes problems. I don't know, but maybe no custom matrix is used with recode. That could be the bug.
btw how can i get the vol infos with mp4box?

[bond]

Originally posted by Videostation
btw how can i get the vol infos with mp4box?use the mpeg4vol tool from mpeg4ip for that

[Videostation]

Here is the vol created by nero directly. hvs-best-matrix was used, but no infos in the vol about that. No matter what custom matrix I've used, the result is every time the same.

Video Object Layer
random_accessible_vol - 0
video object type - 17
is_object_layer_identifier - 1
verid - 5
vol_priority - 1
aspect_ratio_info 1
vol_control_parameters - 1
chroma format - 1
low delay - 0
vbv_parameters - 1
bit rate 2403
vbv buffer size 587
vbv occupancy 49589
video object layer shape - 0
vop_time_increment_resolution - 25000
fixed_vop_rate - 1
fixed_vop_time_increment - bits 15 1000
vol width - 672
vol height - 272
interlaced - 0
obmc_disable - 1
sprite - 0
not_8_bit - 0
quant_type - 1
load_intra_quant_mat - 0
load_nonintra_quant_mat - 0
quarter sample - 1
complexity estimation - 1
resync_marker_disable - 1
data partitioned - 0
newpred_enable - 0
reduced_resolution_vop_enable - 0
scalability - 0

[SeeMoreDigital]

Just finished generating an custom HVS-Best encode with Recode2 to .MP4. It would appear that the matrices have been inputted: -

http://img26.exs.cx/img26/4453/recode2cmsettings3qg.png

However, after de-muxing the stream to .AVI and running it thru' MPEG Modifier, I can confirm that the matrices are not visible: -

http://img29.exs.cx/img29/6903/ndhvsbestdemuxedtoavi2kk.png


I also generated an custom HVS-Best encode with XviD to .AVI. After running the encode thru' MPEG Modifier, I can confirm that the matrices are visible: -

http://img224.exs.cx/img224/6120/xvidhvsbestinavi3dp.png

I then used MP4BoxGUI (v0.14) to mux the custom stream into .MP4

I am able to confirm that I was able to play all of the encodes perfectly using my Sigma Xcard and ADS Media-Link...

I guess I can provide the samples if required.


Cheers

[Videostation]

Thanks for the confirmation.
The problem is, nero seems to use the last settings from "Advanced", before the option "force custom matrix" was activated. If I had chosen h.263, the .mp4 file contained the h.263 matrix, with mpeg the standard mpeg matrix was used, but never the wanted custom matrix.

CU Videostation

[bond]

Originally posted by Videostation
Here is the vol created by nero directly. hvs-best-matrix was used, but no infos in the vol about that. No matter what custom matrix I've used, the result is every time the same.

Video Object Layer
random_accessible_vol - 0
video object type - 17
is_object_layer_identifier - 1
verid - 5
vol_priority - 1
aspect_ratio_info 1
vol_control_parameters - 1
chroma format - 1
low delay - 0
vbv_parameters - 1
bit rate 2403
vbv buffer size 587
vbv occupancy 49589
video object layer shape - 0
vop_time_increment_resolution - 25000
fixed_vop_rate - 1
fixed_vop_time_increment - bits 15 1000
vol width - 672
vol height - 272
interlaced - 0
obmc_disable - 1
sprite - 0
not_8_bit - 0
quant_type - 1
load_intra_quant_mat - 0
load_nonintra_quant_mat - 0
quarter sample - 1
complexity estimation - 1
resync_marker_disable - 1
data partitioned - 0
newpred_enable - 0
reduced_resolution_vop_enable - 0
scalability - 0
this is borked!

or an other possibility is that the nero encodes, altough you set it in the gui, dont use the custom matrix, but mpeg instead. because according to smd's screenshots you also set "mpeg quantisation" in the advanced tab and not something like "custom quant". this would be a gui/encoder bug

another possibility would be that you guys set a nero profile in the gui, which doesnt allow custom matrices, and therefore resets to the normal mpeg one
are you guys sure you set nero to the unrestricted profile?

[SeeMoreDigital]

Originally posted by bond
...are you guys sure you set nero to the unrestricted profile? Here's the "Profile" I've been using: -

http://img220.exs.cx/img220/6039/recode2profiles3dw.png


And here's where Recode2 informs you about "Custom Matrices": -

http://img16.exs.cx/img16/2295/recode2customsetting4om.png


Cheers

[bond]

Originally posted by SeeMoreDigital
Here's the "Profile" I've been using:ok, so it pretty sure that either the nero gui is borked and doesnt encode using the set the custom matrix or the nero encoder is borked as it doesnt signal the matrix (correctly)

either way, its neros fault :D

[SeeMoreDigital]

I think it's probably b0rked :(


Cheers

[Videostation]

Thx for the nice screenshots. :)
I think it's a bug in the gui.

[SeeMoreDigital]

Who's going to bring this observation to the attention of JohnV?


Cheers

[Videostation]

You. :D

[pecus]

Hi all, I was interested in testing my brand new packard bell dvx350 with the cd test suite, but the link is broken. Maybe bond can rearrange with telemachus or find alternative hosting. Or it's just a wrong path. In any case, I'm looking forward to seeing it available again.

[bond]

if someone can and wants to host the files just gimme a shout :)

[bond]

i have created a new version with lots of new mp4 files

the downloadlink is in the first post

[SeeMoreDigital]

i have created a new version with lots of new mp4 files

the downloadlink is in the first postBummer... you've hosted the file with RapidShare!

I would purchase their "Premium" service in order to download your file.... and I don't trust the company at all :eek:

[bond]

why that

[pecus]

@bond: tx for the file, I'll post results.

@SMD: Rapidshare allows you to download the file freely. Try a couple of times if it doesn't work. I'll be sharing it on eDonkey/eMule starting tonight (try Razorback).

[pecus]

@bond: despite the readme, the archive does not contain a burning image, rather files and folders. Is this right? (I used winrar to unpack the 7z archive).

[SeeMoreDigital]

@SMD: Rapidshare allows you to download the file freely. Try a couple of times if it doesn't work. I'll be sharing it on eDonkey/eMule starting tonight (try Razorback).Sadly it does not matter how many times I try or at what time of the day..... I always obtain a meesage similar to this one: -

http://img245.imageshack.us/img245/3742/rapidsharemessage1ri.gif


Cheers

[bond]

@bond: despite the readme, the archive does not contain a burning image, rather files and folders. Is this right? (I used winrar to unpack the 7z archive).yeah, i forgot to update the readme

[masken]

There really should be an overviewable online form or something on Doom9's page where one could view/report results... :)

[bond]

There really should be an overviewable online form or something on Doom9's page where one could view/report results... :)well, you are invited to create and mantain such a page ;)

[masken]

well, you are invited to create and mantain such a page ;)
http://www.home.no/masken/other/smilies/maskens/m-smilie_wiggle.gif ...wish there were more hours in a day hehe ;)