audio-tech

[dull trout]

bro was very much into tube dac until i sell him my hiby fc5, bros never thought of tube since
i guess he liked hiby roll-off filter

[dapper dome]

Yes

To all of the above

Built in sounds maybe quieter but that’s it

[zenith pawn]

the dongle shouldn't sound any different between different computers, unless maybe the OS is messing with it somehow?

[lone flame]

he doesn't mean volume

[zenith pawn]

yeah i dont mean volume

[dapper dome]

Mega quality diff between 3 different machines using the same setup

[dull trout]

could be messy sampling rate

[lone flame]

Maybe apple has now "anti different OS" like Android diff

[zenith pawn]

i think the macbook built in output might use the same dac as the dongle but a slightly stronger amp

[dull trout]

or just usb meme negotiation

[zenith pawn]

it should sound almost identical but be louder

[lone flame]

yeah

[dapper dome]

Like, I would say my desktop gets 85% of the way of what I hear from the macbook

Maybe less

it sounds squished

like sample rate mismatch

I tried literally everything I could think of

[lone flame]

what's the sample rate of dongle

[dapper dome]

removed drivers, downloaded mobo specific drivers

[zenith pawn]

sample rate mismatch shouldn't really do anything as long as the sample rate on your output is higher than the source media/file

[dapper dome]

41k

[zenith pawn]

just use 48 khz on the dongle

[dapper dome]

It reports it can do 48k

sounds the same

16-24b

[zenith pawn]

ok thats what i'd expect

[lone flame]

some DACs are set up for one sample rate. so its possible that it samples up or down

[zenith pawn]

maybe some weird audio processing going on on the windows pc

[lone flame]

maybe

[dapper dome]

But it’s definitely worse on my Desktop, better on my wifes desktop, better yet on the mac

[dull trout]

or gae usb voltage negotation

[lone flame]

describe the diff

[zenith pawn]

i think upsampling is supposed to be essentially lossless (with the tiniest bit of added noise), but downsampling is supposed to be worse

so that's why you just want to use 48 khz on the output

[dapper dome]

Clarity, sound stage. The difference between sounding less clear than airpods pro vs full sound stage and SUPER clear tremble

actual proper bass not just rumbles

it sounds like it’s compressed

[zenith pawn]

i really would suspect some DSP accidentally being enabled somewhere in windows, or with some program you have installed

[dapper dome]

Like my m50x’s sounded better

[lone flame]

does like treble etc gets turned down when massive bass hits?

[dapper dome]

no

[lone flame]

ok so no ducking

[zenith pawn]

https://tenor.com/view/twitter-duck-smoke-smoking-twitter-duck-gif-1365401141084093497

[dapper dome]

I checked all configurables, toggled on/off everything

nothing changed

Updating drivers helped a lot

but still 85% or so

[lone flame]

drivers ALL up to date?

like ALL drivers?

[dapper dome]

Yeah

[lone flame]

even if its for the fucking printer

[zenith pawn]

i can't really remember what all you mentioned already
but did you try linux on the same windows pc (where you're having the issues)? i thought i saw something about linux mentioned

[dapper dome]

No, but I wanted to switch to linux due to having arguably easier time resolving driver issues

And window is just dumb with drivers

The motherboard is the asus x470-f

[zenith pawn]

although i don't think there should even be a driver at all for the apple dongle

it should just be generic usb audio

[lone flame]

no driver

plug and play

[dapper dome]

It does showup as generic, so yeah

[zenith pawn]

so that part shouldnt really be a big deal

[lone flame]

Driver would make it x10 price xD

[dapper dome]

But it still used onboard sound processing

even with the apple dongle afaik

[lone flame]

not possible

bcs. its digital

it takes the shit directly from the CPU

[dapper dome]

So is it possible my dongle is just borked?

[lone flame]

no

if it works on other devices perfectly fine?

[dapper dome]

I guess it works on my mac so

lmao yeah

[lone flame]

so your PC is just borked

or atleast something

[zenith pawn]

i can't really imagine this being anything but a software issue

i don't use windows so not really sure what to try

[dapper dome]

so, I do have voicemeteer potato, nvidia broadcast and virtual audio adapters installed

[lone flame]

it can be Software / Firmware

[dapper dome]

which I wondered about the extra overhead

but driver refresh should ahve resolved those issues

[lone flame]

Voicemeeter

Do you have an AMD CPU?

[dapper dome]

And those aren’t tied to anything

[dapper dome]

Yez

[lone flame]

i know that Voicemeeter doesnt really like AMD

[zenith pawn]

stereotypical amd cpu issues (it's probably not)

[lone flame]

idk if they fixed it

[dapper dome]

indeed it does not

[lone flame]

AMD x Voicemeeter = no

atleast some years ago

idk about now

[dapper dome]

It works for a lot of things, just not all

It can’t do high samplerates

So if you drop say your mic quality down it’s fine

$500 mic running 16bit 41k

[lone flame]

and thats fine

[dull trout]

could be software, ive faced this before with windows when connecting to a hub/directly to board usb c. somehow volume is lower than connecting to the phone so i just reset my windows and problem solved (for now)

thought my usb dac was cooked so i got another one out of spite and it still does the same shit

[dapper dome]

This is a fairly new install, less than 2 months

[lone flame]

if u want we can hop on a call to try an troubleshoot

[dull trout]

no way to find out unless you try but i guess this aint worth the time lol
might as well just cope with it

its not that bad yea?

[dapper dome]

Fun fact, it is nearly 1am for me and my typical workday starts at 5am, so maybe not tonight- but I’ll definitely take-up your offer another day/night

[lone flame]

yeah go to sleep

maybe its all placebo

[dapper dome]

I could just use a secondary boot, I just am fairly certain it’s the onboard chip being literal dog

[dapper dome]

[dull trout]

yeah could be that also, or usb being usb
sometimes good sometimes shit

[dapper dome]

when my non audiocaring wife noticed a difference when I showed her (without mentioning it) I knew I was cooked

[lone flame]

ok

bro got wife

[dapper dome]

And a child!

[dull trout]

nice life

[lone flame]

bro reproduced

[dapper dome]

twenty three year old reproducer

I could just steal my wifes motherboard. Same generation of AMD

[lone flame]

bro is 2 years older and is Level 100 Mafia Boss
Im just a Level 1 Crook in life...

[dull trout]

lmao not like she would notice

[lone flame]

i just have a work
and life with my parents bcs its cheap

[lone flame]

dont think it would change something

[dull trout]

if hes lucky, its just windows, if hes not, its the board

[dapper dome]

I tried on her PC before the mac, that’s how I knew it was my desktop not the headphones

[lone flame]

t the..

https://tenor.com/view/board-break-karate-regular-show-gif-20602208

[dull trout]

or the ear

who knows

[dapper dome]

Could it be BIOS?

🤔

[lone flame]

wouldnt make much sense

BIOS has nothing audio related

iirc

[dull trout]

dont think so, as far as things go bios only controls power up and down

for usb*

[lone flame]

Windows 11?

[dapper dome]

Well, if I’m on an old bios and latest audio drivers it’d screw with something surely

yeah

[lone flame]

so Win 11

uhh

uninstall Realtek

trythat

[dapper dome]

made it worse

[zenith pawn]

usb has a standardized voltage so you wouldn't be changing that

[lone flame]

Windows 11 had 2 or 3 times issues with Audio

[dull trout]

yeth
not that you can in the first place

[dapper dome]

Only after installing the official x470-f board drivers did it even reach 85%

prior was more like 45-50%

(Just raw windows audio drivers, no mobo specific)

[lone flame]

uninstall every audio drivers

and use dongle

try that

[dull trout]

uh, a more practical (read: costly) solution is uh.... 4X AKM DAC WITH 8 STAGE AMP

[dapper dome]

https://rog.asus.com/us/motherboards/rog-strix/rog-strix-x470-f-gaming-model/spec/

[lone flame]

https://www.magnahifi.com/holo-audio-cyan-2-dac-nos-r2r-dsd1024

get that

gud dac

[dapper dome]

I’ll send my pj over

[lone flame]

its R2R

[dull trout]

dac cost more vs my whole audio setup combined

[dapper dome]

I could probably upgrade everything but my GPU for that

[zenith pawn]

oh cool it's only 1.147,93 that means it's only 1 and 14 tenths euro

[dull trout]

real

me when reading credit card statement lmao

[dull trout]

eh could get snowsky melody

i dont personally like it, but its a decent dacamp usb

but perhaps solving the actual issue would help, if you ever find one

[dapper dome]

I picked up a cx usb c dongle which runs at a higher power than the stock apple dongle

which should theoretically clean up the power moreso

I would bet my balls it’s purely just very dirty power

[zenith pawn]

if that doesn't fix it then it's definitely software

[dapper dome]

Either that or yeah

drivers

Could be bother

[zenith pawn]

idk if there's such a thing as dirty power

[dapper dome]

I’ll also update bios

The other thing, which I guess I didn’t mention

[dull trout]

windows being gae part 150million probably

[dapper dome]

Prior I ran my m50x’s off my massdrop ctrl keyboard’s secondary USB-C port

which helped those specific headphones a crapton

didn’t do jack all for the ft1s

[dapper dome]

Another reason to switch to Linux

unfortunately might be worth

All my development environments would actually run properly on Linux instead of using a secondary system to do tests lmao

Or WSL. Second bane of my existence, windows being the first.

[sick terrace]

the dongle is a tiny amp/dac. if it sounds different on the macbook then there's something happening on the software side of things that's changing the sound. buying another amp/dac isn't gonna do anything. also fwiw using the apple dongle on any computer bypasses its internal dac/amp. plugging the headphones directly into the 3.5mm out on the macbook would sound even better than the dongle

[sick terrace]

what are you listening to music through?

could be the quality settings or an eq in the app you're playing from

[lusty sparrow]

Hii, i just lost my cmf buds pro 2, whats the best earbuds i can get for £50 max uk? Or £62 if they come with tracking as i would need to buy a smart tag 2 so i dont loose them again

[dapper dome]

Spotify

[dapper dome]

They’re downloaded at the highest quality Spotify will let me

[dull trout]

what store would you prefer?

[dapper dome]

This was what I assumed, but if that were the case then updating drivers shouldn’t have changed anything unless the headphones were plugged directly into the 3.5mm jack?

[lusty sparrow]

Amazon or argos ideally amazon tho.

[trail whale]

update
dude after replacing a cable connected left channel to R and L and right channel to R
and could not find it out 4 times
my MOBO is perfectly fine for pushing dt990 600ohm btw 🙂

no issues with listening music now at pretty koud levels (after he fixed it finally after fifth RMA)

[dull trout]

https://www.amazon.co.uk/Moondrop-Space-Travel-TWS-Earphone-Black/dp/B0CKW5JY9Y/ref=sr_1_5?crid=1ZVUTJPEGVUVN&dib=eyJ2IjoiMSJ9.Nkt13XiKJZBJN5fiZoh2w1QTrr-bjNJLGo4HAbq8uOrbaqaQGeKnuKlIasr_zq3l_lCZs5hWaqXD5MANbhQzksju3y-YfA6wHKGoS2rqWX39W_G0DjxArKakaSBX3gqYrSI9r9yqQoPIzstwQtXEdlUtp5NhLxb3hfsq2BNAOHeRNDobAZa2_P2-AXS6Z8B6Lz2U7GMEN0frbqWJR-P1ka76X8PkKSH-sO1ZT4fhnRM.QPGdQGQ1FUOkjMzLCQm3bAmP8bca7CNVGENPA6F6zpg&dib_tag=se&keywords=tws&qid=1756909424&refinements=p_36%3A2000-4800&rnid=389035011&sprefix=tw%2Caps%2C205&sr=8-5

https://www.amazon.co.uk/Anker-Soundcore-R50i-Bluetooth-IPX5-Water-Black/dp/B0C9MNT48D/ref=sr_1_21?crid=1ZVUTJPEGVUVN&dib=eyJ2IjoiMSJ9.Fn9IpWgf3QgMePsTKk1jEA1F6JhSLddln0CkvdqQ5R5lMU1W4XtskLtpC1o9fmV5deTnrVfLbZEFut5WTUCGJQUjKqLb67FUCmMd9Wiu5eWbYsbwBAGM1Hb0S3425K6Ncxi9QO8OFErFEc-tsfOHqxFHSQ9m2CDhYovq1DrQ0k9oP63QJFGoNnbcvWn8SUN2jJyzxUO2evsap5A5QRNqJksFC6Dl5mj2Ra911Z4cddc.XxPF3R95tMS586Fz8bDC1makhkTXptPAxER_DMtg9n8&dib_tag=se&keywords=tws&qid=1756909455&refinements=p_36%3A2000-4800&rnid=389035011&sprefix=tw%2Caps%2C205&sr=8-21&xpid=r0HuvUkVrY75W&th=1

https://www.amazon.co.uk/SoundPEATS-T3-Pro-Cancelling-Waterproof-Black/dp/B0F1MW58H1/ref=sr_1_51_sspa?crid=1ZVUTJPEGVUVN&dib=eyJ2IjoiMSJ9.9dTt1HhupQf7iVJBtOTD5JinJ2eEgDogKWkGfMRUoWdbIQ3IJYmvR9JVd5QFrbd3nJoSJfEomFKeU94KvocDV-RB-wMhd79SB1hWTEbMh3GaaYBnlDDbP4VSrtmnvddmFXz2HakeDxYSY6_I8rmrlfZFEXVErdr3gzY1c_Lzg6wmJ6wjuyQ3bfTaFHXuXjgj.U3JJy5YxKqOC1BurAW3ekkbxFLWx7smI5CVymK_RJmQ&dib_tag=se&keywords=tws&qid=1756909492&refinements=p_36%3A2000-4800&rnid=389035011&sprefix=tw%2Caps%2C205&sr=8-51-spons&xpid=r0HuvUkVrY75W&sp_csd=d2lkZ2V0TmFtZT1zcF9tdGY&psc=1

you can pick one of these. my personal recs is the space travel

theyre all quite decent, the space travel definitely edges your old cmf buds pro 2 tho

[sleek lily]

@broken grotto I'm technically still in the rectifier and clamper circuit part of my electronics class, but I'm doing some reading ahead for MOSFETs, and it seems the active region there is about a 1-5% variance from linearity that's expected and amplitude dependent. Could that be a valid explanation for the about 3% variance I saw in gain ratio for the amplifier I tested? I haven't technically gone over this stuff formally yet, so I'm trying to understand it better.

[lusty sparrow]

What about anc? How are the on board controls? (Do they have any) Do they gave apps for control and do tgey use ldac

[lusty sparrow]

Im also wondering if i can get something with in ear detection

I really miss my cmf buds but i got them as a gift and can barely afford new ones

[dull trout]

anc is alright, space travel is hard to beat for their price

and they do have decent on board controls, they use ldac, and have an app

iffy but its there

[lusty sparrow]

Alrightt, what about ied also how come the space travel 2 arent available here?

[warm scarab]

ldac?

is this space travel 2

[dull trout]

oh yes my bad

ldac is space travel 2

[dull trout]

some correction it doesnt have ldac but that shouldnt matter

probably import

the space travel 2 are still relatively new

[lusty sparrow]

10 quid more for the 2s worth it????

@dull trout @warm scarab

[lone flame]

Space Travel 2 are better

[dull trout]

heck yes

[dull trout]

in the store he provided, there isnt any stock

apparently he found another so thats fine

[warm scarab]

i dont know, i also wonder the same thing

[lone flame]

it is woth

[warm scarab]

they are really recent

[lone flame]

yes

its straight up Space Travel but better

[zenith pawn]

It probably changed some software processing. the "driver" may also have software doing DSP

[broken grotto]

I like my stack of schiit

If anything else it gives a nice spot on my desk to plug headphones into instead of having to buy a longer cable.

(And sounds better than the pc motherboard audio.)

[broken grotto]

Do you even know for sure that that amplifier uses MOSFETs?

[sleek lily]

I haven't had a chance to take the amp apart. My fiance uses it so she'd probably be upset if I did that lol

I need to get more amps eventually

[dapper dome]

Yeah Idk if I’ll regret it but I’ll listen to the wiser than I

I do think he is right

$120 for a Modi/Magnum stack

[broken grotto]

Amp model and maker?

[sleek lily]

Monoprice liquid spark

[broken grotto]

so, it's using both JFETS and MOSFETs.

but that variance you see is still measurement-related.

you even saw that with your sine wave tests.

if it was nonlinearity, you would have seen the same with the sine.

[sleek lily]

Both saw about the same variance of 3% in gain if I'm remembering right

[broken grotto]

bruh.

do the math here. make a table.

just to clarify

now you do it for the square wave.

[lean grove]

%diff should be 2-3 sig figs

SMH my head

[broken grotto]

apologies

i think the point still stands tho lol

bottom line, at this point im starting to think you're chasing a unicorn to say, "ah-HAH! I'm right!" but, cmon , 8mV?

[lean grove]

Is that percent difference in the margin of error

[broken grotto]

if you want me to be honest i think the values he was getting was just dumb luck

I expect on really good equipment and skilled operator you'd see pretty consistent unity gain.

[lean grove]

I just don't have a frame of reference for electronic measurements

1-2% might as well be considered identical in my field

[broken grotto]

in other words, yes

I am fairly certain those measurement values were pretty well within the scope's margin of error.

if you've never used a scope before, sometimes (especially on cheaper ones!) they're pretty sensitive to just about anything

you can breathe on it and get an extra 10mV if you really wanted to.

[lean grove]

Or if the room were a few degrees hotter

[broken grotto]

I am all for electronic discourse but I'm pretty sure it was pretty definitive what it was between Golden and I lol

I even was able to recreate the same thing with Multisim, which I know he can do himself.

so at this point im fairly certain it's just chasing a unicorn.

[lean grove]

I've known biooc for many years and he has always found creative ways to be wrong in the face of overwhelming evidence.

So this is just who he is.

[broken grotto]

its fine to be wrong when learning, that's how we learn

but holy cow accept when you're wrong and figure out why you were lmao

especially without a good foundation in analog electronics like transistor power amplifiers and nonlinearities, it's not gonna be a screaming answer on google.

Golden explained it pretty well, if it was a real behavior, then it would have been documented by now, especially in the audio space.

audio is relatively low frequency. I spent most of my analog electronics stuff in radar, which is a much different beast than the low frequency audio; you're looking at X-band and higher (9+ GHz).

weird stuff happens at those much higher frequencies, but those don't often translate to audio because it's just too low of a frequency.

(for reference, take the 20kHz cap on audio for humans. A full wavelength antenna that would pick that up would be 15 kilometers long!)

some audio players use the headphone cable itself as an antenna for FM radio

[lean grove]

Those were the good old days

I feel like a lot of people mystify audio equipment a bit too much.

Some of the best designed amplifiers are a handful of parts.

Things get more complicated when you want an amp that's a few hundred watts but other than that

[broken grotto]

and for that, it works because most headphone cables are about, idk, 1-1.5m long?

which is riiiiiiiight about where FM radio's half-wavelength is.

[lean grove]

Is the half vs full wavelength and efficiency thing

How low can you go realistically

[broken grotto]

1/4 wavelength is pretty common.

[broken grotto]

yeah, full wavelength antennas are most efficient and it decreases as you go smaller.

a half wavelength is most common

well

full wavelengths have their issues

[lean grove]

Other than scraping the underside of bridges?

[broken grotto]

yeah, radiation patterns and stuff like that.

full wavelengths are very good at converting current to radiation

but the downside is for something like FM, an antenna is pretty heckin huge.

electromagnetic radiation... not the other kind

the music-through-air radiation

or comms like CB radio

or HAM if you're into that.

you'll see quarter-wavelength antennas a lot for FM radio too, which are about... 2.5 feet long?

are you an audio engineer?

[lean grove]

I see ones that are about 5-10 feet on some trucks

No I'm a chemist

I took some analog electronics classes to help with instrumentations

Never got beyond transistors, RLC circuits, and op amps.

[broken grotto]

once you're past that stuff, you're into the weeds so

neat

[lean grove]

It's really all I need. All of my instruments are photodiodes, amp meters, voltage meters, and wheatstone bridges with some fancy stuff ahead of them.

[broken grotto]

I did a lot of analog and digital circuitry along with DSP for radars

it was very systems heavy

lots of cool projects tho

[sick terrace]

do you have spotify's eq enabled?

whatever is happening is definitively a software issue somewhere on your computer. the fact that the same, dac, amp and headphones are giving you a different output across devices is all the confirmation you need

[lone flame]

Statement that describes what science is.

like. this is literally science

you have a theory, try it. either youre wrong (mostly) or you're in the right and maybe wonder why? xD

then you find the reason why you're in the right or wrong and of that you can learn more, adapt what you have learned. apply it somwhere else etc etc

[warm scarab]

with the apple usb c?

yeah thats normal

[dapper dome]

I guess I just don’t feel like it’s as good as my prior headphones at the current moment with the existing setup

whereas if I replace the ft1 with the m50x’s , everything else made equal- the m50x’s sound better

but if I move to Macbook pro the ft1s beat it soundly

not even comparable. Ft1s are very nice

[dapper dome]

Yeah, but it doesn’t sound any different toggled on/off assuming a flat EQ

[dapper dome]

Yeah I think you’re right

I’m gonna try through my keyboard on a different USB port

Only tried through the keyboard with the USB-C adapter on a 2.0 port, which probably isn’t ideal

[dapper dome]

@lone flame It was caused by depop being enabled in BIOS. This motherboard shares the filtering logic for the 3.3v/5v rail to prevent/help voltage drops- essentially making all USB power not great for audio

Now it sounds as good as my wife's desktop, but still probably not as good as the Macbook pro

Enabled by default

[lone flame]

bruh

[lean grove]

astonishment

windows fucking sucks

[dapper dome]

I think in this instance it was asus but same diff

[warm scarab]

i think it might just be a volume thing

get a dB meter app

[stuck bolt]

Laughs in Mac

[warm scarab]

do a test tone at 1khz

[stuck bolt]

[warm scarab]

try your best to approximately put your phone or dBmeter at the same distance your ear would be

get the same volume level from both headphones and take note of what volume % this happens in

and then compare the headphones that way

a lot of people compare easy to drive headphones to hard to drive headphones without volume matching and just pick the louder one

same with dacs and amps too

[dapper dome]

It wasn’t a volume thing

also I misremembered which headphones I had prior

They were the m40xs

I feel like you just don’t believe me when I explain the differences. You’re essentially telling me what I’m hearing is placebo and I’m gaslighting myself

Don't get my wrong, I really appreciate the help picking the headphones and generally the feedback- but I feel like you just straight up dont read 80% of the prior conversation or what I'm explaining. Which drives me nuts lol

I'm happy to anwer questions (Repeat or otherwise), happy to give information to help problem solve. I am not happy to essentially be told that what I'm hearing is wrong and I just need to get over myself and what my first impressions are

[lean grove]

its okay dark, klaus is the local clown of the channel

don't take anything he says seriously

https://tenor.com/view/kung-pow-kung-fu-trained-wrong-wimp-lo-gif-27152837

[warm scarab]

do it regardless

also wtf i didnt know we trained nils ai model to be able to send gifs

nils is a failed cleverbot instance we decided to put in the audio chat for spreading misionformation because we found it funny btw he isnt a real person

[dapper dome]

Compare audio -> get them close to the same loudness -> Record % = use that volume when comparing the two?

[warm scarab]

yeah

[dapper dome]

Aight, will do

[warm scarab]

use a web tone generator

at like 1khz

[dapper dome]

the cx dac/amp usb-c adapter comes tomorrow too, I don't actually expect a super insane difference now that my stupid BIOS isn't screwing me but I'm definitely excited. I had a super cheap apple knockoff one that I tried a little bit ago today and holy crap was that horrible lmao

[zenith pawn]

Klaus when someone hears something

[dapper dome]

I am bias towards u-green so I opted for theirs over something probably cheaper/better

[zenith pawn]

they're probably almost all the same conexant DACs anyways

[terse river]

hey guys, im sorta interested in buying a record player, any recs? idk how pricing really works for these, is there really a good point to getting the ones that cost thousands? i wouldn't know

anyway i am probably looking to spend less than 400, but if there is a good reason to go higher ig i could

[warm scarab]

it literally doesnt matter, just make sure it has a counterweight and decide beforehand if you want it to have a phonopreamp built into it or not

[terse river]

whats a phonopreamp?

[warm scarab]

95% of the sound will come from the stylus you put on the turntable so thats the only part that actually needs to be good

[civic beacon]

Don't blindly trust ANY review website, but... this is a good place to start if you want to learn what you should look for in X thing. https://www.nytimes.com/wirecutter/reviews/best-turntable/

[warm scarab]

turntables provide sound in a way where bass is very silent and highs are loud, a phono preamp makes that into a normal sound signal

[civic beacon]

https://www.reddit.com/r/audiophile/comments/10uj21u/can_someone_explain_phono_preamps_a_little/

[warm scarab]

just go on ebay

buy whatever turntable that you think looks nice

with the condition that it has an adjustable counterweight

and maybe an integrated phono preamp, in which case a tt made more recently might be better

if you have a speaker amp there is a good chance there is already a phono input

[terse river]

yeah im using a denon x2700h

oh yeah it has one

ok so ig thatl do that for me

[warm scarab]

yeah you dont need an integrated phono preamp

make sure whatever you buy either has no phono preamp in it or has an option/socket where its not active

actually fuck that it doesnt matter

if it does have a phono preamp you cant turn off you can just use literally any of the other rca inputs

[terse river]

great, so all that matters is that it has an adjustable counterweight then?

[warm scarab]

yeah

[terse river]

cool

ty

[warm scarab]

what you need to research is just what stylus you want

and maybe some accessories to help you set up the cartridge/stylus in a flat manner

[terse river]

theres a difference?

[warm scarab]

yeah thats the entire difference

[terse river]

i thought they all jsut used piezzos

[warm scarab]

and its a huge difference

no

there is also the entire low compliance vs high compliance tonearm debate

but you dont need to get into that

[terse river]

does the type of music i listen to matter?

[warm scarab]

its subjective

perhaps

there is conical, elliptical, microlinear etc

[terse river]

ok im gonna watch some yt this clearly is a bit too in depth for a discord rec lol

[warm scarab]

and within those there are different quality ones

[warm scarab]

go ask in sonic visions

there are some good tt nerds there

you can also ask them for turntable advice but it largely doesnt matter what you get

but they might recommend some bang for buck tt with good suspension and easy to replace dustcovers etc just for convenience's sake

[sick terrace]

the sound that's produced by the little needle/stylus that rides the grooves of the record is extremely quiet. a preamp is boosting the volume of that signal while introducing as little electrical noise/white noise as possible. that signal then runs into an amplifier before being output by speakers or headphones

[viscid tangle]

Thoughts on a Aune ar5000 + Fio k11? I don't really have a budget, I'm just looking for an upgrade from my He400se + apple 3.5mm to type c

[steel escarp]

I'd go for a Fiio FT1 Pro

[lone flame]

question is

what do you expect form a upgrade?

its not like. investing more = better sound

its more likely. try around. find out. with luck you find what you like

[viscid tangle]

I was not a fan of the he400se stock, I did an eq on it. Stock just sounds thin. Vocals sound like they're trapped under instruments rather than their own thing. The EQ for me adds thickness to the bass and makes the vocals pronounced.

So I guess I'd like something with more bass and better with vocals.

(This is my eq graph for the he400se) ^

[umbral charm]

Hey everyone, I need some advice on managing and EQing a pair of headphones. I’ve got Sennheiser HD600 open back and HD620S closed backs, and I want the 620S to sound as close as possible to the 600s. I’ve never done EQ before (in this scenario anyway) and I’m not really sure where to start. Someone suggested using Equalizer APO with the Peace GUI and maybe starting with an existing HD620S preset, then tweaking by ear. But again, having not done it before seems like alot of stress as it's quite important.

On top of that, I want an easy way to switch between the two headphones. Right now they both plug straight into my PC (Having to swap them out as I need to) I’ve heard people mention using a USB plus headphone jack setup, a passive 3.5mm switch box, or a DAC and amp with multiple outputs, but I’m not sure what would be best or which DAC/amp to get.

So basically just looking for how to EQ the 620S to get closer to the 600s and what setup would make switching between them simple without constantly unplugging. Any advice would be appreciated.

[green marsh]

I dont think ar5000 fits your criteria there

I would consider ft1

[lean grove]

Hmmm. If you're into vocals, maybe one of the Sennheiser would be better.

I'm always an advocate for the 6xx, but it doesn't have more bass than the 400se. It is warmer, but not bassier. But really good vocals.

Hd560s or hd550 might be worth checking out.

[viscid tangle]

Sennheiser's sound? Would something like the hd490 pro be good? i just want something that would be a good gap from me he400se. I guess the best option would be to abuse amazon's return policy and try a few.

Seems like the hd490 has a little more bass

[lean grove]

490 comes with different pads. One of the pads adds a decent amount of bass.

But that 2khz dip I think is what you're complaining about with the vocals on the 400se

And it's also on the 490 pro.

Sundara might also be an option. It has better bass extension and doesn't have the 2k dip that most other hifimans have

Better headband too

[viscid tangle]

Ok I will look into it, thank you.

[lone flame]

Hard to accomplish

Or lemme say. Nearly impossible

Firstly. HD620 is 500series chassis

HD600 is dirrerent chassis. Larger ear cups more space inside etc

Also it's closed vs open

So Acoustic impedance diffs alot

You could archieve a somewhat similar tuning

But there are toany factors changing the sound...

[lean grove]

You can use AutoEQ to use the hd600 as a target

But at that point why not just pick an actual target

[viscid tangle]

Yep I see

[umbral charm]

I get what you mean about picking a target, but for me the 600s are my reference that’s why I want the 620s tonality to be closer to them. Where’s a good place to start with that, like what software and how to actually do it? Also, what DAC/amp would make sense for this setup, and how can I make that the EQ I apply to the 620s doesn’t affect the 600s?

[lean grove]

There are dozens of videos online on how to use autoEQ. Headphones.com and crinacle have the best videos imo so start there.

There's a database called squig.link that has AutoEQ integrated it into and it's what most people use. You use this website to create an EQ profile for a headphone and export it to your software of choice.

The software for PC is called equalizerAPO with peaceUI. Its clunky and annoying. There's probably a few other softwares that can do this too.

If you plan on getting a DAC/amp, it makes sense to get one that had hardware level EQ built into it. Something like a qudelix 5k or a qudelix t7i would be a good buy. They have an app that had AutoEQ built into it so it'll be as easy as opening the app and switching profiles.

Topping dx5 ii also has hardware EQ with profiles and has way better specs than the qudelix options.

Because theres no real way to have a headphone amp know what headphones are plugged into it, you will have to switch between profiles somehow. Or just buy two devices that are set to that eq at all times.

[viscid tangle]

I was recommended to get the atom amp+ and pair it with my apple dongle? Or should I just get the fio k11. I'm going to go for the hifiman sundara for my headphones.

[umbral charm]

Hmm. I was less aware than I thought I was then. In that case I might just return them, seems more hassle than anything to have them. It wasn't a necessity just a bonus.

In that case, what amp/dac should I get for the 600s? 😂 Sorry for being such a pain.

[lean grove]

Either is a good option. The atom amp is pretty good. It's downright one of the best amps for high sensitivity stuff like iems. The dongle will be holding it back in terms of power, since it only has 1 volt output. Should still be plenty for the sundara.

K11 is also really good value and has a digital volume knob, which means perfect channel matching across the entire volume range.

If you think you'll be upgrading soonish or listening to sensitive stuff like iems, then get the atom+dongle. If you want something simple that just works, get the fiio.

[lean grove]

How much do you wanna spend and what are your priorities?

[umbral charm]

I’m not entirely sure what an amp/DAC does or what I should look for in one. I just know they’re supposed to be beneficial. I produce Dancefloor DnB and I might add studio monitors in the future (not sure if that changes what I should get). I don’t have a fixed budget, so I’d be happy to hear a couple of suggestions at different price points. Maybe something that’s also future proof if I do get monitors later.

[lean grove]

https://youtu.be/PqlcJnoIcGo
https://youtu.be/HOvubnenXyo

If you're doing pro audio, your needs might be different compared to the consumer hifi stuff that most people in this chat focus on.

So you'll probably want xlr pre-outputs on the back for studio monitors

But you can get away with rca -> dual mono TS cables. Most studio monitors should support that. So rca should be fine.

I don't do pro audio so idk what else you'd need.

But realistically, what you have now should be fine for the hd600. It's not hard to power and its high impedance means its pretty agnostic to whatever you plug it into.

At the high end, if you get a minidsp studio or a rme adi-2 they're really flexible and will be really useful for doing DSP to make sure you get the most accurate sound for your music.

If you just want something simple, get a fiio k11.

[umbral charm]

Thank you. That clears things up. One thing I’ve noticed is the 600s feel a bit lighter in the low end compared to the 620s, though overall I really like them. Would getting a DAC/amp help give an all round improvement, or would it sound more or less the same? You mentioned the Fiio K11 would that be a good neutral choice, or should I look at something a bit warmer if I want a touch more presence

[lean grove]

The 620s does have more bass and more bass extension than the 600. So you're right in thinking so.

Getting a DAC/amp will not change the bass on the hd600 all that much. So it will be about the same.

And the actual role that a dac/amp will have on the acoustics is not huge enough to fix something like the bass rolloff on the hd600. You need EQ to do that.

The difference an DAC/amp will make is adding a subtle smoothness to the sound. Taking away some of the harshness. So it will probably be an all around very minor improvement.

And as for warmth, neutrality, and presence I'm not the person to ask for that. I pick my audio source equipment based on functionality and convenience. Ideally they shouldn't play much of a role in the tonality of the music.

[umbral charm]

Okay thank you. The bass thing hasn't been a major just something I noticed im comparison. My mixes seem to be fine regardless so it's whatever. Referencing exists anyway. I take it I can just leave a dac/amp for now then

[lean grove]

If you're hitting the volume levels you want, and you don't hear any noise, cracking, popping, or distortion then yeah a fancier DAC and amp would be a luxury.

No need to buy something you don't need if what you have works fine.

[sick terrace]

i'd recommend checking out some audio interfaces rather than headphone amp/dacs. you get 1+ headphone outputs, speaker/monitor outs, and minimum 1 analog input for a microphone, guitar or synth. there's also an added benefit of dedicated drivers that'll increase performance in your daw, assuming you're on windows or linux

i don't have any specific recs. just avoid 3rd gen or older scarletts

[umbral charm]

Thanks guess Ill have to try look into it and see if it's worth grabbing one. I forgot about audio interfaces tbh so depends what it does and if it's worth it.

[maiden badger]

I ordered this today. Is it worth it?

[broken grotto]

do you like how it looks?

[lean grove]

Typically you're supposed to ask that question before you buy it.

But I guess you'll be able to let us know after it arrives!

[broken grotto]

most people ask questions for affirmation of what they just bought

that's why half the time it's pointless telling someone if it's a good buy or not, they've already bought it lol

[lean grove]

I was attempting to give a subtle bit of criticism for exactly this

[broken grotto]

it's just how people are, it'll happen until the end of time lol

[lone flame]

Won't work sadly

[lean grove]

i know. but it's free and worth trying just to see what happens.

something will happen. and it will be a learning experience.

[elder thistle]

but you could've gotten real speakers for this price

did you not have desk space?

[stuck bolt]

Me when people buy shitty speakers

[steel escarp]

@lean grove

[maiden badger]

I have heard good reviews on from people here in the server and someone recommended the brand anyways

lol

How many more times you wanna fail that?

[zenith pawn]

most people's standards of good audio ends at whether it succeeds at creating sound waves

[lean grove]

Then why ask us if it's good

[steel escarp]

I got a Bliss

[lean grove]

Why would you do that

[steel escarp]

For my Susvara

I thought my Violectric V222 was endgame. I was dead wrong

The improvement is very substantial

[lean grove]

Any time you need someone to tell you that you're wrong let me know. I'll do it for a lot less than $3000

[steel escarp]

I hear things on that amp in the far away stage that dont even exist in violectric v222

[steel escarp]

Nah its real

I heard like 5 other 3k+ amps. And thought it didn't matter

But Bliss is different

Sound stage is like 40% wider

[lean grove]

That's impressive considering that the headphones are in the exact same spot on your head

[livid ruin]

yeah headphone mount position and seal make huge difference

[steel escarp]

Its 3W of power vs 10W

[livid ruin]

10w wow

that’s too much even for endgame headphones that require end game amps

[lean grove]

I've got some class D speaker amps I can recommend that are more than 10x that

Imagine a 400% soundstage improvement

You'd be able to hear things from the recording booth next door

[livid ruin]

yeah save for a small house/room dedicated for speakers and you’ll have much better experience

instead of spending so much for amps

plus real estate doesn’t lose money unlike amps, because the cheapo manufacturers are improving by year

[steel escarp]

Class D amps are not good for low level signals

[steel escarp]

I want a personal experience. Not speakers.

I have speakers. I cant play them loud or the neighbors complain.

[livid ruin]

foam exists ^ ^

[lean grove]

10 watts isn't a low level signal.

OH you're talking about the susvara. Doy.
Yeah that thing only needs like a watt

Maybe less depending on what you're listening to

[livid ruin]

pretty sure if you build your own room for audio listening it’ll be the best

[lean grove]

Yeah easily less than a watt

[steel escarp]

Ye I got it for Susvara

[livid ruin]

i guess he’s listening at ~125db

[steel escarp]

Yeah to produce sound it needs that much power.

But not to sound good

It needs more for that

[livid ruin]

i don’t think it’s good for da ears

even if it’s for short time

[steel escarp]

You dont turn it up

You just need an amp that has the headroom

[lean grove]

Oh in that case I have some class D speaker amps with tons of power I can recommend you.

[steel escarp]

Speaker amps are not made for headphones

[lean grove]

You can even hook them up in parallel with the amp you have now so you have an extra 90 watts of headroom

[steel escarp]

The noise floor is completely different

[livid ruin]

are you sure you just don’t have your headphones little off on your head, like for my ears if i angle the headphones forward and pull them forward i hear much more treble for a lot of headphones

[steel escarp]

Also designed to run under a different load

[lean grove]

Imagine how much better they'll sound that way. All this extra power that you're not using just never being used

[steel escarp]

The amp itself is also way better

[lean grove]

Oh well why didnt you say that

[steel escarp]

The engineering that goes into holo products is 10x what the competition does

[lean grove]

I thought you just bought the amp for power reasons.

[steel escarp]

Nah. I'd get a topping then

[lean grove]

If you want I can charge you 10x what Holo costs

[steel escarp]

If you make it 10x better

[lean grove]

$30k for a magni with a fancy chassis I carved from solid silver

With a spoon

[livid ruin]

they don’t add any sound signature to the sound right

[steel escarp]

Bro. Legit only the cost of the parts already adds up a looot

[lean grove]

Nah. Even Holo couldn't do that.

[steel escarp]

Bliss is mostly neutral to my ears

I am not buying it to tune the sound

If you get the chance to listen to a Bliss with Sus, I'd recommend it

Btw the holo dacs are a meme

Even tried the may

[lean grove]

Send them to me I'll give it a listen

[steel escarp]

Oversampling literally beats any nos dac

[green marsh]

Ok buddy

[steel escarp]

I'm not your buddy, mate

[lone flame]

Lmao

[lone flame]

Also.
More Soundstage from a dac doesn't sound real to me

[steel escarp]

the Holo Bliss is an amp

[lone flame]

Does it has stuff to improve that

Like the AMP from Zähl

[steel escarp]

some sort of crossfeed?

[lone flame]

Ye

[steel escarp]

no I don't think so

I was just able to hear very quiet details in the background way clearer on my Violectric amplifier

[lone flame]

Like the Zähl H1 has the best crossfeed I have ever heard

[steel escarp]

I think its not quite called crossfeed but something else that they have

[lone flame]

That's mostly Placebo

[steel escarp]

I did an A/B test

[lone flame]

Blindly etc?

[steel escarp]

I legit was not able to hear a lot of stuff with my violectric, even when knowing its there

[lone flame]

Also most people don't know how to A/B correctly

[steel escarp]

I volume matched

[lone flame]

Measured?

[steel escarp]

no, by ear

[lone flame]

Yeah no that's not really how it works

[steel escarp]

and for funs, I turned the violectric up quite a bit more than the Holo Bliss

[lone flame]

Can be 2+dB of a difference

[steel escarp]

to see if it would sound better

but it didn't

[lone flame]

Like. It's possible that your experience is legit

But I doubt it hardly

[lone flame]

How did you A/Bd them.
Blindly

[steel escarp]

it wasn't a minor difference I was hearing

no not blindly

[lone flame]

So it's invalid

A/B has to be blindly. Because Placebo does ALOT

In your head you expected that it sounds different

[steel escarp]

we are comparing a 1200 euro amp to a 3250 euro one

[lone flame]

And that can cause lots of shit going on in our heads

[steel escarp]

I thought it was gonna sound like ass from what I heard from some reviewers

[lone flame]

Funny that it doesn't mean it's better

Price does not correlate with Sound Quality

[steel escarp]

generally not

[lone flame]

Some 👍

[unique wind]

3250 euro for an amp that's objectively inferior to the topping a70 pro which is only $500usd 🥀

[steel escarp]

bro, an SMSL SU-1 has better SINAD than most dacs

[lone flame]

Tbh the A70 and D70 has lots of stuff happening to measure that good

[steel escarp]

yet it sounds like shit

tho if we talking dacs, Holo audio literally makes the best measuring dacs that I know of

[green marsh]

I'm not your mate, pal

[lone flame]

Cyan 2 treble rolloff

[steel escarp]

I did A/B cyan 2 to Chord Mojo 2

[zenith pawn]

TIL non-oversampling dacs exist in 2025

[steel escarp]

with a switch

[lean grove]

Only if you have inferior digital files.

[lone flame]

A/B has to be blindly otherwise Placebo will get triggered with that switch too 🙂

[lone flame]

Wasn't that Centauri?

With the shit performance to 16 bit 44.1khz files

[green marsh]

no

[steel escarp]

no you can use HQPlayer to upsample

[steel escarp]

they possibly fixed it

[lean grove]

Afaik any NOS DAC will have treble rolloff at 44.1khz

[steel escarp]

they released quite a few firmware updates

[sick terrace]

literally

[zenith pawn]

onboard sucks though i think that one is pretty fair

[lone flame]

Apple dongle

[zenith pawn]

except on apple computers

[steel escarp]

the apple dongle is so ass

[zenith pawn]

yet another apple W

[lean grove]

Only if you have low impedance headphones like some loser.

[zenith pawn]

shut up

[lone flame]

No

[lean grove]

I only use my 2000 ohm Sennheisers

[zenith pawn]

i love my iems that can be driven by a potato batteyr

[steel escarp]

in europe we dont get as much output power on the apple dongle as in the USA

[steel escarp]

tube that mfer up

[lean grove]

Sounds like someone's shouting at you at all times but I can confidently listen without worrying about output impedance

[lean grove]

I did but the tube amp had the worst ground noise leakage I've ever heard

[steel escarp]

lmao

[lean grove]

Made it past my fancy measure baiter DAC

[lone flame]

Rip

[steel escarp]

average lowfi tube amp

[lean grove]

Yeah xduoo ta-26s

[zenith pawn]

pretty soon we're gonna see people using tube amps in 2025

:supershocked:

[steel escarp]

people are tho? they are great with dynamic headphones

[lone flame]

I liked some tube amps

Especially on my HD800s

[zenith pawn]

that was the joke i was making

[lean grove]

I should get a ta-66. And a jds labs synapse

[lone flame]

Imo depends. Some tubes sound ass on some headphones

[steel escarp]

sennheisers are good with tubes

[unique wind]

nah not necessarily

[steel escarp]

so is dt880

[unique wind]

motherboard analog audio outs are the lowest latency solution for a pc

[lone flame]

KSC75 with tubes

[unique wind]

a lot of new boards are coming with usb based audio tho so we're latency cucked regardless

[zenith pawn]

simply run everything at faster clocks

[lone flame]

Ok guys back to work bye

[steel escarp]

just get a NOS, no decoding, just sound go brrrr

crazy low latency

[zenith pawn]

why can i not run my usb at 10 GHz

[unique wind]

ive been exclusively using apple dong for a long time now

for my iems

[lean grove]

https://www.lineartubeaudio.com/products/z10e-electrostatic-headphone-amp-integrated-amp

This is the only tube amp I need

[steel escarp]

just get a Feliks Echo 2

[green marsh]

latency cucks out

[lean grove]

Sugma

LTA better

Can power speakers

[zenith pawn]

latency cucks when the 480 hz monitor has 2ms of latency

[steel escarp]

if only Cayin tube amps were not so noisy

[green marsh]

latency cucks when the human reaction time is 200 ms

[zenith pawn]

TWO HUNDRED CHERRY MX SWITCHES

[unique wind]

some nerds are going back to using pcie sound cards in their main rigs in order to get lower latency

[steel escarp]

man

[unique wind]

it's pretty but it's a scam

[zenith pawn]

pcie latency is too high

we should have it connected directly to the cpu lanes

[steel escarp]

in what sense? the component cost alone already justifies the price

you are getting your money's worth

[unique wind]

how so?

it doesnt do anything for you that a cheap dacamp couldnt

[steel escarp]

yes

drive Susvara

[unique wind]

it's pretty yes but it's essentially overpaying for art

[steel escarp]

really well

[green marsh]

i dont think a cheap dac can do r2r nos

[steel escarp]

this is not a dac

[steel escarp]

fiio k11 r2r

tho it measures like shit

not even an SNR of 80

[unique wind]

XDDDDDDDD

[lean grove]

It's not a scam any more than a Ferrari is a scam. Or a rtx 5090. Or any luxury item.

[green marsh]

ferrari kinda is a scam 😔

[unique wind]

those have appreciable objective performance characteristics

overpaying for a dac or amp is not comparable

[green marsh]

cope

[warm scarab]

i think ferrari is kind of a scam yeah

[green marsh]

latencucks need not apply

[unique wind]

frick ferrari fr

i wouldnt buy one even if i had infinite money

[sick terrace]

with dacs and amps you're paying more for measurements you will never be able to perceive outside of a graph

[warm scarab]

if you had infinite money that would make money worthless so you would have zero money

[green marsh]

gota buy 50 other ferraris before you can buy the one youwant

[warm scarab]

this guy is stupid or something i swear

[steel escarp]

no you can get good measuring stuff for relatively cheap

[green marsh]

and you cant do anything cool with it either

[steel escarp]

its not what you are paying for

[warm scarab]

the things id do for an r34

[lean grove]

Why are you buying a Ferrari when a Ford fiesta can get you to work just as quickly?

Or why are you paying $4000 for a 5090 when you game at 1080p?

There is a real difference but any moron can come in and look at it and deem it unnecessarily expensive.

[steel escarp]

only if you spend it all

[warm scarab]

only if you hide the fact you have infinite money

[unique wind]

i will pray that you guys escape the hamster wheel eventually

[lean grove]

If you could go to Ali express and buy the exact same Holo bliss for 1/10th the cost in some no-name Chinese PCB, yeah it's a scam.

[zenith pawn]

i would as long as they didn't blacklist me

[lean grove]

It's literally just gray

[warm scarab]

i dont think holo stuff is a scam because of what you pay for the shit inside

[lean grove]

I'm almost retired from hifi

[green marsh]

hell nah

buy a mclaren

[unique wind]

[green marsh]

or porsche

[warm scarab]

you couldnt get what holo may has in terms of hardware from anywhere else

[broken grotto]

why is everything about "good measurements + cheap?"

[zenith pawn]

i would buy a porsche

mclaren is dumb

[broken grotto]

why not pay more for something just because it looks nice, or you like it, or something along those lines?

[zenith pawn]

porsches are actually made well

[unique wind]

yeah you can overpay for art if you want to, i bought a premium 5090 model because it's pretty

[green marsh]

pushback against the boomer mentality or something

[warm scarab]

porsche makes actually driveable sports cars, mclarens lose their worth while being on the road

[lean grove]

I got a new hobby to drop money on.

[zenith pawn]

fr

[warm scarab]

reddit

you went from reddit to reddit

wow good on you

[broken grotto]

buy garbage quality because "paying more for something," is bad or smth

[unique wind]

klaus is jealous because he's too obese for that bicycle

[warm scarab]

i cant believe we lost a man to fucking cycling of all things

[steel escarp]

I didn't buy it for measurements or looks. but for the sound.

[unique wind]

this is why i say you get scammed

there is no appreciable sonic difference

[steel escarp]

but measurements cant quantify how it sounds

[unique wind]

if you wanna buy it because it's pretty then go for it

[steel escarp]

that is where you are wrong

[unique wind]

but youre deluding yourself and wasting money like an addict

[green marsh]

could be worse

[warm scarab]

its got like 50 pcm filters stacked man

[zenith pawn]

then invent a new measurement to quantify how it sounds :trol:

[warm scarab]

you'd have to be deaf to not hear a difference

[green marsh]

have you tried it

[unique wind]

i dont even know what device in particular he's talking about right now

what is it called and has someone measured it?

[lean grove]

Not even cycling. Bikepacking.

[steel escarp]

the Holo Bliss https://goldensound.audio/2023/02/06/holo-bliss-kte-measurements/

[warm scarab]

tf does that mean

stop being smart idiot

[lean grove]

I'm gonna slowly go long distances with a bunch of shit on my bike.

I've got a 100 mile trip planned in a couple weeks.

[warm scarab]

at least you are in a place where you dont bother others while on your bike

[green marsh]

why does it need to be measured just use your ears

[steel escarp]

measurements only indicate something isn't absolute dogshit

[warm scarab]

i hope nils gets stung by some weird mosqiuto

[zenith pawn]

measure your ears

??

[steel escarp]

they literally dont tell you anything about how good it sounds

[unique wind]

you are not supposed to hear your dac or amp grey that's the entire point

[warm scarab]

@steel escarp i have a joke

[unique wind]

all it's supposed to do is convert from digital to analog and then amplify without altering the sound in any perceptible way

[broken grotto]

i honestly can't tell if klaus and nils just hate each other or if it's mutual insults

[warm scarab]

what do you call a cyclist that goes over long distances slowly with a lot of stuff on his bike @steel escarp

[green marsh]

🤡

[lean grove]

Been the most fun I've had in years. Genuine fun. Climbing up a thin mountain trail with nobody for miles.

[zenith pawn]

nothing ever happens and nobody hates anyone

[steel escarp]

a delivery guy, idk

[warm scarab]

bikepacking 😂

[green marsh]

i hate everyone

[zenith pawn]

my gif picker is a mess

[steel escarp]

keep yourself safe

[zenith pawn]

i couldnt find it in time

[warm scarab]

https://cdn.discordapp.com/attachments/756995857777164360/1352304985765052508/Luna_a.gif

[steel escarp]

disagree

[lean grove]

https://tenor.com/view/cat-cake-cutting-cutting-kitty-gif-11810334948872627297

[unique wind]

nils the gigachad explorer

[warm scarab]

it alters sound though

[broken grotto]

there are actually cases where you want it to

[warm scarab]

lmao

[zenith pawn]

it's a luxury hobby

it's all expensive toys

[warm scarab]

first and foremost assuming we have perfect D/A isnt true

[broken grotto]

and also, just electrically it will have characteristics that modify the output.

[warm scarab]

second of all you dont need perfectly converted D/A to enjoy it

[green marsh]

regardless gray is talking about an amp rn

not a dac so 🤷‍♂️

[warm scarab]

we should hunt down gray

[green marsh]

troo

[warm scarab]

he has gone unpunished for too long

[steel escarp]

converting digital to analogue perfectly is impossible. you require an infinite amount of sinc functions to decode the signal

[broken grotto]

also this thing is expensive, not just literally but they are more than happy to call out some components used

[warm scarab]

grey said the same thing as me

so ill change my opinion instead

we have perfect A/D

[broken grotto]

neat

looks nice

[zenith pawn]

but the question is whether the missing data is physically perceivable by human hardware

[broken grotto]

for reference, the caps they say they're using, Mundorf, i think are nearly $100 a cap.

[steel escarp]

the closest you can get to converting digital to analogue perfectly is by using extremely high taps filters. via HQPlayer. and for using that you'd ideally feed it into an R2R dac that supports NOS. but most NOS dacs measure really badly due to resistor tolerances. So if you also want clean. you'd have to go for a Holo R2R dac

[sick terrace]

yes but there's an appeal to chasing coloration through amplifiers. adding harmonic density to sound is virtually always gonna sound better to our ears, even though it's technically inaccurate. music is recorded, mixed and mastered with coloration/distortion in mind the entire time to produce a more pleasant result and amps are just an extension of that from the listener.

regardless, i think there's better means of reaching that coloration than relying on overpriced amps

[steel escarp]

There have been several studies on this. Current conclusion is that its still unclear what exactly we can percieve

[zenith pawn]

that's why it's a question

we all have our bets placed on what the answer is

[broken grotto]

44.1kHz contains everything you need, under the Nyquist Sampling Theorem. It's more the implementation that matters.

[steel escarp]

yes, but there is no perfect implementation

you can never get the 100% correct transients back

[zenith pawn]

there's no such thing as a true transient, is the point

and afaik it's not believed to be perceptible

[steel escarp]

only if you use an infinite amount of sinc functions

but that would be impossible

but literally no cheap dac uses even close to a good amount of taps

[zenith pawn]

idk the exact specifics or if i can explain it right
everything between your audio file and your ear hairs is supposed to be a linear system, so it's a sum of the individual frequencies that create the sound wave
we only have ear hairs that can perceive up to a certain frequency
so replicating anything above those frequencies shouldn't be possible to perceive

so it doesn't matter if we can't recreate a perfect transient because we just have to recreate the components of a transient below the maximum frequency we can hear

[broken grotto]

a few hundred taps would give you about 100dB of stopband attentuation, which is basically where it stops mattering

even in Radar we never gave a hoot past a few hundred taps.

[lean grove]

https://tenor.com/view/no-one's-getting-spanked-karen-tarleton-i-think-you-should-leave-with-tim-robinson-no-one's-getting-hit-no-one-will-be-slapped-gif-8178224902305421035

[zenith pawn]

someone let me know if this is accurate because i would like to reuse this explanation

[steel escarp]

I am not talking about accurately reproducing information above the nyquist frequency

[lean grove]

Within the relevant bandwidth you can.

[zenith pawn]

transients aren't some special magic signal, we can already recreate the parts of a transient within the 0-20 khz domain

[steel escarp]

yes, but those transients are incorrect

[zenith pawn]

for it to be incorrect that would mean you have to be talking about information above the nyquist frequency

[steel escarp]

you can't reconstruct signal below the nyquist frequency properly without an infinite number of taps

[broken grotto]

well, real quick because I have a feeling about something rn: describe to me what the "impulse," you are talking about is in the time domain.

[steel escarp]

an impulse is a type of transient

[broken grotto]

wow he had that on standby

[steel escarp]

I am talking about an issue in the time domain, yes

not the frequency domain

[zenith pawn]

we start with the assumption that you can't hear anything above 20 khz or so
so we can just remove all the information above 20 khz
you can perfectly recreate that signal, that has the supersonic information removed, by sampling it at the nyquist frequency

[broken grotto]

I wanted to verify that lol

[zenith pawn]

and that's proven

that's not just something we assume is true

[broken grotto]

well

let me summarize what I think he's arguing

[warm scarab]

im unsure what was said but take the opposite of what grey said and youll probably be correct

[steel escarp]

you need an infinite amount of sinc functions to reconstruct it 100% accurately. otherwise the output is taken at the sampled points, and then the stair stepped output is smoothed. but the waveform is malformed then

[lean grove]

Again. Grey. This isn't the case for a bandwidth limited signal

[steel escarp]

it is

[lean grove]

It's literally not.

[broken grotto]

....

[broken grotto]

bro

[warm scarab]

mate

[warm scarab]

.

chat should try mirroring 1% of my intellect

[broken grotto]

that figure shows the concept of sampling rate. that plot isn't even about a "number of taps," that's sample spacing.

[steel escarp]

taps are used to interpolate the signal

during reconstruction

[lean grove]

Did anyone notice how it's the exact same sine wave despite one having more samples than the other

Crazy

[broken grotto]

ehhhhhhhhh. once you satisfy the Nyquist condition, the difference comes down to how accurately the filter is (with finite taps) reconstructing the bandlimited signal.

[zenith pawn]

Lmao

[steel escarp]

I am not even talking about the passband ripple. I am saying you get phase distortion.

[zenith pawn]

whenwe know the signal is band limited, we don't need any of the information between those samples at the Nyquist frequency because it can just be regenerated

[broken grotto]

keep shifting the argument bro

[steel escarp]

bro, the actual line is not what would be outputted. its how its sampled at recording in that image. the first line would be distorted

[steel escarp]

its still about how you cannot get an 100% accurate transient

due to the phase distortion and ripple

[broken grotto]

so you are now arguing that an insufficient amount of taps cause phase distortion?

[steel escarp]

which would only completely disappear when using infinite sinc functions

[steel escarp]

I was from the start

[green marsh]

distortion is a product of the mind

wake up and smell the roses

[lean grove]

No. It wouldn't be. There's a single perfect solution for the line that travels through all those dots as long as you're below the Nyquist frequency.

Which it obviously is.

[steel escarp]

your hardware has to generate that line based off the samples it has

[green marsh]

considering nyquist states you can perfectly recreate a signal if your sample rate is 2x i dont think your arguement stands

dacs aint shit

apple dongle is acoustically transparent

[steel escarp]

this assumes ideal conditions: the signal must be bandlimited, and the sampling must be uniform and precise with infinite precision

perfect reconstruction is limited by factors like noise, finite precision, or non-ideal filters

[broken grotto]

I just want to point this out, https://en.wikipedia.org/wiki/Finite_impulse_response

[steel escarp]

mathematically its possibly, practically its not

[broken grotto]

oh i see golden

[white gate]

is isn't the whole criteria. It says you CAN perfectly reconstruct.....IF you PERFECTLY and infinitely attenuate everything after the nyquist frequency with an infinite-coefficient filter as @steel escarp was describing.

In practice we don't have infinite compute power so this isn't possible. Practically all DAC filters will either roll off early, attenuating treble, or will fail to attenuate fully by the nyquist frequency at all, often both.

[broken grotto]

this script was modified from an age old radar one

[white gate]

Most DACs use 128-1024 tap (coefficient) filters. Which are "fine", but don't technically adhere to nyquist theorem.
Better DACs like Chord, Ferrum etc will have filters that do both leave <20khz content untouched and fully attenuate by 22.05khz by using higher coefficient count filters which requires much more compute power. (Hence the more powerful FPGAs in Chord DACs and the powerful SERCE module in the Ferrum WANDLA).

Then there's objectively much more accurate options like the MScaler with 1 million taps. Or HQPlayer which has filters up to a few million taps as well as quite advanced noise shapers which will also then give you higher dynamic range in the audible band.

Right now objectively the most accurate tool is PGGB, which uses the maximum number of coefficients mathematically possible with the number of samples in the file, AND an absurd noise shaper that has well over 600dB of dynamic range in the audible band even with 24 bit output. But it takes several minutes to process one file

[lean grove]

Grey isn't saying that though. He's saying you need an infinite amount of sine waves to recreate the signal.

[steel escarp]

phase distortion still affects transient accuracy, even for passband frequencies not near the cutoff

[steel escarp]

no infinite amount of sinc functions

mathematically

[white gate]

Technically this is true. An actual 'perfect' reconstruction WOULD require infinite coefficients.
However in practice, infinite compute power isn't a thing, and there is also an additional limit in the number of samples we have available in the source audio itself.

So in practice, we can get as close as possible to a perfect reconstruction using something like PGGB. But then that can't be run realtime.

So instead you can use something like HQPlayer, but that still uses a lot of compute power, and more than could be put into a DAC itself in most cases.

[green marsh]

pc sized dac when 😈

[steel escarp]

maybe one day. someone with enough brain power will invent a new reconstruction method with 100% time domain accuracy

[green marsh]

phasure was on to something

[white gate]

lol no they weren't

phasure is crazy weird

[lean grove]

We're talking about bandwidth limited signals.

[steel escarp]

he knows

[white gate]

yes

[lean grove]

Infinite bandwidth isn't in the question

[broken grotto]

Lemme phrase this from my perspective of DSP doing Radar. In radar, we care about the stopband attentuation and phase linearity. 100dB was "good enough," because it cuts out thermal noise floors and anything further gave no practical gain, it's buried in system noise at that point. In practice, when we hit 100dB the filter was "good enough," and attention shifted to hardware limitations.

FIR Filters can be designed to be Exactly linear phase, which causes no distortion of phase pulses, only a uniform time shift. The tap count determines the transition width and stopband depth, not phase accuracy.

IIR filters achieve a sharp roll-off with fewer coefficients, but phase response is inherently non-linear, which does distort our radar pulses and makes matched filtering harder.

That's why we used FIR filters when phase was a big deal.

[white gate]

infinite bandwidth and infinite coefficients aren't the same thing

[lean grove]

Why would you need infinite coefficients for a limited bandwidth.

If the Nyquist theorem can recreate the signal perfectly

[steel escarp]

to get perfect time domain accuracy in the phase

otherwise you get phase distortion and ripple

[white gate]

The thing is there's more to it than stopband attenuation, particularly in audio where we have such a narrow band between the audible band itself and the nyquist frequency.

You can design filters with shit tons of stopband attenuation but incredibly poor in-band response or attenuation by the nyquist frequency itself. And you can also design filters with essentially perfect attenuation at the nyquist frequency but with terrible stopband attenuation

Stopband attenuation itself is very rarely the issue. Most even quite basic filters will have >100dB of stopband attenuation, the issue in audio is almost always when, and how steep the attenuation is

[lean grove]

Are you talking about ringing artefacts

Or is this something else

[steel escarp]

something else

ringing isn't really that much of an issue if you are not near the cutoff

[broken grotto]

yep, im not gonna disagree, audio is much lower frequency than radar (radar is 9+GHz, enough to give you cataracts if you're right next to the transmitter : D )

hehehehe one transmitter we had was 5kW.

anyway.

[steel escarp]

@lean grove best image I could find

only look at the graphs

[white gate]

The issue would be much less a concern if we used higher sample rates. Even just 48khz.

44.1khz leaves us with only 2.05khz to attenuate by ideally at LEAST 100dB.
48khz would double that to 4khz

88.2khz gives us over 20khz of space between the audible band and nyquist frequency and so even really basic filters would be totally fine

[steel escarp]

okay actually this image is bad

[lean grove]

It looks like regular distortion but the distortion harmonics are offset by a bit.

[steel escarp]

surely there has to be an image that shows the actual analogue signal (within the audible band) that is produced from lower sample rate vs interpolated during reconstruction

[broken grotto]

however, @steel escarp is talking about phase

[white gate]

This is a typical filter response (the fast linear filter in an ESS DAC). It has plenty of stopband attenuation (also keep in mind here that 0dB on the Y axis is normalised to the level of the white noise on the FFT which is about -60dB. So technically this is actually more than 150dB of stopband attenuation.

BUT....it doesn't fully attenuate by 22.05khz.

[broken grotto]

woah

[white gate]

This on the other hand is the Chord DAVE with MScaler

[broken grotto]

my matlab script nailed that lmao

[white gate]

Instant attenuation at 22.05khz

[broken grotto]

yeah, which more comes down to implementation. that's a high quality DAC filter right there.

my concern is the argument about phase

[white gate]

DAC filters are typically either linear or minimum phase, which means that any behaviour that isn't perfectly minimum or linear phase will be as a result of quantization noise. But that's where stuff like Chord's modulator, or PGGB/HQPlayer's noise shapers come in.

[steel escarp]

I give up. I cannot find a picture that shows off the phase distortion I was talking about

[white gate]

The way I check for time domain accuracy is by generating a 1fS signal which is a sum of multiple sinusoids (different frequencies and phase), then running it through PGGB to upsample to say 16fS, comparing the results directly with the same signal generated at 16fS and finding the maximum absolute error. This is all before noise shaping. The current PGGB at 64 bit precision is accurate to the 10th decimal digit. I.e., the maximum absolute difference between the upsampled signal and the signal generated at 16fS rate is 1e-10 or 33 bit accurate. At 128 bit precision the accuracy is within 28 decimal digits or 93 bits and at 256 bit precision it is 55 digit or 193bit accurate.

Which is nuts

[broken grotto]

pretty sharp way to test that. But, what about consumer dacs?

the more "normie," ones.

[white gate]

Consumer DACs it's not possible to test the modulators themselves since you're bottlenecked by analog domain noise.

Even if you have a beyond-physics perfect modulator, no getting around johnson noise.

You could if you want technically run a days-long FFT but to get to proper low levels it's going to literally be days, and you'd need to set something up to even run a billions/trillions of points FFT as well so that's not something I've done 😅

Some manufacturers like Chord are pretty proud of their modulators so they've released some stuff. This is the Chord DAVE 1-bit modulator resolving a -300dB signal for instance

[broken grotto]

sure. im framing everything from radar DSP which i know is a different beast, even though the same theorem apply. Most of the radar DACs i made were aimed for 10-14ENOB.

[white gate]

But the filters themselves are of course pretty trivial, so I measure/test those whenever I measure a DAC

[broken grotto]

I just have a very hard time believing in the "phase distortion," being as audible as he believes, is all.

[steel escarp]

NOS dacs sound softer for a reason. besides the treble rollofff

[green marsh]

many things are inaudible

[white gate]

That's something which is a bit tricky. Personally, I'd argue it IS audible, even though logically it's a bit hard to explain why when you're looking at such incredibly tiny differences. But I have successfully ABX'ed between different modulators where the differences are all below -100dB, and plenty of others have shared their experiences with the different modulator options on dCS, L&P, Sony DACs, or in tools like HQP, questionable as sighted impressions are ofc

[elder thistle]

why do we need a filter at all? won't our ears create the sharp lowpass necessary?

[white gate]

no

Without a reconstruction filter you get -3dB treble rolloff by 20khz

[elder thistle]

howdoes that happen

[broken grotto]

gosh i love signals, systems, and DSP.

[white gate]

Because there aren't enough samples at those high frequencies.
Again, the big thing people forget is that nyquist theorem is based around using a reconstruction filter. Without it, you aren't just 'not removing the aliasing', you're also not adding the extra bits at high freqs that SHOULD be there

[steel escarp]

you get sinc droop

[broken grotto]

put that in a textbook

[steel escarp]

a good reconstruction filter accounts for it

[elder thistle]

nyquist isnt a real frequency though

[green marsh]

ok?

[lean grove]

If it were important you'd think there'd be more graphs of it.

[elder thistle]

there are just mirrors of audible frequencies and inaudible ones around it

[green marsh]

huh

[broken grotto]

that sinc droop is quite a bit bigger over hundreds of MHz in Radars, but our pre-comp in DSP fixes that.

[steel escarp]

the term nyquist frequency means the frequency of the sample rate / 2

its a real frequency

[elder thistle]

if you're to lowpass a simulation of this in music software, you can lowpass below nyquist and preserve the lower frequency

[steel escarp]

I wasn't wrong tho

rare W

[broken grotto]

lmao

it was inconclusive.

[green marsh]

but were you right

[elder thistle]

why would our hearing be different from this lowpass? and why would putting an electronic lowpass slightly above 20k cause any intervention

[broken grotto]

Sampling theorem.

[steel escarp]

sleepy time

https://cdn.discordapp.com/attachments/403909858287550467/1410749910722019428/bocchi-bocchitherock.gif

[elder thistle]

whats that mean

or like what point within it

[broken grotto]

If a signal is bandlimited to a maximum frequency f_max, then it can be perfectly reconstructed from it's samples if and only if the sampling rate, f_s is greater than twice the maximum frequency.

[elder thistle]

this is already the case

why would that play any role between ear lowpass or electronic lowpass

[broken grotto]

the theorem requires a brick-wall filter.

which your ears are not.

[green marsh]

my ears are bricks

[broken grotto]

to prevent aliasing, you use the 20-22kHz region as the transition state for the brick wall, thus preserving the nyquist theorem and preventing aliasing.

[warm scarab]

i rub rocks together for music

[broken grotto]

placing the transition at a lower frequency just steeply attentuates those above it.

[steel escarp]

I didnt know u made rock music

[elder thistle]

no. the theorem only actually requires total filtering to the opposite extent that frequencies exist in the audible band

[green marsh]

?

nyquist aint just for music

[elder thistle]

eg if you can hear to 16khz and nyquist is 24, you must completely filter things out by 32khz

[green marsh]

nyquist is for any signal

[elder thistle]

the ear accomplishes this

nobody hears to 24.1k when nyquist is 22.05

so what makes it not shooting ourselves in the foot trying to electronically lowpass?

[broken grotto]

https://tenor.com/view/huh-gif-1807502725802114204

[elder thistle]

I have a demonstration of this on my pc somewhere

[green marsh]

bro clearly does not know what hes talking about

[elder thistle]

I am well-versed on pcm audio

[warm scarab]

i have a pc somewhere

[elder thistle]

it quotes this because an ideal lowpass is a guaranteed way to give a perfect reconstruction

if you consider brickwall filters per analog electronics perfect enough, then ideal lowpass is not the only way

[broken grotto]

For example, let's assume we're digitizing a radar pulse with a maximum bandwidth of 40MHz.

• By nyquist, we need a sampling rate of 80MS/s. This means our sampling frequency is 40 MHz. (it's real.)
• We assume the radar pulse spectrum is strictly 0-40MHz and nothing above that.
  In reality, real radar pulses aren't strictly band-limited. There will be leakage above 40MHz. If we sample at 80MS/s without filtering, those higher components alias back into the baseband and corrupt the signal.

[elder thistle]

this is why audio interfaces are good about lowpassing their outputs. they'll be sent back to the inputs/adc's

it is not the same as a filtered adc or in this case our ears

[broken grotto]

what we do (and what I did,) was insert an anti-alias low-pass filter before the ADC.

• Passband flat above 38MHz
• transition 38-40MHz
• Stopband > 100 dB attentuation above 40MHz. This is the guard band and filter steepness requirement that turns the theorem into a working system.

[broken grotto]

ideal lowpass doesn't exist. The only mathematical way is a brick-wall low pass. everything else we build is just an approximation.

[elder thistle]

so why would we need an electronic lowpass if our ears also lowpass?

I think I may have a misconception about the nature of the ear's lowpass, or if humans have the audible band affected by ultrasonics

that's what I'm trying to learn

[broken grotto]

there's no aliasing. you just don't hear it. anything above Nyquist aliases back into the baseband.