Is Recency Bias fooling us all the time or manufacturers?

[Guddu]

Someone asked me this, I thought it’s a good one to have some opinion on from all of you…..

Is Recency Bias fooling us all the time or manufacturers?
Question is supposed to be generic!

Take an example- TPA3255 amplifiers
* Every new one coming out is better than previous one - advertisement!
* Many would replace their few months old one, which was best sounding for them few months back, with the new one
* Now new one is absolutely BEST sounding, better than others or previous one……… until another new one is released
* Mono is better than stereo…… but stereo is better value with same output and sound…… but still mono is better…… audiophiles win!

What do you think……
When will it become anything sensible!!!

 

[amirm]

Take an example- TPA3255 amplifiers
* Every new one coming out is better than previous one - advertisement!

There has been significant improvements in each version due to design optimizations.

Many would replace their few months old one, which was best sounding for them few months back, with the new one

I don't see evidence of this. I myself am running RME ADI-2 Pro which I reviewed back in 2019 I think. Once in a while I get asked if someone should upgrade from version a year ago to today and I always tell them no, unless there is a feature they like.

Speaking of features, there are enhancements on that front. We are now seeing EQ. Or even more mundane but useful trigger support.

* Now new one is absolutely BEST sounding, better than others or previous one……… until another new one is released

Members here are mostly careful to know that best measuring doesn't mean best sounding.

* Mono is better than stereo…… but stereo is better value with same output and sound…… but still mono is better…… audiophiles win!

What?

 

[Anton D]

It's a niche hobby to begin with, so I don't think it will ever become 'sensible.'

These are toys, they always get better with each new fantastic release. After a while, it's the oldest ones that get regarded as best!

Enjoy the circle, but feel free to step of at any time.

 

[JSmith]

mono is better

... for testing speakers, if that is what you mean.


JSmith

 

[Guddu]

Please Note - All the comments in original post are from different people asking and mentioning all that over a nice wine influenced gathering, I am just putting it across.
Cause of most of those comments are online and youtube reviews etc.

Those comments were just to put thorugh an example, original question is "Is Recency Bias fooling us all the time or manufacturers?"

==================
*** Just for the sake of clarity

There has been significant improvements in each version due to design optimizations.

While I somewhat agree, but significant improvement should be significant which might not be the case all the time.

I don't see evidence of this. I myself am running RME ADI-2 Pro which I reviewed back in 2019 I think. Once in a while I get asked if someone should upgrade from version a year ago to today and I always tell them no, unless there is a feature they like.

You are technically equipped and well-informed, and such a buyer is always different than many who just ask this question - will x2 would be an upgrade over my current x1. While some get an upfront answer but not all.
It is not a question about ASR reviews and recommendation in that review.
But when two similar products with similar stats/measurements collide then probably this question arise - "is it an meaningful upgrade over previous?"

Speaking of features, there are enhancements on that front. We are now seeing EQ. Or even more mundane but useful trigger support.

Agreed

Members here are mostly careful to know that best measuring doesn't mean best sounding.

I will have to disagree on this one, many might conclude differently on best measurements - intentially or unintentially!

What?

Online or you tube review spread things, these come as question as well as comments.

 

[Guddu]

... for testing speakers, if that is what you mean.


JSmith

For testing speaker- sure.
But not what I meant.

 

[kemmler3D]

Someone asked me this, I thought it’s a good one to have some opinion on from all of you…..

Is Recency Bias fooling us all the time or manufacturers?
Question is supposed to be generic!

Take an example- TPA3255 amplifiers
* Every new one coming out is better than previous one - advertisement!
* Many would replace their few months old one, which was best sounding for them few months back, with the new one
* Now new one is absolutely BEST sounding, better than others or previous one……… until another new one is released
* Mono is better than stereo…… but stereo is better value with same output and sound…… but still mono is better…… audiophiles win!

What do you think……
When will it become anything sensible!!!

I have a simple way to defeat recency bias, if you find yourself attracted to new and shiny things.

Only buy secondhand.

This way, at least by the time you get the new shiny thing, it's 30-50% cheaper.

 

[boxerfan88]

Another way to extract best value, buy the top end outgoing model when prices have been reduced to clear inventory.

For example, buying ESS9038Pro DAC (using the prior tech, and already well optimized) at clearance prices gives the best value, compared to chasing the latest ESS9039Pro DAC. Differences/improvement (if any) will be small.

 

[DanielT]

It is functionality that in many cases is the focus when upgrading. This shows itself most clearly regarding DACs. SINAD, the challenge in terms of audibility has long since been solved, so functions remain and the manufacturers listen to what the market wants. An example:

Amir:
This is a review and detailed measurements of the Topping D50 III balanced stereo DAC with support for parametric equalization.

Topping D50 III Balanced DAC with EQ Review

This is a review and detailed measurements of the Topping D50 III balanced stereo DAC with support for parametric equalization. It was sent to me by the company and costs US $229. Sorry for the quality of the picture/display. It is quite dim compared to my photography light and when I tried to...

audiosciencereview.com

I think that is a good development. The only thing I'm a bit concerned about is whether the pace of releasing new models leads to proper and rigorous endurance tests not being carried out? What about the durability?

When it comes to TPA3255 based amps, just wait until TI releases a new amp chip and you'll see an upgrade wave like no other. It will happen if the new chip provides opportunities for more power and a smooth implementation of PFFB to reduce, remove load dependency.
(that of course if the price is good plus sufficiently low SINAD)

Speaking of using old stuff. I have class AB based amplifiers, one of which is approaching half a century old and it works superbly well (recap done on it, I'll just add).

 

[rcstevensonaz]

I have a simple way to defeat recency bias, if you find yourself attracted to new and shiny things. // Only buy secondhand. // This way, at least by the time you get the new shiny thing, it's 30-50% cheaper.

Wise strategy! Moore's law is still largely in effect for anything in the digital domain where increased transistor density (i.e., compute power) is applicable.

To put in perspective for a DSP chip (and similar for DAC, etc.): consider the net difference from the first DSP chip ever released compared to the most powerful state-of-the art DSP chip that was most recently released. All other things equal*, the capability of the next round of DSP chips released 2 to 4 years from now could be 2x to 4x the delta from the first DSP chip to the current DSP lineup.

Personally, I don't like purchasing second hand since I prefer having the warranty. So the other alternative for digital based products is to (a) purchase at the end of model line when their final inventory is on close-out pricing discounts, and (b) purchase a few notches below that vendor's top-of-the line flagship product. Then, once all of those "flagship features" have been pushed down their product line in 3 to 5 years, repeat the purchase cycle.

* All things equal: the vendor's factories don't burn down or get flooded out, the vendor doesn't decide to "sit on new tech" for several years in order to milk profit out of their existing product line through minor incremental changes, the economy doesn't crash, etc.

 

[MAB]

Is Recency Bias fooling us all the time or manufacturers?

Not any more than Nostalgia Bias.
Records
8 track tapes
Tubes
VU Meters
All of these recently trended to popular.

 

[DVDdoug]

Personally, I consider myself to be practical and economical. I never heard anything wrong with my 1st CD player (of course it was a quantum leap beyond vinyl) and I never felt the desire to upgrade it. ...I still have it but I don't use it much... I wonder if it still works...

Back in the magazine days it used to bug me that they'd only review newly-introduced products and they'd have articles like "Best speaker of the year", with no comparison to last year's best model, which sill might be available.

Not any more than Nostalgia Bias.
Records
8 track tapes

Funny movie moment - There is a movie (made in 1990) called Spirit of '76 where time-travelers come from the far-future back to 1976. (They were trying to go back to 1776.) One of the future people bumps into a partially inserted 8-track tape. It starts playing. It startles her, and she asks, "What's that? They tell her and the guy says, "Don't tell me they don't have THOSE in the future!" It's a fun movie if you remember the 70's (nostalgic comedy).

 

[kemmler3D]

the capability of the next round of DSP chips released 2 to 4 years from now could be 2x to 4x the delta from the first DSP chip to the current DSP lineup.

Perhaps, but also, the capability home users might need from a DSP chip was technically a non-issue years and years ago.

The power of DSP chips isn't governed by Moore's Law, it's governed by whether the DSP chip manufacturers are feeling any competitive pressure on their current offerings, which seems to happen a good deal less frequently than it might in (say) GPUs or CPUs for data centers.

This is made clear when you consider that a low-end Raspberry Pi can handle as much (or vastly more) DSP as most of the purpose-built DSP chips. General purpose processors left DSP chips in the dust a long time ago, and mostly just need to be put on / next to a module with digital audio I/O.

I guess a long way of saying that if you're buying audio electronics with DSP capabilities, don't expect improvements year over year because of improvements in circuit density. At that level they have no direct relationship at all, only tertiary at best.

Topping only VERY recently released a DAC with 10 bands of PEQ on it, a DSP capability that's been available for pennies for many years and is likely present in any $30 pair of bluetooth headphones you buy - computing power is not remotely the limiting factor on things like that at the consumer level.

 

[Anton D]

I have a simple way to defeat recency bias, if you find yourself attracted to new and shiny things.

Only buy secondhand.

This way, at least by the time you get the new shiny thing, it's 30-50% cheaper.

I know a person who uses that as their dating style.

 

[Timcognito]

For TVs yes, for store bought food not so much (more packaging less food, more money)

 

[rcstevensonaz]

The power of DSP chips isn't governed by Moore's Law, it's governed by whether the DSP chip manufacturers are feeling any competitive pressure on their current offerings, which seems to happen a good deal less frequently than it might in (say) GPUs or CPUs for data centers. ...

Good counterpoint. Which is to say, the "All Things Equal" clause is the dominant factor since the chip manufactures are intentionally choosing to not take advantage of it.

 

[kemmler3D]

Good counterpoint. Which is to say, the "All Things Equal" clause is the dominant factor since the chip manufactures are intentionally choosing to not take advantage of it.

True, and there's also the fact that CPU power (on a DSP chip or otherwise) is not really the limiting factor for most consumer DSP applications, it's user interface and measurement that's a bigger hurdle. Any PC from the past 10 years can run as much room correction as you want up to probably 192Khz. It's just that setting up complex room correction (ala Dirac ART or Trinnov) is complex from a software development and UX POV.

 

[Tell]

There has been significant improvements in each version due to design optimizations.

Of course it has been improvements which your measurements shows, but are those improvements actually audible in practice? I'd say probably not for most people

I'm personally planning on upgrading my ten year old DIY IRS2092 to a DIY TPA3255 soon, mostly because of noise (when I'm a few dm from the speaker), but also because of feelgood of having an amplifier with low noise and distortion, and also because I want to build something.
So to answer OPs question I'd say no, I very rarely buy new "shiny" stuff, I've been cure from that affliction for quite some years now.

 

[pseudoid]

I had that recency-fad via peer-pressure that was also coupled with technical improvements and progress: I got sucked into Multi-Channel 'fad' when Audio wed Video. Dozen years of spending oodles on MCH hardware, my recency-bias has turned into buyers-remorse and now my A\V hardware are in divorce proceedings.
But if you want to call my trade from a 5channel Rotel amp to a 2channel Fosi a recency-bias, I will not argue you.

 

[goat76]

One thing I find fascinating and that I think suits the topic of this thread is the urge for the Meta technology in KEF speakers. It seems like many people can't wait for their favorite series of KEF speakers to get the "Meta-treatment", but can anyone, in all honesty, hear a difference that without a doubt is thanks to the Meta plates behind the drivers, or isn't it more likely that the new model is just voiced differently than the older version which causes the difference in sound?

And what will happen when KEF introduces the new and improved Meta2 material plate, will they sell their current Meta speakers for peanuts just as they did with their non-Meta KEF speakers?