Skip to main content
Topic: Quality Level (-1) (Read 3915 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Quality Level (-1)

Hi All,

What's your impression of Quality Level (-1) in Ogg Vorbis (oggenc.exe v1.0)?  I encoded a Pop album (Kelly Clarkson - Thankful) and I'm quite pleasantly surprised by the general quality.  Honestly, I seem to be having trouble telling quality levels apart in Ogg.  For example, I don't notice any artifacts at all with the QL (-1) encode of the album.  Is this unusual?  I also encoded the album with LAME (--r3mix) and can't tell the difference between this encode and the Ogg encode.  In fact I can't even tell the difference when comparing to APEs or WAVs.  Is it that my hearing isn't sensitive enough or is Ogg really that good?  Also, are there any obvious examples of Ogg's flaws?  Pretty much down to 96 Kbps 44 Khz is indiscernable to me from 256 Kbps 44 Khz (in LAME MP3 encodings).  I ask this because while it seems good that I can encode at lower quality and still enjoy my music at decreased file sizes, I feel like I'm missing out on NORMAL hearing.  In any case, any help, comments, and/or advice would be appreciated.

-- alphaneutrino

Quality Level (-1)

Reply #1
Oggenc -q-1 produces a lot of artifacts :
  • Hiss
  • Electronic noise (on low volume parts of music)
  • Stereo issues
  • problem with bass
  • and of course traditional artifacts : pre-echo, distorsions....

Listen with an headphone to perceive some of them.

Quality Level (-1)

Reply #2
Sorry to be that harsh, but you are either deaf or just very untrained/unsensitive to artifacts. I  have to admit that Vorbis sounds pretty amazing at q -1, but it's distinguishable from the original any time and has some serious flaws on some music.

dev0
"To understand me, you'll have to swallow a world." Or maybe your words.

Quality Level (-1)

Reply #3
Quote
Sorry to be that harsh, but you are either deaf or just very untrained/unsensitive to artifacts.

Another possibility is that alphaneutrino just has some very cheap speakers. A few months ago I had a set of good ones die on me, and plugged an old pair in for a temporary replacement... and everything sounded like a 96kbps Xing MP3.

  Seriously, the listening equipment makes a big difference. alphaneutrino, if that's your situation you might consider spending a couple hundred to get a decent set of speakers.

    - M.

Quality Level (-1)

Reply #4
Quote
Seriously, the listening equipment makes a big difference. alphaneutrino, if that's your situation you might consider spending a couple hundred to get a decent set of speakers.

Once again, I disagree with this.  Listening equipment doesn't make all that big of a difference.  As long as you are using headphones, even cheap ones, that should cover the equipment side of things -- the rest is training and/or natural ability.  However, I believe that once you have been trained to hear these artifacts, you can even hear most of them on cheap speakers, at the very least in the case of using something like Vorbis at -1.

At any rate, that's not to say that one shouldn't strive towards higher end equipment or something, but it's certainly not the magic bullet for hearing artifacts like many people seem to think.

@alphaneutrino:

I suggest giving this page a visit: http://www.ff123.net/training/training.html

Quality Level (-1)

Reply #5
I think there's a point where your speakers become the be-all and end-all of quality.  I have a set of $2 speakers (no kidding) that will make anything sound awful.  I don't think even you Dibrom could ABX a xing or ogg q0 from the original on them, they both sound like horrific mess.
< w o g o n e . c o m / l o l >

Quality Level (-1)

Reply #6
Quote
Once again, I disagree with this.  Listening equipment doesn't make all that big of a difference.  As long as you are using headphones, even cheap ones, that should cover the equipment side of things -- the rest is training and/or natural ability.  However, I believe that once you have been trained to hear these artifacts, you can even hear most of them on cheap speakers, at the very least in the case of using something like Vorbis at -1.

I agree with you.
Nevertheless, with good equipment, perceiveness is more easy. Even when you're trained...

I accidentally severly wounded my set of BeyerDynamic headphone. I tried to perform the AAC-128 kbps test my poor earbuds (Sennheiser MX500), and the worse artifacts of the test were difficult to hear. Some artifacts (temporal one) are not difficult to hear, because cheap headphone have resolution enougH to restitute them. But some others (I want to say 'frequential one') are really hard to detect in poor conditions. Though they are not impossible to perceive, I find it much more difficult with poor equipment. When you don't know how an artifact sound (alphaneutrino case apparently), I believe it must be near impossible to detect subtle one (but oggenc -q-1 artifacts are not really discreet).

Quality Level (-1)

Reply #7
Quote
Quote
Seriously, the listening equipment makes a big difference. alphaneutrino, if that's your situation you might consider spending a couple hundred to get a decent set of speakers.

Once again, I disagree with this.  Listening equipment doesn't make all that big of a difference.  As long as you are using headphones, even cheap ones, that should cover the equipment side of things -- the rest is training and/or natural ability.  However, I believe that once you have been trained to hear these artifacts, you can even hear most of them on cheap speakers, at the very least in the case of using something like Vorbis at -1.

Dibrom, I'm not trying to imply that artifacts are inaudible (or indistinguishable) on poor speakers/headphones. I fully agree that higher-end equipment is not a "magic bullet." I am, however, strongly implying that with higher-quality speakers the difference will be much more noticeable to untrained ears.

  Cost isn't a necessary factor - I've done some mastering on a $20 pair of "Labtec" headphones that got rave reviews among those who heard the results... and some of those people were "trained" to listen. (And I carefully refrained from telling them what to ignore, since I knew the source material inside and out, and could have pointed out enough patches to make their ears burn.) But most people are not as aware of the sounds they hear, since they do not as a normal part of their routine listen with critical ears.

  Alphaneutrino, if your initial impression was based on the speakers which came packed with your monitor (and I'm not trying to criticize, I'm just using this as a convenient example since you are the one who brought up the subject), then exposure to the same sounds on a higher-end system might make the differences more noticeable.

    - M.

Quality Level (-1)

Reply #8
Thanx for your replys everyone.  I checked out http://www.ff123.net/training/training.html and was able to notice some of the artifacts such as the lo-freq dropout and the chirping, but beyond that I was none the wiser.  With regard to what I'm using for listening, I use Grado SR-60 headphones (which were highly touted all over the Internet).  Anyway, is the situation that I'm deaf or am I just untrained?  What do you guys think?  I've noticed that in QL (-1) neither chirping nor lo-freq dropout occurs.  Would it be possible for me to post a one of my encoded files just for you guys to show me where you hear artifacting or is this against the rules?  Gimme some ideas on this if possible.

-- alphaneutrino

Quality Level (-1)

Reply #9
Another thing to consider, alphaneutrino, is that female pop music *generally* is easier to encode than something like metal, rock, or jazz. Though I don't listen to pop and have no idea who Kelly Clarkson is, I would be willing to bet that stereo-panning effects, drum solos, crashing cymbals, and acoustic guitars aren't the centerpiece of the album  Try encoding something like Black Sabbath, Rush, or Led Zeppelin at -q -1 and you will (should) hear tons of distortion on high-pitched sounds like cymbals, loss of stereo image, etc. And these things should be readily apparent, even without the aid of good speakers, etc., as Dibrom mentioned. You don't need expensive phones or speakers to hear a cymbal crash or thick, 1970's-style hi-hat turn into a metallic-sounding mess....it's very easy to hear at -q -1 or -q 0. (And before any Xiph Zealots jump on me, I'm not knocking Vorbis....it does amazingly well for 45kbps, but we can't expect miracles, you know!)

Quality Level (-1)

Reply #10
Quote
Would it be possible for me to post a one of my encoded files just for you guys to show me where you hear artifacting or is this against the rules?

Anything under 30 seconds is generally fine.

 
SimplePortal 1.0.0 RC1 © 2008-2019