I’ve only been away from HA for a week and wtf!? there’s a new recommended LAME setting.
HA Recommended LAME Settings
samples where aps or ape has a problem

APS has been transparent for me for a long time (an understatement since --alt-preset cbr 160 already sounds very good to me). I trust the ears here, which is why I’ll use --alt-preset standard –Z from now on.

A few hours ago, the usual weekend boredom kicked and I said to myself, “Hell, let’s re-rip and re-encode everything.” I most probably won’t hear the difference, and the quality improvement is not a big step as moving from remix to APS, or APS to MPC. I just want the warm fuzzy feeling that my MP3s are going to be better.

Just wondering if I’m the only one who’s bored and sick enough to do this.

Made a typo with the poll question. Can't figure out how to change it.

What I meant was now that -Z is recommended will you reencode?

Maybe it's time I update the 3.90.2 compile and just default this switch.... heh.

What if you encode using --alt-preset standard -Y? Is it now supposed to be --alt-preset standard -Y -Z?

I wouldn't say so, cause the primary goal of adding -Y to aps was to lower the bitrate at the cost of some quality, while the primary goal of -Z was to achieve a higher quality at the cost of a little more bitrate.

dev0

I think that's a great idea. I'd hate to make a typo and use -Y or -K instead of -Z. It would also discourage people from adding their own switches like -m s or -q 0.

I don't see any reason to reencode transparent encodings. -Z switch corrects or reduce some specific artifacts. Therefore, you should reencode the few tracks that might gain in quality, and not the others. If you're happy with your --aps encoding, don't throw them away.

I'll be re-encoding because I had transcoded only about 8 albums to mp3 from my FLAC archive. If I was looking at re-encoding more than, say 300-500 songs, I wouldn't bother, especially if I had to re-rip them.

-Y lowering quality in APS??

not a good idea i think

-Y cuts off more high frequencies and thus generates smaller files - for, say, usage on portable players (which is what many people here are using it for), it's a good choice.

agree. and i wont switch to APS -Z cause ive never found a problem in any tracks with APS. instead of tring to get a warm fuzzy feeling ill use my time to listen to my music

I'm surprised by the number of people sticking to APS now that it's no more recommended. Are you all, like Marteataca, in a "not-broken-don't-fix-it" mood ? Or have you got other reasons ?

I feel so lucky I use lossless and haven't gotten a hardware player yet so I didn't encode anything yet

I'll re-encode because I haven't encoded too many albums yet since I discovered this forum. The extra KB that it adds to a file isn't that big of a deal. If I was that pressed for space, I wouldn't be concered about quality and just use lower bitrate encoders or switches.

Can anyone actually hear the difference between aps and aps -Z?
I'm a real quality freak, but I sure as hell ain't re-encoding my aps
fiiles... Do I smell placebo?

To keep things simplified (and to prevent people from starting to add custom experimental options to the command line again), I'm going to update the compile to where alt-preset standard and extreme will use noise shaping 1 and bump the version number to 3.90.3.

It seems that the atmosphere has changed some since the alt-presets were originally introduced and that people aren't as concerned with space as they once were, so this seems to makes sense.

nope. people do hear a difference. again see this.

using -Z increases quality but in the expense of bitrate. With the 6 CDs I've reripped so far there has only been a size difference of 1-1.5 megs per CD. Encoding time has also been significantly faster. So i don't see any reason why to still use just plain aps.

I did not say that I still cling to aps.

What I'm saying is that re-encoding 100s of mp3s for a slight difference in
quality (given that aps is already near transparent) sounds like madness to me...

it is.

Maybe it would be sensible to introduce --alt-preset portable (or something else) as a mapping for --alt-preset standard -Y, becuase -Y reduces the bitrate quite significantly (172kbps; random sample) compared to aps (229kbps) or aps -Z (231kbps).

dev0

I think that is a good idea. Would anyone else argue for different switches for --alt portable though?

The -Y switch has no effect on most classical music. I can't imagine an official preset that have no effect for some users. IMO, it is a better thing to develop a real VBR preset, different from APS, than adding one limited to some musical or signal genre.

I'm quite surprised at the results of this poll.

I'm slightly surprised that 8 people (who I assume haven't heard any problems with their aps mp3s yet) are going to re-encode them all with aps -Z. I guess it depends how similar your music is to that where aps has been shown to have a problem, and how many mp3s you have. Maybe this is placaebo - once you know that someone else has heard a problem with aps, you'll never think that your mp3s sound quite as good again! (even though your mp3s almost certainly sound fine, and probably don't include artefacts).

I'm really surprised that 16 people won't use aps -Z for future encodings. I know it only fixes the occasional artefact, and it appears to be a 1% size increase. But if you'd rather save space than remove artefacts, 128kbps mp3 is for you Seriously, I guess "if you've never heard a problem, then there's no reason to switch." I just didn't think that this attitude would have brought people to aps in the first place.

Anyway, whichever camp you're in, you can download all the samples in the original thread and listen for yourself:
birds, glass_short and erhu demonstrate the benefits of aps -Z vs aps, whereas
vangelis, fugue and awe32 demonstrate the limits of both aps and aps -Z.


FWIW I voted for the second option in this poll. And if I was encoding just for me, and space was an issue, then I'd consider aps -Y -Z, because I know -Y worsens things that are beyond my range of hearing, whereas -Z corrects very occasional problems which are well within my hearing abilities. With only the usual space concerns, I'd use aps -Z. I do have one harpsichord CD, and if I had to mp3 it, I'd use api.

Cheers,
David.
EDIT: typo

By looking at the recommended presets, the -Z switch doesn't seem to improve the quality on the "fast" modes. Is that the case?

I will try re-encoding 3 or 4 of my favorite albums just to try aps -Z out and see if I hear any difference, I will also use the new setting for any music I encode form now on as I am a "do whats recommended" kind of person. However there is no way I am re-encoding my whole collection (I just finished encoding most of it this month ! originally -r3mix and more recently aps) that would just be too sad, not to mention totally crazy. I want to have time to listen to some of the music now !!

Well, I've only ripped 4 albums since finding this forum and I left the waves on my machine so re-encoding with -Z isn't a big deal.

Wasn't this -Z switch around when Dibrom (and others) did the initial testing and creation of the -alt-preset settings? Why are we just now figuring out that this is a better switch when it's part of the 3.90.2 compile? I could see if this -Z switch was part of the new lame compiles then we should be hyped about it, but I'm really disappointed that this switch has been there for some time but is only now recommended. I've got hundreds of CDs encoded with APS and to me they sound fine. Now I'm finding out -Z is/could be better. Frankly, my confidence in the testing procedures is now shattered. APS has been the standard for quite some time, and overnight it changed to APS -Z when that option was apparently there all along.

That's a bit harsh. It's only now been realized that there are samples in the real world that it can make a difference. And I'm sure that new samples will crop up that exhibit other types of artifacts that will require an additional change to the preset at some time in the future. You've also always had the option of --alt-preset insane. No changes seem to happen there.

Unless you'd like to test every possible wave sample?

So if one stays with 3.90.2 or one of the offical compiles of 3.90 or later, the recommended commandline is --alt-preset standard -Z, correct?

If you go with the updated 3.90.3, you would simply use --alt-preset standard. In fact, using the -Z switch with this compile would result in noise shaping 1 being turned off, right?

Heh. Unless I'm mistaken, this might get confusing for whoever is maintaining the list of recommended settings.

@dibrom:

if you update 2.90.2:

please make --preset switches in addition to the --alt-preset settings. so that --preset standard is the same as --alt-preset standard.

just to achieve conformity with current lame releases.

tnx a lot,
amano


Edit: and could it be possible to add an acm codec in this new bundle?

Perhaps you're right. I guess I'm a littled miffed because I just encoded hundreds of CDs using APS without the -Z setting only to find out there's a new recommended setting that I could have been using all along throughout my project. No big deal. After thinking about it a little more, I'm glad improvements are still being made to the best mp3 encoder we got. My apologies for the rant and my sincere thanks to the community for the continued efforts to improve.

It's not really better in a broad sense, it's better in a few select cases. Mostly cases that showed up quite a bit after the alt-presets were no longer being worked on by me.



The testing procedures that apparently you no longer have confidence in are the same procedures that discovered the remaining few issues with the alt-presets. Hell, without these testing procedures, there wouldn't even be an alt-preset standard as a basis for this small tweak.

Maybe if you would read up on the issue a little more and not take things out of context so much, you wouldn't be so angry about this matter.



No, it didn't overnight change to -Z. In fact, I never changed to recommended settings at all, someone else did. Since I no longer have much of an interest in the LAME project, I didn't bother to dispute this much. I also don't really have time to explain the matter to people who are now going to be confused by this whole thing (like you apparently are), so I figured I'd just not even try.

I'm going to update the default settings to use noise shaping 1 instead of noise shaping 2, so that this issue can be laid to rest. As others have already pointed out, there's no need to reencode your material unless you have heard a problem. This issue with -Z does not happen very often because --alt-preset standard is already tuned well enough to avoid the majority of problems which were present beforehand (which is also why these later samples were not known about originally).

I'm not going to update or change anything else beyond modifying the noise shaping mode. Someone else can do this if they want to but it won't be me.

3.90.2 is the only compile that has ever been recommended for use with the --alt-presets by Hydrogenaudio. 3.90.3 will be the next compile. Possibly 3.94 will be after that. As for the other compiles, yes, maybe things will be confusing for some people, but this issue has been beaten to death many times already on this board in some of the larger discussions about the problems with the LAME project. Put quite simply, I'm not going to worry about it.

If you use the --alt-presets and want to be "sure", then the best thing to do is stick with the Hydrogenaudio compiles until further notice.

If noise shaping 1 will be default, remember to make it so that --alt-preset standard -Z with 3.90.3 doesn't switch it back to noise shaping 2.

If you read my follow up post, you would see that I admit I overreacted and that I'm grateful for the continual testing efforts.



You assume I didn't read up on this. Just so you know, I read all the threads and clearly understanding the matter. You should not make assumptions about my knowledge, although I admit I said nothing to prove I know what's going on.



I never asked for your recommendation, nor did I ask for your explanation. Don't assume my post was targeted toward a response from you. It was a general rant, in which I promptly apologized.

Dibrom, you've got my deepest respect for all you've done. I'm honored I've finally triggered a response from you in one of my posts, however I wish it was under more positive circumstances. I'll crawl back into the hole I've been lurking in for years....

Daffy

I did read your post, but took your reaction as an opportunity to make a point about this matter. I have a feeling that you are not alone in your sentiments and so I thought that making a statement about the matter might help to clear some things up.



I may have jumped the gun a bit on that, though I fail to see how someone fully informed on this matter could feel the way you apparently do/did. I think a few of the factors in this matter go back quite a long ways and that there is more to it than just some of the more recent posts. I'm not sure how well you have been following this issue, but only a cursory examination of the current discussions will not provide enough information to fully understand the matter.



You mentioned the "recommended settings". These have traditionally been recommended by me (in regards to the alt-presets that is). Since you didn't make this distinction clear in your message, I took it upon myself to do so.

As for an explanation, it was geared not soley towards you, but moreso towards an explanation of the matter at hand to all who are reading this thread.



I didn't. But I see no need to refrain from the discussion either.



Thanks I guess, though compliments aren't really necessary. I don't really see a problem here myself, I was just discussing a few points. I'm sorry if you've been offended somehow.



Well that's up to you, but I don't see why that's necessary.