Yep... I'm having trouble getting this track to encode well. This track is setup with a siren sound that tracks left, center, right constantly. Additionally the lead singer's voice begins far left moves to center, then right and back to center again. Meanwhile there is a guitar hard right.

So far all the attempts I've made have produced poor results. Those seperate typically into one of two results. In the first there is an almost total loss of stereo seperation. Not much difference from mono. In the second, there appears to be good stereo seperation but, in the case of the vocal it goes strong left, and as it moves to center it fades to inaudibility and at some point as it moves to left it returns to full volume.

I've been using Grip (cdparanoia, lame) on a linux laptop with a variaty of options.

I'm hoping someone here has figured out how to get a good result for this song.

Glenn

Interesting. This sounds like a great "poster child" song for encoders to strut their stuff.

I don't know which encoders give the best stereo separation, and I don't know much about how the encoders do the separation, but here's what I do know:

There is an option called "Joint Stereo" for most encoders. I have only a vague understanding of how it works. I think it conserves bits by not encoding the tracks as two separate entities, but rather devotes bits to the stereo separation only as needed.

Did you ever try encoding with Joint Stereo turned off? This would require a higer "base" bitrate to acheive the same quality level as an equivalent Joint Stereo file (nearly double, I'd guess), but the tracks would be two separate entities.

Now you've got me wondering about this. My car's stereo image is awful, and I always assumed it was just because of the way my speakers were set up (nothing in the dash yet). Maybe it's just that, or maybe the stereo image isn't as good with MP3s. I don't know. Maybe I won't know for sure until WAV support is available for the Empeg?

___________
Tony Fabris

Some of the last tests I did were at max (320) bitrates, in both joint stereo and otherwise. Turning the joint stereo on and off seems to be the major difference I noted. Encoding as stereo & 320kbps seemed to produce the best result, This was far from satasfactory compared to listening on the cd player. Even using cheap head phones.

Glenn

Now we're back to the old "previewing the tune on two different pieces of equipment" problem.

Hang out for 1.1, when you'll get WAV support on the empeg, then you can preview the tune in both formats on the same device.

Or, if you have a CD burner, use the WinAmp disk writer plug-in to decode your encoder tests to WAV and re-burn them as CD audio tracks. Then you can preview everything the same way.

___________
Tony Fabris

There is an option called "Joint Stereo" for most encoders. I have only a vague understanding of how it works. I think it conserves bits by not encoding the tracks as two separate entities, but rather devotes bits to the stereo separation only as needed.

I don't know the exact algorithm either but I've seen a simple explanation of what joint stereo does that sounds plausible. Suppose you want to store two 16 bit numbers A and B. One option is to store them separately, that takes 32 bits. Another is to store A+B (17 bits) and A-B (up to 16 bits but potentially much fewer if A and B are close). That's essentially what the MP3 encoder does in joint stereo mode, except it has a limited number of bits to work with (determined by the chosen bitrate). It just allocates more bits to the sum of the channels than to the difference, hoping that the stereo separation isn't too high.

So yeah, try it without joint stereo and see if that helps.

Best,
Borislav

Now we're back to the old "previewing the tune on two different pieces of equipment" problem.
-----

Didn't realize that comparing the sound of an original audio cd, on the equipment it was meant to be played, against the endoced resultes played on and empeg was such a Hot topic.

I wasn't carping about the player. I was reporting a lack of success on getting an acceptable conversion for a unique track. The specific problem being either, a whacky left right encoding, or a lack of spatial (left-right) resolution.

All I really meant to say was that this track played on a cheap portable cd player and headphones, evidenced a level of spatial resolution that I couldn't come close to replicating on the player.

My conversion attempts did go the the extreem of bitrates and to using both stereo and jstereo.
Stereo merged the chanels too much. Jstereo on the other hand had strong channel seperation, but failed to meet the challenges of this track. It simply could not deal with three differently pitched audio sources (high siren, male vocal, and guitar rif). Jstereo seemed to lock in on the fenquencies of the strongest source untill some other source overwhelemed it, then there was a dramatic shift left or right.

Time for a technical question. I'm not sure where this question will lead. Is it possible to rip the channels individually, then recombine them? In other words, can the encoder be prevented from any kind of channel mixing? Do encoders do that, mix the streams?

G_E_N
_L_N >encoder > GLENN

Ok, I've been digging into the entire process I used, to rip and encode my tracks.

I found esdplay loaded on my RedHat based laptop. Playing the wav file thru it shows some of the same problems. I think this eleminates Lame as a source of the problem. GIGO. That leaves either cdparanoia or the hardware as the garbage source. I'm suspecting the hardware. I'm going to try ripping on another machine next.

Meanwhile, Does any one have suggestions for a new laptop. One that rips music with excellence. I've been looking for a reason to buy a hot new Laptop, this just put me over the top.

Glenn

DAMN, I feel so stupid for not realizing this before.

I've had exactly the same problem- Stereo image problems with ripping. There's even a mention about it in one of the FAQs that quotes me. I'm so sorry I didn't bring it up before. I can't believe I had the answer to your question all along and just didn't think of it. (/me kicks himself in the head a few times.)

I've seen this happen with bad rips, where the stereo image flip-flops in places where it's not supposed to. As you surmised, it's the rip at fault, not the encoder. Good detective work.

I only have experience with desktop CD drives for ripping. I have a particular drive which will cause this exact problem unless I use WINDAC32, with very specific settings, to rip. Other rippers, despite their claims to produce perfect rips (such as EAC), cause errors on this particular drive.

I don't know about laptop drives and ripping. I know that some desktop drives specifically advertise their rip speeds and are marketed for that purpose. But I've never seen a laptop advertise it's ripping capability.

Does anyone know if there's a database on the internet where people log their rip results with various drives?

___________
Tony Fabris

I think this is important, so I put this one into the FAQ, with some more comments of my own.

Re-editing the original question it for the FAQ, I noticed this line:

, in the case of the vocal it goes strong left, and as it moves to center it fades to inaudibility and at some point as it moves to left it returns to full volume.

That should have tipped me off right away. I can't believe I missed that. (/me kicks self in head again.)

I've gotta read posts more carefully before I spout off.

___________
Tony Fabris

Don't kick too hard, Its usually happens when some one, like me, has a problem and makes an initial misdiagnosis. That tends to send everyone off chasing the wild goose. I've had my share of those experiences.

Glenn

For better of worse I've committed to a solution to my problem of ripping & encoding. I was able to borrow an Apple Wallstreet equipped with Sound Jam, and test ripping an encoding my problem track. The results were, compared to before, excellent.

So...

I've picked up an Apple Powerbook g3 series (wallstreet) from eBay. $865 wasen't too bad for a machine that originally sold for around 3g's. Overall I really like the Wallstreet, I would have preferred it have usb and firewire but for the price I can live with pcmcia cards, should they be needed.

Here is what's planned.

Hardware upgrades:
PowerLogics 500mhz BlueChip replacement cpu. (OS X compatable)
An MCE Xcarét Pro 2000 Expansion Bay CDRW.
At least 256mb ram.
20 or more gb drive internal
swapable 20gb (or more) drive for music files

Software: to start
MacOS 9.x
Sound Jam MP plus
jEmplode

For the near future, I'm leaning towards a single boot system based on OS X.
Thinking about lame. Also should more subtil problems with ripping crop up a Plexor scsi drive would be an EZ addition.

--Glenn

I welcome all to comment on this plan.

Sounds like a fun little machine you've got there. That's a lot more trouble than I would have gone to to get good rips (it's usually just a question of software and settings, or at worst, needing to buy a new CD-ROM drive), but if what you've got works, then definitely don't fix it.

___________
Tony Fabris

I never really like CPU upgrades in any laptop, price/preformance wise, it's just easier to get the laptop with the speed you want in the first place.
unless I went laptop only for my home computing needs, the portable CD burner would be kidna a waste.
RAM, the more the merrier :)
drive upgrades are ok, depending on the jump in space. normaly the older G3's have 4-6 gig drives in them, so a 20gig would be great.

12gig red mk2 -- 080000125

unless I went laptop only for my home computing needs, the portable CD burner would be kidna a waste.

Ya nailed it in one. I like laptops. The desktop I have stashed away in the closet would run at 4mhz, if it was plugged in.

I'm with you on the cpu upgrade, except the Wallstreets are wonderfully modular in this reguard. Cpu, L2 cashe, and memory sockets are all on a daughter card about the size of a playing card. Its a transparant upgrade.

Sounds like a fun little machine you've got there. Sound Jam in this config is ripping and encoding at around 5x speeds. It fluxuates between 4.6-5.4x This is at a 256kbps rate. --Glenn