Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: 3.97b2 vs. 3.90.3: Kick drum impact? (Read 12147 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

3.97b2 vs. 3.90.3: Kick drum impact?

I'm not making a statement here, I'm asking a question (TOS #8)

Does anyone think that there is a difference in the impact of kick drums heard between 3.97b2 -V 0 and 3.90.3 --alt-preset extreme?

I've switched over to 3.97b2 as recommended here, done some ABX'ing in foobar, and am not sure if at this point it's just psychosomatic. But even when I'm not consciously aware of what version a file is encoded with, I'll notice a difference in how visceral lower frequency sounds are. When I go to check afterwards my perceptions are confirmed.

Lately I've been comparing many familiar tunes by transcoding from FLAC to -V 0 / 3.97b2  and --ape / 3.90.3.

On the whole it seems to me as though 3.97b2 -V 0 is more accurate to FLAC or wave originals. Which I suppose is better. But this difference in lower frequency sounds keeps on coming up. 

Can anyone please offer some technical input that discredits or validates this perception?

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #1
Without ABX logs, your suspicions are meaningless. If you say your percepcions are confirmed when you have this "feeling" and check the info, then you could very well complete an ABX session and post the results.
we was young an' full of beans

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #2
Why do I need a ABX log for asking a question?

I don't see the problem in getting a technical explanation why this perceived difference may or may not be possible. There wasn't an answer to the question in the searches I made on the subject.

Here's some examples:

e.g.

"Nothing has changed from 3.90.3 to 3.97b2 in the way lower frequency sounds have been handled" or "There were some changes to the psychoacoustic model that places higher frequency sounds perceptually closer to the listener, this may account for why they seem more pronounced than lower end frequencies. We realized that the presentation of the --alt-preset standards were not as accurate as they could have been to the original WAV files"

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #3
Quote
Why do I need a ABX log for asking a question?

I don't see the problem in getting a technical explanation why this perceived difference may or may not be possible. There wasn't an answer to the question in the searches I made on the subject.

Here's some examples:

e.g.

"Nothing has changed from 3.90.3 to 3.97b2 in the way lower frequency sounds have been handled" or "There were some changes to the psychoacoustic model that places higher frequency sounds perceptually closer to the listener, this may account for why they seem more pronounced than lower end frequencies. We realized that the presentation of the --alt-preset standards were not as accurate as they could have been to the original WAV files"
[a href="index.php?act=findpost&pid=373653"][{POST_SNAPBACK}][/a]


Its impossible to say without testing, so thats why ABX tests are required

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #4
Quote
Why do I need a ABX log for asking a question?

[a href="index.php?act=findpost&pid=373653"][{POST_SNAPBACK}][/a]



Because that answer is going to be that it's placebo effect unless you have proof that you actually hear a difference
"You can fight without ever winning, but never win without a fight."  Neil Peart  'Resist'

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #5
Quote
Quote
Why do I need a ABX log for asking a question?

[a href="index.php?act=findpost&pid=373653"][{POST_SNAPBACK}][/a]



Because that answer is going to be that it's placebo effect unless you have proof that you actually hear a difference
[a href="index.php?act=findpost&pid=373669"][{POST_SNAPBACK}][/a]


I see, thank you. I guess I was hoping for some technical basis beyond the placebo effect with which I could discredit the question. I looked through all of Gabriel's posts since 3.97 came out and couldn't really find anything that would settle the matter.

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #6
ABX and your questions may be answered.  If not, post the results.
"You can fight without ever winning, but never win without a fight."  Neil Peart  'Resist'

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #7
Its probably the 3.90 anxiety syndrome due to you switching versions.

To confirm your suspicions, just lower the quality level (v0-v2-v4) and the 'problem' should become more pronounced. Like wise going to -api should make it better. Then there is abx as others suggest. Easy.
wavpack -b320hhcs.5
lame --preset cd -f --lowpass 17

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #8
Quote
ABX and your questions may be answered.  If not, post the results.
[a href="index.php?act=findpost&pid=373744"][{POST_SNAPBACK}][/a]



Well, as mentioned I have ABX'd, between 3.97b2 -V 0, FLAC, original WAV, and 3.90.3 --ape. That was part of my testing before switching. It was after that point that I decided -V 0 was closer to the original FLAC and WAV in detail, if the viscerality of lower frequencies was not the same. Highs seemed more pronounced in exchange for lesser bass / kicks. I understand these kinds of subjective perceptions don't mean a lot here.

I'll record the results this time around but was looking for the input of people more knowledgeable than me about what changes have occurred between the versions, so that they could offer reasons for why I might be hearing what I am beyond placebo effect. Reassurance in other words.

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #9
Quote
I'll record the results this time around but was looking for the input of people more knowledgeable than me about what changes have occurred between the versions, so that they could offer reasons for why I might be hearing what I am beyond placebo effect. Reassurance in other words.
[a href="index.php?act=findpost&pid=373757"][{POST_SNAPBACK}][/a]

If you want people to answer your question on why you might be hearing is not placebo, you should first confirm that what you are hearing is not placebo. Otherwise this is just a waste of time.

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #10
Quote
...-V 0 was closer to the original FLAC and WAV in detail

HINT: FLAC is lossless and cant be anything else buy identical to the original wav file.

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #11
Quote
Quote
...-V 0 was closer to the original FLAC and WAV in detail

HINT: FLAC is lossless and cant be anything else buy identical to the original wav file.
[a href="index.php?act=findpost&pid=373825"][{POST_SNAPBACK}][/a]



Dazzling, thanks. What was meant that some tracks were compared to a FLAC copy of themselves, other tracks were compared to a WAV copy of themselves, etc.

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #12
Could you please tell me what you used as sample music?
Maybe even upload a sample or two?
Pusk is the new Start.

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #13
Quote
Well, as mentioned I have ABX'd, between 3.97b2 -V 0, FLAC, original WAV, and 3.90.3 --ape. That was part of my testing before switching. It was after that point that I decided -V 0 was closer to the original FLAC and WAV in detail, if the viscerality of lower frequencies was not the same. Highs seemed more pronounced in exchange for lesser bass / kicks. I understand these kinds of subjective perceptions don't mean a lot here.
[a href="index.php?act=findpost&pid=373757"][{POST_SNAPBACK}][/a]

I think you are avoiding the ABX question: Did you, or did you not ABX successfully between either
Lame 3.90.3/3.97b2 APE vs. WAV
or
Lame 3.90.3 APE vs. 3.97b2 APE ?

If not, don't be surprised if all your perceptions are put down to placebo.
The way you write, it seems that you didn't actually manage to ABX them successfully, but had some vague feelings about the sound characteristics. Please clear this up.
Proverb for Paranoids: "If they can get you asking the wrong questions, they don't have to worry about answers."
-T. Pynchon (Gravity's Rainbow)

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #14
Quote
Quote
Well, as mentioned I have ABX'd, between 3.97b2 -V 0, FLAC, original WAV, and 3.90.3 --ape. That was part of my testing before switching. It was after that point that I decided -V 0 was closer to the original FLAC and WAV in detail, if the viscerality of lower frequencies was not the same. Highs seemed more pronounced in exchange for lesser bass / kicks. I understand these kinds of subjective perceptions don't mean a lot here.
[a href="index.php?act=findpost&pid=373757"][{POST_SNAPBACK}][/a]

I think you are avoiding the ABX question: Did you, or did you not ABX successfully between either
Lame 3.90.3/3.97b2 APE vs. WAV
or
Lame 3.90.3 APE vs. 3.97b2 APE ?

If not, don't be surprised if all your perceptions are put down to placebo.
The way you write, it seems that you didn't actually manage to ABX them successfully, but had some vague feelings about the sound characteristics. Please clear this up.
[a href="index.php?act=findpost&pid=373889"][{POST_SNAPBACK}][/a]


I understand differently.
- He did ABX 3.90.3 vs. 3.97b2
- He could, beyond doubt, hear a difference between them.
- He could not tell which is which. Only that they are different.
- 3.97b2 is closer to lossless than 3.90.3

He asks:
Could there be a technical explanation for this phenomena, like the two versions of LAME behaving differently in certain situations?

I would very much like to test this for myself, if only he would provide some test samples.
Pusk is the new Start.

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #15
Quote
Quote
ABX and your questions may be answered.  If not, post the results.
[a href="index.php?act=findpost&pid=373744"][{POST_SNAPBACK}][/a]



Well, as mentioned I have ABX'd, between 3.97b2 -V 0, FLAC, original WAV, and 3.90.3 --ape. That was part of my testing before switching. It was after that point that I decided -V 0 was closer to the original FLAC and WAV in detail, if the viscerality of lower frequencies was not the same. Highs seemed more pronounced in exchange for lesser bass / kicks. I understand these kinds of subjective perceptions don't mean a lot here.
[a href="index.php?act=findpost&pid=373757"][{POST_SNAPBACK}][/a]


I don't think "viscerality" means much in general.  Viceral refers to your gut or stomach, or metaphorically to your gut feelings.  Do you mean how much you felt the bass in your stomach?  Or something else?

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #16
Quote
I understand differently.
- He did ABX 3.90.3 vs. 3.97b2
- He could, beyond doubt, hear a difference between them.
- He could not tell which is which. Only that they are different.
- 3.97b2 is closer to lossless than 3.90.3

He asks:
Could there be a technical explanation for this phenomena, like the two versions of LAME behaving differently in certain situations?

I would very much like to test this for myself, if only he would provide some test samples.
[a href="index.php?act=findpost&pid=373923"][{POST_SNAPBACK}][/a]


Thank you. This is closer to my experience than the suggestion that I have not ABX'd anything and made claims without testing.

It went something like this for me:

1. Hydrogenaudio recommends 3.97b2 and the new presets, including --vbr-new. I had been using 3.90.3 --ape up until this point so decide to investigate.

2. I downloaded 3.97b2, ripped some tracks from CD to WAV, then encoded copies to FLAC, 3.90.3 --ape and 3.97b2 -V 0, 3.97b2 -V 0 --vbr-new to compare them. A variety of tracks were used for this purpose, somewhere around 15-20. I ripped a few familiar favourite CDs whose sound I know well for comparison too. I used EAC, Speek's Multi-frontend, the latest version of FLAC and Razorlame for the various conversions involved. I compared both on my speaker set up as well as my headphone setup (Beyerdynamic DT770 Pros with a custom meta42 amp, driven off an M-Audio Sonica at 24 bit decoding in Foobar)

3. The first set of ABX comparisons was to determine which mp3s sounded closer to the original. I tested to see if I could tell a difference between original WAV / FLAC  vs. 3.97b2 -V 0, and original WAV / FLAC vs. 3.90.3 --ape. The general conclusion I reached from this is that 3.97b2 -V 0 was harder to distinguish from the original file than was 3.90.3 --ape.

4. The second set of ABX comparisons involved 3.90.3 --ape vs. 3.97b2 -V 0.  I could usually tell a difference between the two, in that higher end detail (melody, hi hats / cymbals, vocals) seemed better with 3.97b2 -V 0. Lower end frequencies, like how strong kick drums or bass lines sounded, seemed more pronounced with 3.90.3 --ape. It seemed to me that 3.97b2 favoured higher end detail to the original, while 3.90.3 favoured bass impact with the highs more rolled off. The original wav or FLAC files possessed both attributes (higher end detail, impact on kick drums and bass lines).

5. I concluded that overall accuracy to the original, as well as the (on average) higher bitrates that 3.97b2 -V 0 produced was probably smarter to go with. Not to mention the fact that 3.97b2 had been subjected to the scrutiny of so many here and that a recommendation to use it does not come lightly. I reasoned that if 3.90.3 --ape was "artificially" producing more lower end presence, it was still less accurate to the original file. Perhaps it is the case that 3.97b2 eliminated an imbalance that I had simply gotten used to. I didn't do enough comparisons on -V 0 --vbrnew vs. the others to validate switching to it.

6. I kept many of these files (encoded in 3.90.3 --ape, FLAC, WAV, and 3.97b2 -V 0) around and they would occasionally pop randomly into my playlist. Without knowing which encoding type was playing when, I would notice that bass seemed more pronounced with some files than with others. Upon checking, it would usually confirm prior observations.

The important point is this: I didn't take screenshots of the listening trials in 3 & 4, and given 5, I continued using 3.97b2 -V 0. But point 6 has come to bear a few times since making the decision in 5. Which is why I asked a question on whether or not it was possible.

I understand now that no one wants to entertain this question until I reproduce results observed in points 3 & 4. I still contend that asking a theoretical / technical question doesn't necessarily need the burden of proof that making a statement does, but I respect the philosophy behind the rule.

I think it is interesting that some have PM'd me to say that they have thought the same, but have not spoken up since the idea would be met with the rejection any subjective perceptions seem to receive here by reflex.

I think I will try to reproduce results in points 3 & 4 if it will elicit the technical commentary. I have been working long hours lately and usually feel too tired to analytically listen when I get home. I understand some see this as avoidance. But just as the placebo effect can work to taint people's perceptions, a legitimate question can be ignored if perceived as nothing different than subjective / baseless accusation.

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #17
Quote
I don't think "viscerality" means much in general.  Viceral refers to your gut or stomach, or metaphorically to your gut feelings.  Do you mean how much you felt the bass in your stomach?  Or something else?
[a href="index.php?act=findpost&pid=374022"][{POST_SNAPBACK}][/a]


When listening on my speakers, the impact of kick drums seemed more pronounced, in that they seemed to move more air from the subwoofer or be heard more loudly. On my headphones I found similar results. I am kind of reluctant to keep answering this question as it emphasizes the subjective things I noticed that led me to ask the question I did.

I wonder if it would have been better to have just asked "is there a technical difference in the way that low frequency sounds are handled in 3.97b2 -V 0 vs. 3.90.3 --ape", without saying anything about some of the reasons why I wanted to ask this question.

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #18
Quote
3. The first set of ABX comparisons was to determine which mp3s sounded closer to the original. I tested to see if I could tell a difference between original WAV / FLAC  vs. 3.97b2 -V 0, and original WAV / FLAC vs. 3.90.3 --ape. The general conclusion I reached from this is that 3.97b2 -V 0 was harder to distinguish from the original file than was 3.90.3 --ape.

4. The second set of ABX comparisons involved 3.90.3 --ape vs. 3.97b2 -V 0.  I could usually tell a difference between the two, in that higher end detail (melody, hi hats / cymbals, vocals) seemed better with 3.97b2 -V 0. Lower end frequencies, like how strong kick drums or bass lines sounded, seemed more pronounced with 3.90.3 --ape. It seemed to me that 3.97b2 favoured higher end detail to the original, while 3.90.3 favoured bass impact with the highs more rolled off. The original wav or FLAC files possessed both attributes (higher end detail, impact on kick drums and bass lines).

Sorry to sound harsh, but these 2 comments are useless (especially the statement and speculations highlighted in bold), unless you provide proofs that supports the statement "you really hear something different".

Quote
I think I will try to reproduce results in points 3 & 4 if it will elicit the technical commentary. I have been working long hours lately and usually feel too tired to analytically listen when I get home. I understand some see this as avoidance. But just as the placebo effect can work to taint people's perceptions, a legitimate question can be ignored if perceived as nothing different than subjective / baseless accusation.

A statement needs support in order to be treated seriously. Everyone is waiting.

Quote
I wonder if it would have been better to have just asked "is there a technical difference in the way that low frequency sounds are handled in 3.97b2 -V 0 vs. 3.90.3 --ape", without saying anything about some of the reasons why I wanted to ask this question.

LAME is open source, so if you want the technical details, read the source codes. If you don't know programming, then you are too far away to understand the technical details.

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #19
The condescension and redundancy is unwarranted.

Quote
A statement needs support in order to be treated seriously.


A question only needs curiosity, and some attempts to try and answer the question oneself before eliciting the input of others. If it pains you to even entertain the question, just skip it over. It's a lesser investment of time on your part than belabouring what has already been said and agreed with.

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #20
Quote
The condescension and redundancy is unwarranted.

Quote
A statement needs support in order to be treated seriously.


A question only needs curiosity, and some attempts to try and answer the question oneself before eliciting the input of others. If it pains you to even entertain the question, just skip it over. It's a lesser investment of time on your part than belabouring what has already been said and agreed with.
[a href="index.php?act=findpost&pid=374129"][{POST_SNAPBACK}][/a]

William is only trying to explain why nobody here has given you the answers you want. You explained your situation, and in great detail I might add, but the fact remains that you haven't posted any test results, you mentioned that you took these tests, and you mentioned what happened in each of these, but still nobody will be convinced of your claims until you post some ABX results, 21 posts in, and you are already defending your position. You also mentioned that you didn't take screenshots of the testing, you don't need to take screenshots, I'm sure whatever ABX program you used (foobar?) has some way of exporting the test results.
we was young an' full of beans

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #21
Quote
William is only trying to explain why nobody here has given you the answers you want. You explained your situation, and in great detail I might add, but the fact remains that you haven't posted any test results, you mentioned that you took these tests, and you mentioned what happened in each of these, but still nobody will be convinced of your claims until you post some ABX results, 21 posts in, and you are already defending your position. You also mentioned that you didn't take screenshots of the testing, you don't need to take screenshots, I'm sure whatever ABX program you used (foobar?) has some way of exporting the test results.
[a href="index.php?act=findpost&pid=374138"][{POST_SNAPBACK}][/a]


I understand and agreed already that tests are an understandable roadblock to moving along the discussion.

I will provide ABX test results when I have the time / fresh ears to do so. The rudeness and redundancy can hold off in the interim.

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #22
@ poorimpulsectrl : When you can, post short samples where you think this is noticeable. This would allow others to check if they have similar results.


@ William and skelly : So... tell me... if i just fake an ABX test (listening to two completely different samples) and post it, I will be believed more than if i *say* i did an ABX and how i did it?
We know that people come here sometimes claiming things because they know no better.
But comming from a 3 and a half years member, and user of HA recommended encoders and settings... it's rude to act this way.

(Edit: spelling)

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #23
@ poorimpulsectrl : I'd really like to listen to such a sample too.
lame3995o -Q1.7 --lowpass 17

3.97b2 vs. 3.90.3: Kick drum impact?

Reply #24
Quote
,Mar 23 2006, 09:45 AM]@ William and skelly : So... tell me... if i just fake an ABX test (listening to two completely different samples) and post it, I will be believed more than if i *say* i did an ABX and how i did it?
We know that people come here sometimes claiming things because they know no better.
But comming from a 3 and a half years member, and user of HA recommended encoders and settings... it's rude to act this way.

(Edit: spelling)
[a href="index.php?act=findpost&pid=374344"][{POST_SNAPBACK}][/a]

Yes, you're right, I overreacted and would like to apologize to poorimpulsectrl for being a hard-ass.

As for your question JAZ, ABX'ing different samples and using the results to validate a claim is rather sneaky, but proof is proof. And I'm sure anyone who does that will get banned

EDIT: Anyone who does that and gets caught!
we was young an' full of beans