IPB

Welcome Guest ( Log In | Register )

> Upload forum rules

- No over 30 sec clips of copyrighted music. Cite properly and never more than necessary for the discussion.


- No copyrighted software without permission.


- Click here for complete Hydrogenaudio Terms of Service

6 Pages V  « < 2 3 4 5 6 >  
Reply to this topicStart new topic
Winamp FhG AAC encoder with gapless fix, Updates from the 5.62 release
nu774
post Dec 12 2011, 06:56
Post #76





Group: Developer
Posts: 476
Joined: 22-November 10
From: Japan
Member No.: 85902



I'm not a good, trained listener like the guys here, but I think FhG encoder VBR preset 2 is very good, too.
I've had an impression that QuickTime tends to move spatial locations of the sound for lower bitrate setting (sound at slightly off from center is encoded into dead center). For this reason, QuickTime AAC is sometimes too easily ABX-able only by spatial information change, even for poor listener like me.
Though I didn't tried much, FhG seems not suffering from this kind of problem.

Go to the top of the page
+Quote Post
nu774
post Dec 12 2011, 07:20
Post #77





Group: Developer
Posts: 476
Joined: 22-November 10
From: Japan
Member No.: 85902



[attachment=6785:11_Homicide_5sec.flac]
Uploaded a sample to show the effect I wrote in the last post (though this is not a special example at all).
Try with qaac --he.
Go to the top of the page
+Quote Post
IgorC
post Dec 12 2011, 14:38
Post #78





Group: Members
Posts: 1506
Joined: 3-January 05
From: Argentina, Bs As
Member No.: 18803



QUOTE (C.R.Helmrich @ Dec 11 2011, 18:47) *
By "last versions", do you mean 3.2.3 and later?

Yep

QUOTE (C.R.Helmrich @ Dec 11 2011, 18:47) *
Were the average bit-rates of version 3.2.2 (the one of the 96-kb test) higher or lower than those of the Apple encoder on your test set at ~64 kb?

It was 3.2.4 and bitrate was just slightly higher than of Apple (+0.4%) while FhG produces 65 kbps on my large set of albums and Apple 66 kbps.
3.2.2 produced around 70 kbps on average.
Go to the top of the page
+Quote Post
darkbyte
post Dec 12 2011, 21:08
Post #79





Group: Members
Posts: 107
Joined: 14-June 11
Member No.: 91517



Maybe it's not the appropriate topic, but can we get an estimate when the FhG AAC encoder in Winamp will support HD-AAC encoding? rolleyes.gif

Btw, the VBR Preset 2 setting sounds really good considering that this is an AAC+ with SBR which i don't really like laugh.gif Although i've only listened to it through my amplifier yet. This way AAC+ always gives better result than listening to it using my phone. SBR's problems gets magnified with headphones by me.

This post has been edited by darkbyte: Dec 12 2011, 21:11
Go to the top of the page
+Quote Post
darkbyte
post Dec 13 2011, 20:07
Post #80





Group: Members
Posts: 107
Joined: 14-June 11
Member No.: 91517



Looks like i've killed this topic aswell. ohmy.gif
Go to the top of the page
+Quote Post
C.R.Helmrich
post Dec 13 2011, 22:53
Post #81





Group: Developer
Posts: 681
Joined: 6-December 08
From: Erlangen Germany
Member No.: 64012



QUOTE (darkbyte @ Dec 13 2011, 21:07) *
Looks like i've killed this topic aswell. ohmy.gif

Don't worry, you didn't. It's just that I (or we) can't answer this question at the moment. And thanks for your verdict of VBR mode 2!

Thanks, Igor! Then I'll leave the bit-rate tuning as it is.

Chris

This post has been edited by C.R.Helmrich: Dec 13 2011, 22:54


--------------------
If I don't reply to your reply, it means I agree with you.
Go to the top of the page
+Quote Post
darkbyte
post Dec 14 2011, 20:05
Post #82





Group: Members
Posts: 107
Joined: 14-June 11
Member No.: 91517



QUOTE (C.R.Helmrich @ Dec 13 2011, 23:53) *
QUOTE (darkbyte @ Dec 13 2011, 21:07) *
Looks like i've killed this topic aswell. ohmy.gif

Don't worry, you didn't. It's just that I (or we) can't answer this question at the moment. And thanks for your verdict of VBR mode 2!


Oh, i see smile.gif Sorry, just really like to test HD-AAC and keep asking it everywhere nowadays. laugh.gif

About the VBR2 thing: i've tested it today while walking to work. With headphones SBR artifacts are easily noticable (especially where it adds sinusoids to the replicate content, or it missuses high volume noise in a band to replicate a complex waveform). However it's not that bad as i've expected. Maybe some tuning can be made to hide these annoying artifacts (i think it would be still better to replace complex waveform with some low level noise than encode a loud artifact which has nothing to do with the original waveform). I wish i could understand the complex algorithms behind HE-AAC so i can do some tune on my own. Maybe one day, i'm a programmer so there's hope smile.gif
Go to the top of the page
+Quote Post
Gainless
post Dec 14 2011, 21:47
Post #83





Group: Members
Posts: 169
Joined: 28-October 11
Member No.: 94764



QUOTE (C.R.Helmrich @ Dec 13 2011, 22:53) *
Thanks, Igor! Then I'll leave the bit-rate tuning as it is.

Chris


Is it maybe possible to do some improvement on the high frequencies?
I've tested a sample with hi-hats on VBR mode 2 and the difference is really obvious...

This post has been edited by Gainless: Dec 14 2011, 21:47
Go to the top of the page
+Quote Post
C.R.Helmrich
post Dec 15 2011, 09:15
Post #84





Group: Developer
Posts: 681
Joined: 6-December 08
From: Erlangen Germany
Member No.: 64012



QUOTE (Gainless @ Dec 14 2011, 22:47) *
Is it maybe possible to do some improvement on the high frequencies?
I've tested a sample with hi-hats on VBR mode 2 and the difference is really obvious...

Probably not. As darkbyte mentioned, VBR 1 and 2 use SBR on the high frequencies, which is a parametric coding tool. Although it can get very close to the high-frequency input signal in terms of quality, it never fully reaches transparency in these frequencies. But you save quite a lot of bitrate. You can hear for yourself what it sounds like without SBR by choosing "AAC-LC Constant Bit Rate" encoding at 64 or 68 kbps in Winamp.

An SBR encoder is a quite complicated thing, darkbyte smile.gif But we'll check whether some fine-tunings are still possible.

Chris

This post has been edited by C.R.Helmrich: Dec 15 2011, 09:17


--------------------
If I don't reply to your reply, it means I agree with you.
Go to the top of the page
+Quote Post
hernaaan
post Dec 15 2011, 21:36
Post #85





Group: Members
Posts: 6
Joined: 7-March 09
From: Buenos Aires
Member No.: 67711



Is there a way to encode using FhG AAC codec out of Winamp? I mean, a command line encoder or something.

I'm interested in testing/comparing it but not very keen on installing third party software though.

EDIT: Sorry, just read this.

QUOTE (C.R.Helmrich @ Nov 17 2011, 18:47) *
And to be clear: the command-line encoder you are talking about is neither developed nor supported nor endorsed by Fraunhofer (Fraunhofer sells its own command-line encoder which has dozens of switches, incl. quality). Nullsoft, in particular benski, have put the encoder and a lot of work into Winamp so that people like and use their software. So please appreciate that by using the encoder through Winamp.


I may not totally approve this in terms of compatibility but, anyway, codec can be used for free.

This post has been edited by hernaaan: Dec 15 2011, 21:45


--------------------
$ lame -V3 -q2
Go to the top of the page
+Quote Post
ZinCh
post Dec 15 2011, 21:42
Post #86





Group: Members
Posts: 171
Joined: 28-September 06
Member No.: 35705



yes
Go to the top of the page
+Quote Post
lvqcl
post Dec 15 2011, 21:56
Post #87





Group: Developer
Posts: 3206
Joined: 2-December 07
Member No.: 49183



QUOTE (hernaaan @ Dec 16 2011, 00:36) *
I may not totally approve this in terms of compatibility but, anyway, codec can be used for free.


EULA:
QUOTE
3. RESTRICTIONS ON USE. Licensee may not: (i) modify or create any derivative works of the Software


This post has been edited by lvqcl: Dec 15 2011, 22:00
Go to the top of the page
+Quote Post
ZinCh
post Dec 15 2011, 22:12
Post #88





Group: Members
Posts: 171
Joined: 28-September 06
Member No.: 35705



so, if you follow terms - you cannot use this codec at all, because winamp itself have unresolved lgpl licensing problems described on second page of this thread
Go to the top of the page
+Quote Post
hernaaan
post Dec 15 2011, 22:26
Post #89





Group: Members
Posts: 6
Joined: 7-March 09
From: Buenos Aires
Member No.: 67711



QUOTE (lvqcl @ Dec 15 2011, 17:56) *
EULA:
QUOTE
3. RESTRICTIONS ON USE. Licensee may not: (i) modify or create any derivative works of the Software


'for free' stands for the price. I mean, free as in 'free beer', not as in 'free speech'.
(Apple and FhG are pretty much the same this way. Nero, not that much.)

I didn't came here to start any flame war, anyway.


--------------------
$ lame -V3 -q2
Go to the top of the page
+Quote Post
Brand
post Dec 15 2011, 23:36
Post #90





Group: Members
Posts: 312
Joined: 27-November 09
Member No.: 75355



Is it necessary to use the --ignorelength command when encoding with Foobar? Could it do any harm if it's used, slower speed perhaps?
I've read the description, but I'm not familiar with the technicalities, so I'm not sure how it applies to practical usage. (I encode mostly from FLAC, BTW.)
Go to the top of the page
+Quote Post
kode54
post Dec 16 2011, 02:08
Post #91





Group: Admin
Posts: 4498
Joined: 15-December 02
Member No.: 4082



It is a good idea, since foobar2000 outputs UINT_MAX for the length fields, which will probably be processed wrong.
Go to the top of the page
+Quote Post
Brand
post Dec 16 2011, 12:09
Post #92





Group: Members
Posts: 312
Joined: 27-November 09
Member No.: 75355



Thanks.
One thing I noticed is that when converting with Foobar the files are a tiny bit larger than when converting with Winamp. The difference is less than 0.01 MB for a track and it varies.
I used this command in Foobar:
CODE
--vbr 4 --ignorelength - %d

(I also tried without --ignorelength and it's the same as with it)

This post has been edited by Brand: Dec 16 2011, 12:09
Go to the top of the page
+Quote Post
mudlord
post Dec 18 2011, 04:00
Post #93





Group: Developer (Donating)
Posts: 797
Joined: 1-December 07
Member No.: 49165



QUOTE (hernaaan @ Dec 15 2011, 14:36) *
Is there a way to encode using FhG AAC codec out of Winamp? I mean, a command line encoder or something.

I'm interested in testing/comparing it but not very keen on installing third party software though.

EDIT: Sorry, just read this.

QUOTE (C.R.Helmrich @ Nov 17 2011, 18:47) *
And to be clear: the command-line encoder you are talking about is neither developed nor supported nor endorsed by Fraunhofer (Fraunhofer sells its own command-line encoder which has dozens of switches, incl. quality). Nullsoft, in particular benski, have put the encoder and a lot of work into Winamp so that people like and use their software. So please appreciate that by using the encoder through Winamp.


I may not totally approve this in terms of compatibility but, anyway, codec can be used for free.


Its a shame thier product is so badly protected.
IsDebuggerPresent()? Give me a break.
Go to the top of the page
+Quote Post
Gainless
post Jan 13 2012, 23:46
Post #94





Group: Members
Posts: 169
Joined: 28-October 11
Member No.: 94764



I think I've found a new problem sample. The total ABX result is not that great, but after a reset I got a 9/10.
When you've downloaded the sample you'll notice a more or less subtle hi-hat after the kick starts. It sounded kinda off and harsh to me, which was the reason I started the test. I think there is already a difference when the kicks start, but I mainly focussed on the hi-hats that start at around 16 seconds and "respond" to the kicks. Everything was converted with the latest Winamp btw.

Someone may do a re-test? smile.gif





Attached File(s)
Attached File  ABX___IM___Hush_Mail.rar ( 3.54MB ) Number of downloads: 129
 
Go to the top of the page
+Quote Post
Gainless
post Jan 14 2012, 12:08
Post #95





Group: Members
Posts: 169
Joined: 28-October 11
Member No.: 94764



Edit:
Forgot to mention it, the sample is encoded with the VBR 5 mode.
Go to the top of the page
+Quote Post
C.R.Helmrich
post Jan 14 2012, 14:18
Post #96





Group: Developer
Posts: 681
Joined: 6-December 08
From: Erlangen Germany
Member No.: 64012



QUOTE (Gainless @ Jan 14 2012, 00:46) *
...the hi-hats that start at around 16 seconds and "respond" to the kicks.

These are quite tough for an encoder, since they are actually composed of many extremely short clicks. See spectrogram of the first of those hi-hats in the FLAC file. However, the encoder preserves them very well, except for above 16 kHz or so.


Attached Image


I've ABXed the sample. Even if you disregard the last vote (I got tired), I'm still hovering around 20% guessing, so no significant ABX from me.

CODE
foo_abx 1.3.4 report
foobar2000 v1.1.10
2012/01/14 13:37:16

File A: C:\Unrar\Infected Mushroom - Hush Mail (sample).flac
File B: C:\Unrar\Infected Mushroom - Hush Mail (sample).m4a

13:37:16 : Test started.
13:39:17 : 01/01  50.0%
13:40:05 : 02/02  25.0%
13:41:07 : 02/03  50.0%
13:42:38 : 03/04  31.3%
13:43:43 : 04/05  18.8%
13:46:05 : 05/06  10.9%
13:47:38 : 06/07  6.3%
13:49:42 : 06/08  14.5%
13:50:07 : 06/09  25.4%
13:51:17 : 06/10  37.7%
13:53:18 : 07/11  27.4%
13:53:49 : 08/12  19.4%
13:54:44 : 08/13  29.1%
13:54:51 : Test finished.

----------
Total: 8/13 (29.1%)

Chris


--------------------
If I don't reply to your reply, it means I agree with you.
Go to the top of the page
+Quote Post
Gainless
post Jan 14 2012, 21:42
Post #97





Group: Members
Posts: 169
Joined: 28-October 11
Member No.: 94764



QUOTE (C.R.Helmrich @ Jan 14 2012, 14:18) *
QUOTE (Gainless @ Jan 14 2012, 00:46) *
...the hi-hats that start at around 16 seconds and "respond" to the kicks.

These are quite tough for an encoder, since they are actually composed of many extremely short clicks. See spectrogram of the first of those hi-hats in the FLAC file. However, the encoder preserves them very well, except for above 16 kHz or so.


Attached Image


I've ABXed the sample. Even if you disregard the last vote (I got tired), I'm still hovering around 20% guessing, so no significant ABX from me.

Chris


Thanks for your effort, Chris!
What do you think, is there any room for improvement on this issue (if you can call it like that)?
Go to the top of the page
+Quote Post
C.R.Helmrich
post Jan 15 2012, 00:06
Post #98





Group: Developer
Posts: 681
Joined: 6-December 08
From: Erlangen Germany
Member No.: 64012



QUOTE (Gainless @ Jan 14 2012, 22:42) *
Thanks for your effort, Chris!
What do you think, is there any room for improvement on this issue (if you can call it like that)?

Other than pumping more bits into high frequencies at VBR 5, no. I'll think about it. In the meantime, would you mind listening to the attached file at normal listening level (i.e. the same volume as during casual listening to music) and tell me how many tone pulses you can hear?

Attached File  hf_test44.wav ( 258.44K ) Number of downloads: 179

Chris


--------------------
If I don't reply to your reply, it means I agree with you.
Go to the top of the page
+Quote Post
Gainless
post Jan 15 2012, 12:54
Post #99





Group: Members
Posts: 169
Joined: 28-October 11
Member No.: 94764



QUOTE (C.R.Helmrich @ Jan 15 2012, 00:06) *
QUOTE (Gainless @ Jan 14 2012, 22:42) *
Thanks for your effort, Chris!
What do you think, is there any room for improvement on this issue (if you can call it like that)?

Other than pumping more bits into high frequencies at VBR 5, no. I'll think about it. In the meantime, would you mind listening to the attached file at normal listening level (i.e. the same volume as during casual listening to music) and tell me how many tone pulses you can hear?

Attached File  hf_test44.wav ( 258.44K ) Number of downloads: 179

Chris


The first one is pretty clear, the second one is already very quiet, and after that I hear nothing at all.

I guess that's not really good, isn't it?

This post has been edited by Gainless: Jan 15 2012, 12:54
Go to the top of the page
+Quote Post
C.R.Helmrich
post Jan 15 2012, 13:49
Post #100





Group: Developer
Posts: 681
Joined: 6-December 08
From: Erlangen Germany
Member No.: 64012



QUOTE (Gainless @ Jan 15 2012, 13:54) *
The first one is pretty clear, the second one is already very quiet, and after that I hear nothing at all.

I guess that's not really good, isn't it?

That's a common misconception here on HA. You listened at normal levels (i.e. not cranking up the volume), so your performance is perfectly normal, maybe even above average for your age. At normal levels I get the same result as you. Bottom line: you might have heard artifacts below 16 kHz, so I'm not going to put more bits into high frequencies.

Is this "issue" a show-stopper for you? Apparently this was very hard to ABX. Can you live with the encoder's current performance on this item? Or are you really looking for utmost transparency?

Chris


--------------------
If I don't reply to your reply, it means I agree with you.
Go to the top of the page
+Quote Post

6 Pages V  « < 2 3 4 5 6 >
Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 16th April 2014 - 15:25