new Open Source mp3 Encoder from Helix Community Options

[Rüdiger]

I found a problem with the new piping feature of the Helix encoder, I guess.

Using J.River Media Center 11, it is possible to use Helix as internal encoder (means ripping directly to mp3) by renaming the exe to Lame.exe and using the advanced command line parameter box.

Unfortunately, I don't have any benefit with regard to speed when using Helix in this scenario. The drive (52x CD-RW) doesn't spin up to max speed when using Helix encoder altough CPU load is only 30%, it rips only at 7x.

Using original lame.exe instead the drive speeds up correclty (you can hear it, it's quite noisy) and the speed goes up to 15x, faster is not possible because LAME creates so much cpu-load ;(

I guess it cannot be a problem of MC with my drive, since it works with lame-encoder piping and the drive itself can rip Audio-CDs with up to 32x (when using wave).

So I think it's a problem with the piping that keeps MC from speeding up the drive.
I also posted this behaviour in the MC forum.
http://yabb.jriver.com/interact/index.php?topic=29097.0
It would be really great if you could try to solve this problem together with J.River team.

Thanks a lot for this updated "Xing" encoder!!
Rüdiger

EDIT: Just found out, the Liteon has a feature called SMART-X t "optimize" DAE speed depending on used application, but since MC runs at fast speed with Lame, it really must be some kind of incompatibility between Helix and Smart-X. Nero DriveSpeed shows that using Helix Speed is limited by Smart-X to 8x, whereas with Lame it's theoretically up to 48x. Hope this helps

This post has been edited by Rüdiger: Aug 6 2005, 14:28

[QuantumKnot]

I use this command line successfully with CDex:

%1 %2 -V100 -X2 -U2

Be sure to untick "On the fly encoding" or CDex will crash completely.

As for EAC, I too get the error code. Solution for me was to untick "Check for external programs return code" under compression options.

I'm using "on the fly encoding" in CDex with the following command:

- %2 -X2 -U2 -V80 -HF2 -F19000

The trick is ticking the "send wav header to stdin". After you select that, then away she goes

[DigitalDictator]

Has anyone tested the CML rev12 (Only for test) binary yet?

[Enig123]

Guruboolez's test (mentioned http://www.hydrogenaudio.org/forums/index....howtopic=36062# ) uses rev12 with default settings.

[karl_lillevold]

All the contributed changes have been checked into CVS on helixcommunity, incl. the VS .NET standaline makefile. I also modified the ribosome build system Umakefil to use the pre-built asm optimized .obj files on Win32/x86, such that the encoder built from this should be the same speed as the one built from the standalone VS .NET makefile.

Thanks to everyone who contributed suggestions and improvements!

[LoFiYo]

Hello Folks,

I'm wondering which switches are safe to use? I've been using only -X2 and -Vxx and sometimes -HF.

Also, when RealPlay says ~128kbps VBR, the equivalent in this encoder seems to be -v56. Can someone from Helix confirm this?

Edit: 8/20/05

This post has been edited by LoFiYo: Aug 21 2005, 04:20

[lex_nasa]

I've been using the following settings:

-V135 -X2 -U2 -TX8 -HF2 -F19000

I wanted good sound quality whilst still playing on all iPods without skipping... LAME sound quality is good, but it's iPod compatibility is not.

This post has been edited by lex_nasa: Sep 21 2005, 15:59

[QuantumKnot]

hmm...I've been using CML MOD r11 with CDex and been doing on-the-fly encoding. I noticed there's a new CVS version on rarewares so I grabbed that. But now CDex doesn't work with it. It just says it had an error sending data to the external encoder. It works normally (without on-the-fly). I replaced it with the old r11 encoder and it works again. Did something break in the stdin input option of hmp3.exe ?

[john33]

IIRC, the CVS version was supposed to be no more than the incorporation of the patches, etc., provided here. However, I didn't personally check to see whether anything was changed that would have broken any functionality. I'll compare the two versions later.

[john33]

Hmmm, can't see anything obvious. I'll look closer later.

[QuantumKnot]

Hmmm, can't see anything obvious. I'll look closer later.

Thanks. Maybe it is a compiler issue?

[lex_nasa]

stdin doesn't work for me either... had to use %s in fb2k.

[john33]

Hmmm, well that was interesting - I'm actually surprised it functioned at all!!

I found and fixed the bug, and a new compile with amended source is at Rarewares now.

For anyone interested, take a look at the first line under the leading comment block of the previous 'tomp3.cpp' and you'll see '/*'. A comment starter that was left from a previous comment block that was removed along with the closing '*/'. As usual, looking for something complicated when it was, and usually is, something completely silly!!

[lex_nasa]

double hurrah, works like a dream, thanks!

[DigitalDictator]

I don't know if I should mention it here but I found a sample with serious drop-outs (or whatever it is...) on the right channel at -V65. I have otherwise failed every attempt to ABX anything by HELIX at -V65 (and LAME -V 5!). But I just noticed this one while out running with my portable! I'm like, what is that noise?!? So I had to ABX it as soon as I got back home. I immediately re-encoded it with LAME and the problem was gone. I think I have to reproduce the file and see if it's still there...

[lex_nasa]

I guess that's why we're testing it, if LAME solves your problems then LAME is the way to go.

This post has been edited by lex_nasa: Sep 28 2005, 17:23

[QuantumKnot]

Hmmm, well that was interesting - I'm actually surprised it functioned at all!!

I found and fixed the bug, and a new compile with amended source is at Rarewares now.

For anyone interested, take a look at the first line under the leading comment block of the previous 'tomp3.cpp' and you'll see '/*'. A comment starter that was left from a previous comment block that was removed along with the closing '*/'. As usual, looking for something complicated when it was, and usually is, something completely silly!!

Thank you so much!! I can't wait till I get home and try it.


EDIT: It works!! No progressive stats are printed out though, but who cares. As long as it encodes

This post has been edited by QuantumKnot: Sep 29 2005, 12:25

[lex_nasa]

Is anyone brave enough to do some ABX vs LAME at equivalent bit rates to V2?

This post has been edited by lex_nasa: Oct 4 2005, 17:26

[lex_nasa]

Try -V150 -TX8 -X2 -HF2 -U2 -F19000, and compare with LAME 3.97b1 -V1 -vbr-new. Sorry to hark on, but this thing should be tested, I'm just sorry that I don't have time to do a properly documented ABX.

I apologise if I'm breaking the HA rules by not qualifying my demands by a proper ABX, but concensus is not always a good thing, it generally results in mediocrity, so treat me with the disdain of an outsider and I'll be happy.

This post has been edited by lex_nasa: Oct 13 2005, 18:06

[level]

CML rev12 (Only for test) binary uploaded as usual, at the last post.
http://www.hydrogenaudio.org/forums/index....t=25&p=317595&#

* Add a switch -SBT that can set the short_block_threshold (default is 700). This value can range to negative values.

** Short block detection logic changed a little.


Any further testing is welcome. For the problem sample level offered http://www.hydrogenaudio.org/forums/index....c=35531&st=100# , can someone (level) find a -SBT value to let the distortion gone?

With SBT set to default, we can see if the block-switch logic change make any sence. This kind of result is also appriciated.

After doing many ABX tests with distortion_guitar.wav and Originalsmall.wav I was able to improve the performance of encoder with these samples.

COMMENTS:

distortion_guitar.wav: the combination of SBT and TX switches improved remarkably the quality. The adequate values of SBT and TX did this sample very hard for ABX (almost transparent) [SBT=450; TX=0]

Originalsmall.wav: This sample is more ABXable, but, with SBT=450 and TX=0 the quality is a lot better. For a setting of -V120 -SBT450 -TX0 the result is the same as with -V150 without additional switches. This sample is easy ABX with aoTuV_b4 @q6.


MY COMMAND LINE: -V120 -X2 -HF2 -SBT450 -TX0

A explanation: This command line is the result of many ABX tests with many of my music and mainly with distortion_guitar.wav and Originalsmall.wav samples.

SBT switch: SBT<700 improves the performance; however, values of SBT<450 doesn't improve the sound; at least for my ears.

TX switch: It changes remarkably the sound quality. The strange thing is that to greater values of TX the bitrate increase and the sound quality is worse . I tested values in the range from 0 to 12; the best value (less artifacts with distortion_guitar.wav) was TX=0.

V switch [with SBT450 + TX0]: with my two difficult samples I found that for values V<120 the sound quality was worse in comparison with -V120. For values of V>120 the sound quality doesn't improve significantly.

HIGH FREQUENCY switch: I eliminated the -F19000, because seems that this adds something of extra phasing (very little) to distortion_guitar.wav; although I'm not sure, I need more testing for to be sure. However; pure -HF2 (without -F combination) has a cutoff in 19500 Hz; which the -F19000 switch is redundant.

MUSIC (The samples)
I tested [-V120 -X2 -HF2 -SBT450 -TX0] in my few free time the last two months with many of my music:

1) Kraftwerk - Computer World (217 kbps) [electronic]

2) Kraftwerk - Pocket Calculator (219 kbps) [electronic]

3) Marty Friedman - Tibet (196 kbps) [Instrumental rock]

4) Marty Friedman - Trance (189 kbps) [Instrumental rock]

5) Michel Petrucciani - Play Me (202 kbps) [Piano jazz]

6) Michel Petrucciani - Looking Up (198 kbps) [Piano jazz]

7) Cat Stevens - Wild World (170 kbps) [Acoustic guitar with male vocals]

8) Cat Stevens - Morning Has Broken (197 kbps) [Acoustic guitar with male vocals]

9) Santana - Black Magic Woman / Gypsy Queen (214 kbps) [Latin rock]

10) Santana - Waiting (218 kbps) [Latin rock]

11) Peter White - Café Mystique (233 kbps) [Smooth jazz]

12) Peter White - Long Ride Home (227 kbps) [Smooth jazz]

13) Marty Friedman - Angel (189 kbps) [Instrumental rock]

14) Yanni - The Mermaid (167 kbps) [New age]

15) Iron Maiden - The Ides Of March (198 kbps) [Heavy metal]

16) Iron Maiden - Wrathchild (205 kbps) [Heavy metal]

17) Pet Shop Boys - Left To My Own Devices (219 kbps) [Disco pop]

18) Pet Shop Boys - So Hard (217 kbps) [Disco pop]

19) Pet Shop Boys - It's Alright (220 kbps) [Disco pop]

20) Pendragon - The Voyager (205 kbps) [Symphonic rock]

21) Pendragon - Queen Of Hearts: (ii)...A Man Could Die Out Here... (221 kbps) [Symphonic rock]

22) Metallica - For Whom The Bell Tolls (205 kbps) [Metal]

23) Carpenters - Goodbye To Love (214 kbps) [Ballad with female vocals]

24) Yanni - Farewell (195 kbps) [New age]

25) Yanni - Almost A Whisper (168 kbps) [New age]

Range: 167 --- 233 / Average: 204 kbps

For my ears [-V120 -X2 -HF2 -SBT450 -TX0] was in ABX total transparent with all of these 25 songs of different music.

GENERAL CONCLUSION:
IMO based in my exhaustive ABX tests; Helix work as a champion (at least to my set of 25 songs). Of course, this is only my impression based in my ears.

Could please someone confirm this?

Another thing... my ears are not tuned for classical music, my test was based in my music, mainly rock, jazz and pop.

Could please someone with good ears in classical music to check whether my combination of switches is good or not? maybe... guruboolez?
I would appreciate much this.

[level]

MY COMMAND LINE: -V120 -X2 -HF2 -SBT450 -TX0

GENERAL CONCLUSION:
IMO based in my exhaustive ABX tests; Helix work as a champion (at least to my set of 25 songs). Of course, this is only my impression based in my ears.

Could please someone confirm this?

Anyone?

[woody_woodward]

MY COMMAND LINE: -V120 -X2 -HF2 -SBT450 -TX0

GENERAL CONCLUSION:
IMO based in my exhaustive ABX tests; Helix work as a champion (at least to my set of 25 songs). Of course, this is only my impression based in my ears.

Could please someone confirm this?

Anyone?

Could you tell us what the "-TX" parameter is? The documentation (such as it is) left me in the dark. Thank you.

[DigitalDictator]

would these switches improve the quality at lower bit rates as well? For examle -V60? (around 140 kbps, for metal at least)

[level]

Could you tell us what the "-TX" parameter is?  The documentation (such as it is) left me in the dark.  Thank you.

Sorry.. but.. I don't have idea that this switch does. The documentation available doesn't say anything with respect to this switch. However, it produces a really impact in the sound quality.
It reduces (as already mentioned) dramatically the warbling phasing problem in the distort_guitar.wav sample, where mainly, Helix has problems.

Could please Karl tell us what the "-TX" switch is?

[Rüdiger]

Guess I found a BUG

On some music files (wav), the encoder creates mp3s, that play skippy on some players. I tested J.River Media Center, Media Player Classic, FFDShow Audio Decoder and Winamp.

It plays fine on Winamp and FFDSHOw, and skippy on the two others.

Any idea what's going on? I dunno if it is a problem of the encoder or of the player?? All I noticed that the files have very low bitrate, compared to other files created with the same parameters: "-v90 -x2 -u2 -HF2"

You can get a sample of the skippy file here:
Skippy File

Please help me, this bug is making me anxious about my music.

Thx
Rüdiger