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: LAME and GPSYCHO (Read 1742 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

LAME and GPSYCHO

Hello everyone!

I was playing around with LAME 3.93.1 (yes old I know) and playing with the --abr settings.

So I decided to try --abr 175 -h %s %d in foobar2000.

From what I understand in this version --abr or -b uses the GPSYCHO engine instead of the newer psychoacoustics model.

But anyway, I have to say, I'm quite impressed by the results.  I was shocked because I know this is an old encoder and a long abandoned psychoacoustics model.

But based on listening, I am trying to figure out why the GPSYCHO model was abandoned, because I think it sounds really good!

Am I missing something??/


Re: LAME and GPSYCHO

Reply #2
I don't mean any disrespect, but since I wasn't comparing it to the present psychoacoustic model, I didn't think I needed to cmpare or have some test results.

What I am saying is that it sounds really good to my ears and I was wondering why it was abandoned and not just improved upon as I think in its current state it sounds quite good, that's all.

Re: LAME and GPSYCHO

Reply #3
What I am saying is that it sounds really good to my ears and I was wondering why it was abandoned and not just improved upon as I think in its current state it sounds quite good, that's all.

Presumably the newer models were better.

Re: LAME and GPSYCHO

Reply #4
I don't mean any disrespect
I very much doubt @Heliologue was offended by it. It's just that, as you may know already, non-TOS#8-compliant, SQ-related claims like that, are usually politely challenged around these backwaters.
• Listen to the music, not the media it's on
• The older, the 'lossier'

Re: LAME and GPSYCHO

Reply #5
What I am saying is that it sounds really good to my ears and I was wondering why it was abandoned and not just improved upon as I think in its current state it sounds quite good, that's all.
I think that instead of "abandoned" we should read "phased out". AFAIK (and if memory serves me well) Naoki Shibata is one of the names that spring to mind as being responsible for the newer psychoacoustic model.

Since neither him nor Gabriel Bouvigne hang about here anymore, you'd have to rely either on other developers to chime in properly or on good olde Google for that.
• Listen to the music, not the media it's on
• The older, the 'lossier'

Re: LAME and GPSYCHO

Reply #6
If the intent is to figure out why GPSYCHO was replaced with a different model, the implied question is "How do you measure or compare psychoacoustic models?" and "sounds good" is, unfortunately, a meaningless label.  If "sounds good" means "effectively transparent at some desired bitrate" -- sure, I suppose it is.  And presumably, the newer psychoacoustic model achieved that transparency at a lower "cost" (compute, bitrate, whatever) as measured by blind listening tests.

You might find this old thread interesting: https://hydrogenaud.io/index.php/topic,51009.0.html

Re: LAME and GPSYCHO

Reply #7
It was about the move to VBR (APS, APE etc).  the new model was apprently more suited for that while the old better for cbr (Johnv).
The new model did bring  smart joint stereo , So that was a plus for all modes.

To get something similar with current (and  old) versions, You can try --abr 175 -f .  Imo high abr mode is a gem.  I even archive downloads  with --abr 200 -f --lowpass 17  . 

Re: LAME and GPSYCHO

Reply #8
You might find this old thread interesting: https://hydrogenaud.io/index.php/topic,51009.0.html
Indeed! Reading the OP's excuses to justify their placebophilia back then gives me fits already. Some things never change!🙄
Quote
It seems that my hearing is "tuned" to gpsycho , To me it gives the right sound.
Maybe it has something to do with the kind of music im listening to or maybe its the shape of my ears ?
• Listen to the music, not the media it's on
• The older, the 'lossier'


Re: LAME and GPSYCHO

Reply #10
It was about the move to VBR (APS, APE etc).  the new model was apprently more suited for that while the old better for cbr (Johnv).
The new model did bring  smart joint stereo , So that was a plus for all modes.

To get something similar with current (and  old) versions, You can try --abr 175 -f .  Imo high abr mode is a gem.  I even archive downloads  with --abr 200 -f --lowpass 17  . 

This is more what I was looking for, thank you.  I'm wasn't saying the curren psychoacoustics model in LAME is bad, in fact it sounds great too, but so does GPSYCHO, so I was wondering why it replaced. 

Obviously the newer model had benefits, but I just w asn't sure what they were.

Thank you everyone for the replies (even the sarcastic ones, I knew sojme would come!  Have a great day all :))

Re: LAME and GPSYCHO

Reply #11
@shadowking  - I'm just curious, is there a particular version of LAME you recommend overall for this?  I'm asking because I've read that LAME 3.98.4 is good, and so is 3.99.5, but I've read that 3.100 has regresstion in audio quality on these forums after some research?  What about 3.101b3?  Thanks! :)

Re: LAME and GPSYCHO

Reply #12
@shadowking  - I'm just curious, is there a particular version of LAME you recommend overall for this?  I'm asking because I've read that LAME 3.98.4 is good, and so is 3.99.5, but I've read that 3.100 has regresstion in audio quality on these forums after some research?  What about 3.101b3?  Thanks! :)

I use 3.100.1  3.99.5 is identical to 3.100 in quality.  3.98.4 also has the flaws.

I would try --abr 175 -q4 or -q7 with the current version to see if you prefer that or 3.93.1

 

Re: LAME and GPSYCHO

Reply #13
@shadowking  - I find I prefer 3.100 or 3.99.5 -q7 for --abr 176.

When trying the VBR modes I prefer -q4

I decided to switch to 3.100 since it's a more modern/recent version of the LAME encoder.