dBpowerAMP Music Converter R9 Beta 2

[Spoon]

This is the next beta (beta 2) of dBpowerAMP Music Converter, it implements two new options, the more important being the 'Professional Frequency Conversion', this is an option and has to be switched on in the dMC Configuration program. Installing this dMC implements it in Wave and mp3 (Lame 3.92), installing the two other codecs add WMA and Ogg.
The second option is also on dMC Configuration - that is an option to write both V1&V2 ID Tags (I caved in after public pressure ). Download dMC R9 Beta 2:

http://www.dbpoweramp.com/beta/dMC-r9-beta2.exe

1.4MB

WMA v8 Codec:

http://www.dbpoweramp.com/beta/dBpowerAMP-codec-wmav8.exe

2.2MB


Ogg Vorbis (RC3) Codec :

http://www.dbpoweramp.com/beta/dBpowerAMP-codec-Ogg.exe

411KB

==================
Complete R9 beta 2 changes:

Audio CD Input - Send a (WM_USER + 101) message to Audio CD Input it starts ripping
Audio CD Input - Options Menu gets a new 'Character Replacement' option window - now possible to set the characters that are replaced and with what (certain characters such as * / cannot be used in filenames)
Audio CD Input - Added Track right click menu pictures
Audio CD Input - Power Pack Option - Compilation button gets a menu for multidisc albums - so can set a track offset number like 10..11..12
CDA Input Codec - Supplies more info for popup info tip, now more compatible with dAP

Codecs - Added 'Professional Frequency Conversion' option to dMCs Configuration page, the following codecs are compatible with this option - Wave, mp3 (Lame), Ogg Vorbis, WMA v8

dMC Config - for Mp3 - Added option to create V1 and v2 id tag files

mp3 Lame - uses Latest Lame 3.92

bug fix - when ripping from a CD, it makes sure that delete source files is not checked.
bug fix - If using 'Auto Rip' - now will not show the options page - will just convert
bug fix - If using 'Auto Rip' and 'No freedb Messages' and the Disc is not found in freedb the disc is now auto ripped

[bweeks_hcjb]

Hi Spoon,

I've done some preliminary tests, and I am happy with the results. There is no observable distortion added when converting sample rate (repeating tests I already did with the beta wave.dll). I have run tests on both WAV and LAME MP3 encoding.

I have also realized I made a mistake on my earlier tests. The filter is operating at fs/2, as would be most desirable, and I haven't observe any aliasing with the filter there. It seems good and sharp. For instance, a swept 20Hz to 20KHz sine wave sampled as 48 KHz, when resampled to 32 KHz, is cutoff sharply at 16KHz, which is what should happen.

My confusion arose from looking at a LAME MP3 both last time and this time, resampled to 32KHz at various bit rates. For instance, a mono 64kbps exhibits a sharp rolloff at 14 KHz. A joint stereo 96kbps exhibits a sharp rolloff at 12 KHz. I'll post another reply on this topic momentarily.

But as for the SRC part of things ... looks really good. I still want to test Ogg Vorbis, WMA, and Daren's MP2 and AIF encoders. Do Daren's encoders also respond to the Professional Frequency Conversion flag in dMC Configuration?

[bweeks_hcjb]

LAME Encoder Band Limiting

Following up on a note made in my last reply, I've encountered an undesirable "feature" in the LAME 3.92 encoder implementation that exists both in this Beta release of dMC and existed in Release 8d:

The LAME encoder is choosing a bandwidth for output files based on the bit-rate and channel mode selected. The sampling frequency selected seems to have no impact.

Some examples:
Joint stereo, 96kbps -> ~12KHz bandwidth
Joint stereo, 128kbps -> ~15KHz bandwidth
Joint Stereo, 192kbps -> ~19KHz bandwidth

Mono, 64kbps -> ~14KHz bandwidth

I tried the Joint stereo, 96kbps at 32KHz, 44.1KHz and 48KHz, all with the same result.

I can see that bandwidth is increased with bit-rate/channel. But this seems to be an undesirable affect. I checked the LAME project page briefly, and didn't turn up anything about a default setting or variation with bit-rate ... though there is a low-pass filter setting available: --lowpass. Is the dMC implementation twiddling with this? I'd sure like it not to, or at least have the ability to over-ride.

Thanks for looking into this.

[MODatic]

DMC is not twiddling with it. Those are Lame lowpass defaults. Higher Bitrate/available bandwith=less higher frequencies cut.

When you select the frequency in the config dialog, that is the setting of what the wave file is or what it will be resampled to before it gets encoded.

Lowpassing is required because at all bitrates, MP3 can not handle encoding the whole frequency domain without wasting a good deal of bandwith and thus sacrificing quality in other difficult areas of the song/s.

If you want know how bad mp3 can get without lowpassing, then try the [COLOR=darkblue]Blade[/COLOR] encoder since it is incapable, but please I beg of you, don't switch to it if quality is in your best interest.

[Spoon]

Darren is not using the Configuration on/off switch, if he wants to I could tell him where to find it.

[MODatic]

Weird things

I've been having a problem with r9b2. When ripping a CD with rip to ram enabled with the system priority set to anything else but normal on the ripping progress screen, the CD will either be ripped extremely slow (sub 1x) or not at all. Switching back to normal mid-rip will immediately give a huge speed bost or at least let DMC begin ripping the CD.

OS: Same as yours Spoon :rolleyes:

edit: BTW, rip & encode was also checked.

[daren]

Hey Spoon,

I found the "dMCProfFreqConversion" key, value = 0. Does this
mean that the default value is to use ACM freq conversion??

I haven't been able to find a GUI switch for this key....

I could include the switch in my codecs (for overall compatibility)
but I think there should be a checkbox in the codec settings
window (for each codec, or for all codecs) to make it clear that
there is a much better way to do the sample rate conversions,
than with the Windows ACM.

To be brutally honest, I'd rather drop the ACM method all
together... unless anyone can think of a good reason to keep
it in (it's quicker, but the quality is very rough).

Daren.

[bweeks_hcjb]

LAME Encoder Band Limiting

Thanks for your replies, Spoon and MODatic.

I was looking at the Lame command line interface spec, and it allows you to set explicitly a lowpass filter frequency (--lowpass) or ask for full bandwidth (-k). This may result with quality degradation, but I'd like to have access to experiment with such parameters. Is there a way you could open up access to sending or over-riding command-line parms in the LAME encoder?

In the meantime, I might play with a CLI version of the LAME encoder, and go to the LAME project to read some documentation about what decisions it's making about lowpass cutoff frequency vs. bit-rate vs. sample-rate.

[Spoon]

>I haven't been able to find a GUI switch for this key....

It is default off, and is switched on in the dMC Configuration program (Release 9).

[Spoon]

I have updated the Ogg Vorbis codec, the previous one would crash for CBR and ABR with any frequency other than 44100 (this was nothing to do with the new freq implementation, rather ogg it's self was crashing...)

[Spoon]

Superceeded by beta 3.

[Unregistered]

Re: dBpowerAMP Music Converter R9 Beta 2

Forum administrator:

This is a test. A Google search auto-created this "reply" page. It will be rather odd if this actually posts to a forum I've never visited.

Feel free to delete this page, if I am unable to do so.

[Spoons]

Re: dBpowerAMP Music Converter R9 Beta 2

if you created this using google then please show us the URL you did it from.
thanks.

[ChristinaS]

Re: dBpowerAMP Music Converter R9 Beta 2

if you created this using google then please show us the URL you did it from.
thanks.

Do you have access to server logs? Look at entries at about that time, by IP address. If you've got awstats (if it's been run yet) you may find accesses by search engine keywords. Or not.

[Spoons]

Re: dBpowerAMP Music Converter R9 Beta 2

yeh i found one google link that opens a reply with a quoted message ina thread.

not sure if it's an issue though.