AVInaptic. Free utility which analyzes multimedia files (Linux and Windows)

[fsinapsi]

Hi.
I am not sure this is the right spot to announce this, so please excuse me if I'm wrong.
I'd like to present you my program AVInaptic, a free utility which reports many technical informations about multimedia files.
Where can you download the program?
Here: http://fsinapsi.altervista.org
The page is written in Italian but the program can display messages in English, too. I hope you can understand what I mean, my English is so bad...
For Windows users: all you need to download is the file avinaptic-yyyymmdd-full.zip
(note the suffix "-full")
For Linux users: dowload the file avinaptic.tar.gz and check for missing libraries.
My program is not aesthetically pleasing, but can extract many informations from AVI, MP4, MKV, and can do a "DRF analysis" of ASP and AVC streams.
Please feel free to send me suggestions and bug reports.

[buzzqw]

here is some image from parsing an X264+AAC+MP3 into MKV

http://img338.imageshack.us/img338/949/avinaptic1nq3.png (http://imageshack.us)

http://img170.imageshack.us/img170/6213/avinaptic2xa3.png (http://imageshack.us)

http://img218.imageshack.us/img218/7125/avinaptic3ux2.png (http://imageshack.us)

http://img208.imageshack.us/img208/3139/avinaptic4zj3.png (http://imageshack.us)

BHH

P.S mirror link http://www.64k.it/andres/data/a/avinaptic-20070306-full.zip

[buzzqw]

some other analysis and nifty tools

http://img241.imageshack.us/img241/7943/avinaptic5vg1.png (http://imageshack.us)


http://img291.imageshack.us/img291/5705/avinaptic6cc7.png (http://imageshack.us)

http://img100.imageshack.us/img100/4663/avinaptic7ms1.png (http://imageshack.us)

http://img139.imageshack.us/img139/1369/avinaptic8ya3.png (http://imageshack.us)

http://img90.imageshack.us/img90/5000/avinaptic9pb2.png (http://imageshack.us)

http://img530.imageshack.us/img530/373/avinaptic10ez8.png (http://imageshack.us)

http://img86.imageshack.us/img86/5964/avinaptic11xl8.png (http://imageshack.us)

http://img152.imageshack.us/img152/1253/avinaptic12ui4.png (http://imageshack.us)

BHH

[gekan]

Thanks. Very useful! :)

Eee... what does DRF mean?

[HeadBangeR77]

Looks very promising :) - gonna download and check today.
cheers,
HDBR77

Update:
It's working, and it's working like charm (so far at least :p)! I really like the amount of information it provides - as if the most important info from G-Spot, MPEG Modifier and MediaInfo (debugging mode) was put together in just one application. Many, many thanks!

Example:

[ About file ]

Name: Full-2pass-D6of9-VHQ4.avi
Date: 3/03/2007 17:24:20
Size: 185,117,426 bytes

[ Generic info ]

Play duration: 00:04:51 (291.332708 s)
Container type: AVI OpenDML
Number of streams: 1
Type of stream nr. 0: video
Audio streams: 0
JUNK: VirtualDub build 27021/release

[ About quality ]

Resolution: VERY HIGH (1024 x 432)
Width: multiple of 16 (GOOD)
Height: multiple of 16 (GOOD)
Average DRF quality: MEDIUM (3.342734)
Standard deviation quality: HIGH (0.670846)
Std. dev. weighted mean: HIGH (0.317105)

[ Video track ]

FourCC: xvid/XVID
Resolution: 1024 x 432 (*)
Frame aspect ratio: 64:27 = 2.370370 (~2.35:1)
Pixel aspect ratio: 1:1 = 1
Display aspect ratio: 64:27 = 2.370370 (~2.35:1)
Framerate: 23.976023 fps
Number of frames: 6985
Stream size: 184,811,809 bytes
Bitrate: 5074.934704 kbps
Qf: 0.478486
Key frames: 188 (0; 104; 259; 379; 425; ... 6850)
Null frames: 0
Min key int: 1
Max key int: 240
Avg key int: 37.154255
Delay: 0 ms

[ About MPEG4 encoding ]

User data: DivX503b1393p
User data: XviD0046
Packed bitstream: Yes (*)
QPel: Yes (*)
GMC: No
Interlaced: No
Aspect ratio: Square pixels
Quant type: MPEG custom (*)
Custom intra quant matrix:
8 10 11 12 12 13 14 15
10 11 12 13 13 15 15 16
11 12 12 14 15 15 16 17
12 13 14 15 15 16 17 18
12 13 15 15 16 17 18 19
13 15 15 16 17 18 19 19
14 15 16 17 18 19 19 20
15 16 17 18 19 19 20 20
Custom inter quant matrix:
10 10 11 12 12 13 14 15
10 11 12 13 14 14 15 16
11 12 12 14 14 15 16 17
12 13 14 14 15 16 17 18
12 14 14 15 16 17 18 19
13 14 15 16 17 18 19 20
14 15 16 17 18 19 20 20
15 16 17 18 19 20 20 20
Number of frames: 6985
Drop/delay frames: 0
Corrupted frames: 0

I-VOPs: 188 ( 2.691 %) #
P-VOPs: 3485 ( 49.893 %) ############
B-VOPs: 3312 ( 47.416 %) ############
S-VOPs: 0 ( 0.000 %)
N-VOPs: 0 ( 0.000 %)

Max consecutive B-VOPs: 2
1 consec: 2034 ( 76.094 %) ###################
2 consec: 639 ( 23.906 %) ######

[ DRF analysis ]

Average DRF: 3.342734
Standard deviation: 0.670846
Max DRF: 5

DRF<2: 0 ( 0.000 %)
DRF=2: 777 ( 11.124 %) ###
DRF=3: 3045 ( 43.593 %) ###########
DRF=4: 3155 ( 45.168 %) ###########
DRF=5: 8 ( 0.115 %)
DRF>5: 0 ( 0.000 %)

I-VOPs average DRF: 2.792553
I-VOPs std. deviation: 0.405478
I-VOPs max DRF: 3

P-VOPs average DRF: 2.788809
P-VOPs std. deviation: 0.409557
P-VOPs max DRF: 4

B-VOPs average DRF: 3.956823
B-VOPs std. deviation: 0.214809
B-VOPs max DRF: 5

This report was created by AVInaptic (06-03-2007) on 6 mar 2007, h 13:42:56

@ buzzqw:
I really appreciate your enthusiasm for your fellow's application, but could you just copy and paste the information using code tags? Taking several snapshots of on window isn't the most efficient way of posting results. ;) ;) ;)

[fsinapsi]

Eee... what does DRF mean?

DRF=Detail Removal Factor

[buzzqw]

@HeadBangeR77

an image is worth thousand words :)
i would like to show the program and varius options, i have followed the development of this software from its start and Frank had done a marvelous work !

No program so for can analyze the drf of avc stream, expecially inside mp4 or mkv container

@gekan

drf is usually a "measure of quality", the lower the better (take with a grain salt)

BHH

[HeadBangeR77]

an image is worth thousand words :)
i would like to show the program and varius options, i have followed the development of this software from its start and Frank had done a marvelous work !

No program so for can analyze the drf of avc stream, expecially inside mp4 or mkv container.
True enough. ;) For my limited uses the possibility of analysing streams inside matroska container is the greatest of virtues of this application (btw. I didn't know AVC command line options are stored in the stream! :eek:). I don't know the guy, I haven't followed the development of this tool, but I think I'm starting to share your enthusiasm. :)

drf is usually a "measure of quality", the lower the better (take with a grain salt)
Large grain of salt, since it depends on a matrix and source's characteristics. Yet if we analyse our own encodes it is very useful imo. :)

Example:
Average DRF quality: MEDIUM (3.342734) from my sample above is in fact a very high quality considering the matrix used. ;)

[foxyshadis]

I was just asking mean if he could include just some of this in avidemux, but this is the one to give anyone looking for raw statistics in the future. =D It even works with mkv, which is pretty rare, although per-track bitrate isn't supported, like most tools. Probably a lavf limitation. Great work, fsinapsi.

Minor design issue, the histograms are sized by dividing by the total number of frames, not the peak, so they're always very small with lots of dead space. (Avidemux has the same problem.)

The only real issues I have are in how much of a pain GTK usability can be, but that's just puff.

It'd be possible to include the matrix in the DRF quality calculation, probably, though for all the effort you can put into having a computer tell you what your eyes are going to see, you might as well just watch ten seconds of it. :p

[LoRd_MuldeR]

Great work fsinapsi, thanx for posting your tool :)

[fsinapsi]

It even works with mkv, which is pretty rare, although per-track bitrate isn't supported, like most tools.

this is a part of a report, before DRF analysis:

[ Video track ]

Codec ID: V_MS/VFW/FOURCC
Resolution: 640 x 360
Frame aspect ratio: 16:9 = 1.777777
Pixel aspect ratio: 1:1 = 1
Display aspect ratio: 16:9 = 1.777777
Framerate: 23.976000 fps

[ Audio track ]

Codec ID: A_MPEG/L3
Channels: 2
Sample rate: 48000 Hz


and this is the same part, after analysis:

[ Video track ]

Codec ID: V_MS/VFW/FOURCC
Resolution: 640 x 360
Frame aspect ratio: 16:9 = 1.777777
Pixel aspect ratio: 1:1 = 1
Display aspect ratio: 16:9 = 1.777777
Framerate: 23.976000 fps
Stream size: 180,222,645 bytes
Bitrate: 1017.877805 kbps
Qf: 0.184262

[ Audio track ]

Codec ID: A_MPEG/L3
Channels: 2
Sample rate: 48000 Hz
Stream size: 28,327,680 bytes
Bitrate: 159.991641 kbps

[foxyshadis]

Ah! I see now. In that case, you might enable bitrate graph for mkv, once you have the statistics. ;)

[Theliel]

wow, very good work!! gspot work fine with avi and mpg, but for mp4,mov... dont work fine.

i was waiting time ago for a good alayzer for mp4/mov :)

ty.

[fsinapsi]

Ah! I see now. In that case, you might enable bitrate graph for mkv, once you have the statistics. ;)

ok, I'll do it

[HeadBangeR77]

It'd be possible to include the matrix in the DRF quality calculation, probably, though for all the effort you can put into having a computer tell you what your eyes are going to see, you might as well just watch ten seconds of it. :p
For XviD encodes there's a matrix present in the statistics (see my sample above) - I don't know however, how much effort would it cost to implement this for AVC.

Btw. imho values like
Standard deviation quality: HIGH (0.670846)
Std. dev. weighted mean: HIGH (0.317105)
are much more interesting for me than an average DRF /quantizer. ;)

@ fsinapsi:
Being old-fashioned guy, who still uses old Win2k theme under WinXP and doesn't like all those graphical bells & whistles, I really love the simple, compact look of your GUI! :D

[Kurtnoise]

Nice tool....:)

This kind of tool is really missing especially in Linux world. I'll test it as soon as possible.


No program so for can analyze the drf of avc stream, expecially inside mp4 or mkv container
probably because it's useless...

[Sharro]

:thanks: :thanks: :thanks:

Also to HeadBangeR77 that brought to my knowledge this wonderful software!

[buzzqw]

probably because it's useless...

are you sure ???

how many times you have seen the drf analyzer log ?
personally some thousand times...

BHH

[elguaxo]

Great tool!!! And works with MKV/H264. Thanks!!!

[Pure]

Thanks, checking it out!

[BigDid]

Nice tool, working fine.

Instead of firing DrfAnalyser than Mpeg4 modifier (or reverse) I can get the same infos and more with Avinaptic :)

The name is strange though ;) Keep the good work.

:thanks:

Did

[Solo_Senshi]

Awesome program you made there.:thanks:

I just love all the info it gives you.:D

[elguaxo]

fsinapsi, I don't know if this is a bug. I run AVInaptic on one of my MKV/x264 encodes. Under About H.264 encoding deblocking is reported as -2:-2, but I know I used -2:-4. Any hints?

Thanks again for this great app.

[fsinapsi]

fsinapsi, I don't know if this is a bug. I run AVInaptic on one of my MKV/x264 encodes. Under About H.264 encoding deblocking is reported as -2:-2, but I know I used -2:-4. Any hints?

in the present version, my program reports only the following informations from the bitstream:
- the number of SPSs (sequence parameter sets);
- the number of PPSs (picture parameter sets);
- for every SPS: num ref frames;
- for every PPS: entropy coding flag (CAVLC/CABAC), weighted prediction flag (yes/no)

Any other information is copied from the "Unregistered user data".
User data are just comments the codec can leave inside a bitstream...

[elguaxo]

Thanks fsinapsi. I tried it with other encodes and the 'unregistered user data' is ok.

That particular encode, where the reported deblocking by the user data is wrong, was done with MeGUI/x264 rev.593. Perhaps it was a bug of that revision of x264. Other encodes that I did with newer revisions of x264 are ok.

[fsinapsi]

Minor design issue, the histograms are sized by dividing by the total number of frames, not the peak, so they're always very small with lots of dead space.

you can raise the value Misc->Preferences->Length_of_percent_line

[fsinapsi]

For XviD encodes there's a matrix present in the statistics (see my sample above) - I don't know however, how much effort would it cost to implement this for AVC.

added


Name vid-4g.mp4
[...]
User data eavc 1.2.0.14
User data MPEG-4 AVC / H.264 encoder
User data Copyright (c) 2004-2005 ATEME
User data http://www.ateme.fr/
User data v54am2LwWeWBRxLp0vJwFA
SPS id 0
Num ref frames 1
Custom intra4X4 luma
13 14 15 16
14 15 16 17
15 16 17 18
16 17 18 19
Custom inter4X4 luma
13 14 15 16
14 15 16 17
15 16 17 18
16 17 18 19
Custom intra8X8 luma
9 10 11 12 13 14 15 16
10 11 12 13 14 15 16 17
11 12 13 14 15 16 17 18
12 13 14 15 16 17 18 19
13 14 15 16 17 18 19 20
14 15 16 17 18 19 20 21
15 16 17 18 19 20 21 22
16 17 18 19 20 21 22 23
Custom inter8X8 luma
9 10 11 12 13 14 15 16
10 11 12 13 14 15 16 17
11 12 13 14 15 16 17 18
12 13 14 15 16 17 18 19
13 14 15 16 17 18 19 20
14 15 16 17 18 19 20 21
15 16 17 18 19 20 21 22
16 17 18 19 20 21 22 23
PPS id 0
Entropy coding type CAVLC
Weighted prediction No


I found this file (vid-4g.mp4) in the "divxtest cd" (www.divxtest.com). The name of this test is just "Custom Matrix Compatiiblity Test"...

[akupenguin]

That particular encode, where the reported deblocking by the user data is wrong, was done with MeGUI/x264 rev.593. Perhaps it was a bug of that revision of x264. Other encodes that I did with newer revisions of x264 are ok.

The userdata SEI is correct.
r597 | pengvado | 2006-10-29 17:09:21 -0700 | 2 lines
fix parsing of --deblock %d,%d (beta was ignored)

[Brother John]

Though I loathe GTK with all my heart, this is exactly what I need to find out the exakt codec version used in all of my older Matroska backups. Great work, fsinapsi!

[bob0r]

http://fsinapsi.altervista.org/

Ultima release di AVInaptic (10 apr 2007)

:cool:

[DarkZell666]

Ehm ... I just opened the magic.mgc file that's included with a text/hex editor, and there are some wierd things listed inside.

Here are some examples :

Quake I save: ddm4 East side
Hewlett-Packard
the_Dismal_Oubliett
Mathematica notebook


So ... what does this have to do with avinaptic ? :D
Did it get there by mistake or what ? Windows tells me it's a "Media Catalog File". Dude what's up ? XD

[buzzqw]

avinaptic will try to analyze everithing... but when failing will tray to use magic.mgc to extract some informatio (try opening a .pdf or doc..)

avinaptic use its own parser and routines but only for video and audio (and srt...) the unknow file are handled by magic.mgc

BHH

[survivant]

I'm not able to run the application in my Windows 2000 at the office. I run the application from a command prompt and nothing appear, no window, no error.. nothing.

The process is not running in the taskmanager too.

[buzzqw]

have you downloaded the full package ? http://www.64k.it/andres/data/a/avinaptic-20070410-full.zip

BHH

[survivant]

yep. Even try the version of March. do the same thing. and try of the mirrors.. do the same thing

[buzzqw]

if not launched from command line is fuctioning?

BHH

[survivant]

On windows 2000, I can't get it to work. If not even creating the avinaptic.db file.

At home on Windows XP, it's work.. but the program close when I try to open the movie generated from automkv (4.08gigs), but when I open smaller file it's work.

[buzzqw]

i tested avinaptic on over 12gb files (but was dv avi file...)

i will report to fsinapsi (the authort of AviNaptic)


BHH

[smok3]

possible bug:
when run from command line


example:
avinaptic file.mp4


it will bug me with questions everytime i run it from different path (and it will also make a *.db file in that new path).

[buzzqw]

tested with full path ?

like c:\avinaptic\avinaptic.exe c:\test\test.mp4 ?


BHH

[smok3]

buzzqw: tnx, that did the trick :)

p.s. also, whats up with PAR display for avi files? (i thought avi doesn't have such info..?)

[SealTooGreat]

p.s. also, whats up with PAR display for avi files? (i thought avi doesn't have such info..?)
There is no DAR signalling in mpeg-4, only PAR. So, for example, XviD in AVI has PAR. Take a look at this thread Aspect Ratio(s) in MPEG4 (http://forum.doom9.org/showthread.php?p=562048#post562048).

[fsinapsi]

p.s. also, whats up with PAR display for avi files? (i thought avi doesn't have such info..?)

in avinaptic DAR = FAR * PAR.
AVI Open DML specifications introduced a chunk called "Video Property" (VPRP) or "Video Property List", which has two fields called "aspect.x" and "aspect.y". DAR = aspect.x / aspect.y. So, if such info is present, PAR will be derived as DAR/FAR.
Otherwise, PAR will be eventually extracted from the bitstream, and DAR derived as above.
Informations which are present at container level come before informations extracted from bitstreams.
So, if an audio stream is stored as VBR, it will be reported as VBR (note: a CBR audio stream can be stored in an AVI as VBR).

[audyovydeo]

ciao fsinapsi & grazie per questo utilissimo tool.

I would like to know how "partial" is the "partial average bitrage graph", ie : how many frames does it analyse ?

Obviously I'd like to have the graph of the total number of frames, but if I know how many frames it uses right now I can tailor my test clips accordingly.

grazie
audyovydeo

[fsinapsi]

I would like to know how "partial" is the "partial average bitrage graph", ie : how many frames does it analyse ?
the meaning of "partial" in "partial bitrate" is like in partial sums (http://en.wikipedia.org/wiki/Series_(mathematics)):
Sn = a0 + a1 + ... + an.

I tried to reply your email, but i received the following error:
Remote MTA xmxatip.excite.com: SMTP diagnostic: 554 5.7.1 Service unavailable; Client host [212.216.176.118] blocked using blackholes.excite.com;
Your message could not be delivered due to complaints we received regarding the IP address you're using or your ISP. See http://blackholes.excite.com/ Error: WS-04
Please send me another address, and I'll send you a more detailed answer (in Italian).

[GimmeAreason]

Thanks a lot for this great tool. I can stop with Gspot now as this is better for me.
Cheers!

[buzzqw]

update ! 30/07/2007

http://www.fsinapsi.altervista.org/code/avinaptic/index.html

mirror (windows) : http://www.64k.it/andres/data/a/avinaptic-20070730-full.zip

mirror (linux) http://www.64k.it/andres/data/a/avinaptic.tar.gz

changelog

- better export of bbcode tags
- fixed some errors

thanks Frank!

BHH

[fsinapsi]

update 16-09-2007

http://fsinapsi.altervista.org/code/avinaptic/index.html


fixed some bugs (e.g. the number of channels of AAC audio streams);
added more infos about MP3 (and MP2) audio tracks in AVI.


Here is an example of a report created by the last version:

[ About file ]

Name: basic-instinct-2.avi
Date: 10/09/2007 14:18:56
Size: 10,018,018 bytes (9.554 MB)

[ Generic info ]

Play duration: 00:02:11 (131.3 s)
Container type: AVI OpenDML
Number of streams: 2
Type of stream nr. 0: video
Type of stream nr. 1: audio
Audio streams: 1
ISFT: MEncoder Sherpya-MinGW-20060312-4.1.0
JUNK: [= MPlayer junk data! =]

[ Relevant data ]

Resolution: VERY LOW (224 x 176)
Width: multiple of 32 (GOOD)
Height: multiple of 16 (GOOD)
Average DRF quality: HIGH (2.214785)
Standard deviation quality: HIGH (0.505009)
Std. dev. weighted mean: HIGH (0.504816)

[ Video track ]

FourCC: XVID/XVID
Resolution: 224 x 176
Frame aspect ratio: 14:11 = 1.272727
Pixel aspect ratio: 1:1 = 1
Display aspect ratio: 14:11 = 1.272727
Vprp aspect ratio: 4096:2395 = 1.710229
Framerate: 20 fps
Number of frames: 2626
Stream size: 7,722,336 bytes
Bitrate: 470.515521 kbps
Qf: 0.596737
Key frames: 131 (0; 59; 83; 100; 111; ... 2569)
Null frames: 55 (1; 2; 3; 4; 5...)
Min key int: 1
Max key int: 136
Avg key int: 20.045801
Delay: 0 ms

[ Audio track ]

Audio tag: 0x50 (MPEG (MP1/MP2))
Bitrate (container): 128.016 kbps VBR
Channels (container): 2
Sample rate (container): 44100 Hz
Chunks: 5039
Stream size: 2,106,300 bytes
Preload: 496 ms
Max A/V diff: 526 ms
MPEG frames: 5039
Play duration: 00:02:12 (131.631 s)
Bitrate (bitstream): 128 kbps CBR
Chunk-aligned: No
Emphasis: none
Mode: stereo
Delay: 0 ms

[ About MPEG4 encoding ]

User data: XviD0041
Packed bitstream: No
QPel: No
GMC: No
Interlaced: No
Aspect ratio: Square pixels
Quant type: H.263
Number of frames: 2626
Drop/delay frames: 55
Corrupted frames: 0

I-VOPs: 131 ( 4.989 %)
P-VOPs: 2439 ( 92.879 %)
B-VOPs: 0 ( 0.000 %)
S-VOPs: 0 ( 0.000 %)
N-VOPs: 1 ( 0.038 %)

[ DRF analysis ]

Average DRF: 2.214785
Standard deviation: 0.505009
Max DRF: 5

DRF<2: 0 ( 0.000 %)
DRF=2: 2120 ( 82.490 %)
DRF=3: 357 ( 13.891 %)
DRF=4: 84 ( 3.268 %)
DRF=5: 9 ( 0.350 %)
DRF>5: 0 ( 0.000 %)

I-VOPs average DRF: 2.274809
I-VOPs std. deviation: 0.510252
I-VOPs max DRF: 4

P-VOPs average DRF: 2.211562
P-VOPs std. deviation: 0.504523
P-VOPs max DRF: 5

This report was created by AVInaptic (http://fsinapsi.altervista.org) (16-09-2007) on 16 set 2007, h 13:05:54

Note: In order to obtain a more accurate and complete report, you have to request the "DRF Analysis".
As you can see, the audio track is MP2 128 kbps CBR, but it was muxed as 128.016 kbps VBR.

[Schrade]

What is the definition of DRF? Is there somewhere you can point me to some info on it so I can read about it? My searching wasn't very fruitful.

[fsinapsi]

What is the definition of DRF?
DRF = "Detail Removal Factor" or "quantizer".