From randyd over at the r3mix tech forum:
I went back and encoded bloodline with A-P-Extreme and very surprised when I heard the 'ringing' that ff123 had described (obviously in the left channel). It was absent in the A-P-Standard file, but very pronounced in the A-P-Extreme file. To make sure of this I tested it and got 16/16 with .05% chance of guess. It was also present on the A-P-Insane file but BARELY audible and not like Extreme. I got 14/16 with a .5% chance of guess. BTW Insane encoded all the way at 320k. I suggest that one of you encode the bloodline file with A-P-Extreme and test this for yourself. Its a high pitched sound (somewhere between 15k to 19k or so) but very pronounced so you shouldnt have trouble hearing it.
Hrmm.. I'll have to check this, since this is one of the files I originally tuned on. It's possible that with one of the later modifications I made, something slipped by which would allow this to occur though, but that would be a bit of a surprise.
I have no idea when I'll be able to get to this for sure though since I'm very busy atm and probably will remain so for another week at least.
I just wonder if the guy is truely hearing the ringing or just the "line frequency" about 15.6kHz present in the original as well (but with slight difference). This is because he wrote:
I'm not sure if its the 'ringing' that ff123 describes: It sounds more like high pitch static, and is obvioiusly only can only be heard in the left (quiet) channel
Another reason for my sceptism that it's ringing he hears is that it's pretty unlikely he will hear left channel ringing with --alt-preset insane, since the frequency response is beautifully solid upto 20kHz where it cuts (until the music starts).
Then again, alt-preset standard also has small amount of over 16kHz ringing which should be audible just as well if one is able to hear ringing in --alt-preset extreme's case. Yet another thing is that he should be then able to notice that --r3mix rings very much more (starting under 14kHz) than aps, which seems he couldn't.
My guess is that this has something to do with that 15.6kHz line frequency because ringing doesn't sound like "high pitch static", ringing doesn't sound static at all.
Originally posted by JohnV
Yet another thing is that he should be then able to notice that --r3mix rings very much more (starting under 14kHz) than aps, which seems he couldn't.
Hmm.. he didn't test it on --r3mix
Well, a spectrum analysis might identify what is going on there.
The sample vilbel (long version) has also some dropouts that are sharper in extreme than standard (not all).
It's the case with the first or second one (I don't remember) of the short version (badvilbel).
Originally posted by tangent
Hmm.. he didn't test it on --r3mix
Well, a spectrum analysis might identify what is going on there.
But he said he had tested --r3mix.
I found the most differences with the bloodline files. Both the A-P-S and R3mix seemed to add slight artifacts that you could hear in the left channel just before the rest of the music kicked in.
If what he hears is ringing, --r3mix should sound very very much worse.. Then he says afterwards that it was absent in aps.
I can't of course say for sure what he hears... but "high pitched static sound" is not ringing.
Hmm... the part where he mentioned he first tested --r3mix and -aps is a bit confusing I guess, since I'm not sure what he was refering to in his post.
But you're right, it's probably not ringing he's hearing, but some other form of artifact which would be interesting to check out.
I suggest that before you discredit what I stated that you GO ENCODE BLOODLINE IN ALT-PRESET-EXTREME AND LISTEN FOR YOURSELF. I'm not hear to bring down Lame, I was just over at the R3mix forum trying to find out if Alt -Preset-Standard was REALLY better than R3mix or just hype. I took it upon myself to do some OBJECTIONABLE listening tests to see how much better it really is, and YES it does have advantages over R3mix except for filesize. It just so happens that I ran into a file that I heard the 'high pitch static' NOT ringing in the left channel (yes I DID encode it in R3mix and heard similar artifacts) and picked it 15/16 against the wave. I figured that Extreme could handle the file without a hitch. WRONG. There were slight artifacts with Standard but nothing like the high pitched ringing that I heard with Extreme, which I easily picked 16/16. Like I said, I'm not here to bring down Lame, I like Lame, where else can you get the best mp3 encoder on the planet and not pay for it??? FHG ?? Xing ?? I dont think so..... So GO ENCODE BLOODLINE AND LISTEN FOR YOURSELF
BTW I dont know if was ringing, static, line frequency, ear wax buildup or whatever, I HEARD something, and NOT in R3mix or A-P-Standard....
"High pitch static" is clipping, and bloodline is a clipping sample.
I don't see it clipping (though it reaches 0.0 db) in APS or APE. Either you've got some DSP, equalizer, or volume boost somewhere in your playback, either you've got better ears than me.
--scale 0.9 should solve the problem.
I have listened it, and I'm not trying to discredit your results, I was just wondering that if you hear static sound that is not in r3mix or aps, but is also in api, it can't be ringing. Also because of one non-public incident here, I've become even more sceptical recently..
But anyway, if you can't even say what you are hearing, it's not very helpful (ringing and static are quite the opposite). Secondly, pretty much the only possible static high pitched sound in the left channel before the music starts is that 15.6kHz line frequency, so maybe you are hearing something wrong with that.
Finally I think you said you now actually hear this distortion with r3mix,aps,ape and api, but with ape it's very clear, but you can't really say what..
This is just somewhat confusing. Try --alt-preset extreme --lowpass 16 to rule out over 16kHz ringing.
Originally posted by Pio2001
"High pitch static" is clipping, and bloodline is a clipping sample.
I don't see it clipping (though it reaches 0.0 db) in APS or APE.
Left channel before the music starts is very low amplitude and cant be clipping.
I'm sorry but I went back and encoded it with APE again and can still hear the ringing. Is it possible that I introduce aritfacts by decoding it back to wave for test porposes??? I cant get the ABX program to recognize an mp3 file so I decode for that reason. Someone let me know if theres a way around that.
Umm, so now you think it's ringing what you are hearing??
No, decoding to wav doesnt introduce any additional distortion, it will sound the same as mp3.
Did you try --alt-preset extreme --lowpass 16
?
before you say "try listening before you decode" I already did. I was going by tangent's general rules of thumb of ABXing. Listen first, and if you think you hear a difference test it. I thought I heard ringing, I tested, but I'm not sure if decoded hurts a file or not.
You decode in everycase, when you listen or just when you write a wav-file. If you use WinAmp, same decoding engine is used, so there's absolutely no quality difference if you just decode.
Now, could you test --alt-preset extreme --lowpass 16
and say what you think?
Aw, during the silence ? I just hear noise, with a faint echo from the other channel.
Analysis shows the same thing with both APS and APE :
The noise is -55 db RMS.
Two tones, 15640 and 15750 Hz, are 27 and 23 db above the surrounding frequencies.
But since I can't hear them, I can't judge.
ok 16k low-pass does seem to hinder the ringing but I can still pick it from the wave. I DO stand corrected on one thing so blast me if you want. There is ringing in the r3mix file but I didnt notice it the first time because a: I wasnt familiar with it until the Extreme file and b: I was concetrating on a different artifact. Yall go ahead and lay into me...
Ok, you hear ringing. But you say you hear also different type of distortion. What's this other type of distortion like then? Is it this static high freq sound or what?
And what kind of distortion in your opinion is present with api?
Is this other kind of distortion clearly more audible with ape, than with r3mix,aps or api.
ok, tested api it does NOT ring but i still pulled it 13/16 from the wav. I believe the closest thing describeing the sound would be a very high pitched tape hiss. It is not loud or annoying at all, just present. Note that this is in the quiet channel fo a stereo clip. There is no way I could distingquish it in a song that starts both channels evenly. BTW, I'm definitely not an audiophile or claim to be. I just seem to pickout high frequency artifacts. Besides, my hearing aint THAT good. Normal is 20 - 20k and mine reaches 19k right and 21k left so I guess thats about average.
JohnV, the 15.6kHz line frequency thing would also be in the original if it was there, wouldn't it?
Randy, what are you using to decode to wav?
Originally posted by tangent
JohnV, the 15.6kHz line frequency thing would also be in the original if it was there, wouldn't it?
Of course.
So, the distortion which was originally bothering randyd with a-p-e, is this very high pitched "tape hiss", not ringing? And it's most apparent with a-p-e? Can anybody confirm this?
http://www.ff123.net/samples/bloodline.flac (http://www.ff123.net/samples/bloodline.flac)
ok fellas, I may have a new explanation for the sounds that I hear. When I first encoded bloodline with R3mix and A-P-S I used the entire wave I converted from the flac file. Its 15 secs long. When I wanted to try to improve the sound by encoding with A-P-E I cut the file down to 8 secs to keep just the 'important' part. Is it possible that when I shortened the file (cool edit) I messed it up just enough so that A-P-E produced the ringing I heard (and no the shortened file did not have the ringing). On a hunch I went back and encoded The ENTIRE uncut file and did NOT hear any ringing this time. If this is the case I apologize for the stress i'm putting you guys through but I DID hear the artifacts I described whether A-P-E's fault or not. I can email you the file if you want to hear for yourself..
Hi all, first timer here at hydrogenaudio. I saw this thread and just had to respond. First off randyd, I think you are correct in your most recent post by stating that it may not be Alt-Preset-Extreme's (or A-P-E as you referred) fault but maybe the fault of the file you encoded. I do not think, however, that it is the fault of the wave editor you used as long you only shortened it and did not do anything to affect the sound (eq, normalize, etc). By reading your posts and through personal experience I know you had to decode the mp3 back to wave for the Abx tests. I think its very possible that when you decided to encode to A-P-E that you may have accidently used the wave that was already encoded with Alt-Preset-Standard. Back last year I did exactly that when I was comparing an R3mix file to a 256k CBR file. I couldnt understand why the 256k file sounded so rough until it dawned on me what had happened. It had the same "hiss" sounds that you described but I didnt hear any ringing though (my hearing range is not as wide as yours ) I would suggest that you go back and re-encode your test waves and name the files as you go to avoid any confusion. Then come back and tell us what you hear.
**MAXX**
Ok maxx, I did you like you said. I encoded every test file again and just to make sure I even reencoded the decoded APS back to APE. You guys aint gonna like the results though. The other night when I encoded the uncut bloodline file I either didnt listen well or 'hoped' I didnt hear ringing, because today I still do. The 'ringing' is most prevelant in the APE file with a hint in R3mix and possibly APS. API sounded fine. I believe the 'high pitch static' and 'hiss' I describe are indeed the line frequency that is in the music to start with. All the encoders just seem to enhance it even API to a very slight degree, NOT bothersome, just there. I also notice that the 'ringing' can be heard best just before the line frequency noise comes (About a second)and is kinda overshadowed after that. Might this explain something? l did notice that I could open the files in windows media 7.1 and cut the highest treble on the eq and ALL negatives went away. I dont know if this sais something for my hearing or about my system but thats what happened. For reference: SB live platinum, Sony mdrv700 headphones, Iwill KK266 motherboard with AMD 1.4GHz. Again, Does anyone want me to email them that file. (Dibrom???) I could be f**ckin something up, couldnt imagine how, but possible. BTW the maxx, the mp3 that got encode twice DID sound quite rough. I couldnt imagine many files that could handle that....