I once owned a Gato 250s. Fabulous. I have since upgraded to the 400s. With no regrets and "a big grin to pass the time of day with"

Paul, once again you've hit it out of the park... great job Sir.

Paul, you've created a new conundrum. Now you need a video explaining why customized FPGA Dacs are superior to off the shelf implementations. Not to say that there's no talking about R2R ladder Dacs in the conversation in comparison to sigma delta dacs. LOL!

@russellzauner

18 күн бұрын

The clock speed/sample rates of FPGA got fast enough and devices quiet enough - plus the toolsets are massive for designing not only circuits and devices but running emulation and test cases. The bonus of FPGA is you can easily build instrumentation into the device that runs along with but outside your actual design to give you real time debug output so instead of taking massive amounts of build and prototyping time to dial in your amp you can rapidly cycle through different aspects on the actual hardware as well as run simulations Another advantage is you can map out your clock jitter, latencies, etc, simpler than in an ASIC/DSP, where you can get most of that information directly from the device or with simple implementations - it's easy to add something like ring generators to send clock/timing correction if you can simply reflash your PROM instead of having to include instrumentation externally or adding said dual ring generators to your ASIC to cough up a differential that equals your standard clock jitter in that process technology as a value that can be used to do same (but if it's broken, surgery on an ASIC is not even the same thing as on an FPGA). The DAC is just the engine for the output signal - there is a lot of signal conditioning that goes on after before you get to the final boss of the signal chain, the speaker (I forget how many bits one typically loses with dithering/fidelity removal)...but there are capacitors on all the cells to keep them stable, so your DACs that straight up use divider/capacitor stages would appear more similar in behavior if you were checking for any sort of repetitive and or periodic noise/glitches. A sigma delta can offer a smoother output since it's already got an amplifier stage and it's regenerating/shaping signal, but now you're regenerating/conditioning signal within your converter instead of just taking the samples, getting that digital value, and passing that on to the amplification/conditioning stages; I think from a rapid development viewpoint the design becomes easier to work with using FPGA and that's been my experience on many projects where I have options in prototyping to use FPGA or build discrete (and some engineers want both, because you can email a customer/partner/peer a simple textfile, they can change the extension, and load it up on their own FPGA board rather than send them a new board each time a change is made).

@lasskinn474

18 күн бұрын

everyone should just go back to covox speech thing.. you think there would be market for 'high end' resistor sets? maybe a kickstarter. Paul really goes off track here with talking just techno mumbo jumbo about heat and processing power, when neither of those are a problem on the chips. it's not like a dac chip runs out of processing power or uses tens of watts.... sheesh

@tristanjones7735

18 күн бұрын

​@@lasskinn474 No, paul is 100% correct. You are right in what you said, but you are wrong in how you are thinking. For the most part there are 4 main ways to make a dac. R2R, current Segment, single bit delta sigma, and multibit delta sigma. R2R is a series of resistors on a chip, and current segment dacs are kinda rare, so we won't bother talking about them. R2R dacs have a problem. Despite what an R2R chip may claim to do, there are no real r2r dacs with a real world resolution above 16 bits. In fact, most dacs only have a real world resolution of 16 bits or less. The only way you can have resolution higher than 20 bits in the real world is to use delta sigma. The problem is that delta sigma is an approximation of the real audio source. Some level of signal processing needs to be done in order to use delta sigma. The more processing you can perform, the more accurate the signal will be. You are correct that small DA chips do not run hot, and don't suck down tens of watts. That's because they can't; not because they otherwise wouldn't. In fact if you could throw several GPUs at the delta sigma process, it would be much more resolving than an fpga.

@thinkIndependent2024

17 күн бұрын

​@@tristanjones7735 I Agree my favorite setup is Upsample filter through DAC#1 Coax into DAC#2 Ultra Clean Power can produce true Analogue sound with this setup

@ThinkingBetter

17 күн бұрын

@@russellzauner FPGAs can be used in the development of a new DSP chip, but as DAC chip supplier, obviously there is a huge benefit in producing actual dedicated chips. What you lose is the FPGA upgradeability of it, which is nice but in itself doesn't have anything to do with audio fidelity.

Always been impressed with Analog devices. But your view and explanation grow on me - i always find it trust your answers, and agree with the once i know anything about. Keep the videos going out. Excellent thank you.

Only Burr Brown !

i love old burr brown chips. very close to r2r implementation.

@edmaster3147

15 күн бұрын

1702 implemented with DC coupled output stage and silver mica filters as specified by BB, I/V bandwith at video level. Makes FPGA sound dull lol

AWESOME ANS WISE WORDS!! Thanks Paul!! Spot on sir. 😊

Fantastic question and very informative answer. I'd love to hear more about DAC output stages and power supplies

@adrianadrianp5305

18 күн бұрын

I admit to experimenting with adding LPS solutions with different choices of capacitors on numerous Blu-ray players and did indeed find the analogue output stage was the most influential

@housepianist

18 күн бұрын

I think that would be ideal because I don’t think - based on my experience - that these different DACs sound all that much different from each other. When you look at the various circuitry implemented with these DACs - voltage applications, output stage, amp classifications, etc. - you definitely have to consider those factors in its overall sound and not just the DAC itself. The only way one can truly hear any differences is having different DACs placed within a set circuit and compare them. Otherwise, there are simply way too many variables to take into consideration to what one may be actually hearing.

This is slightly off-topic. I'm a British 🇬🇧 Hi-Fi enthusiast of 45 years experience. There is a LOT of snobbery and snake oil in Hi-Fi. I have a Hisense 4K TV. I've used my 45 years of knowledge in video and audio to adjust the settings. Picture and sound (while obviously NOT Hi-Fi) are excellent by TV standards. This is where it gets interesting. The other day, I played some FLAC files from my Minisforum GK41 Mini PC through the TV (which serves as a 4K 50" monitor). The vocals on Buddy Holly's True Love Ways, were stunning (I've never heard them sound bad) through the lowly TV speakers! I've never heard such good sound THROUGH A TV! Cue the Hi-Fi snobs telling me I need my ears syringing? ... Err, I don't! Good sound, is good sound, no matter WHAT it emanates from! (Remember, I'm NOT saying it's Hi-Fi!)

I don't think DACs matter, but implementation matters. Perhaps tuning is a better word? I have some devices with ESS that sound "musical" and some that sound bright and shouty in comparison.

Hi, thanks for that

Many moons ago, I had an Alpine ERA-G320 eq and DSP. It had a BBE circuit. It was adjustable. It could be contoured to increase effect at high or low frequencies. Utterly impressive. That EQ created a soundstage in my car that rivaled many home systems. Miss that thing!!

Would love to hear your opinions on ladder dacs and their supposed warm sound

Yes, let's hear more about encoding the fpga.

Everything matters, and so do the chips. I had three ESS and two AKM DACs in price ranges from 900€ to 2900€ and I can say with confidence that they have a different sound signature, which always shines through. Whether the overall sound is full, fluid, well-balanced, low distortion, what have you, depends a lot on the circuitry around the chips.

Personally I think that would be a great series of videos, that is going through the signal / processing chain to explain what is happening and myth busting Digital signal input Re clocking / jitter removal Filtering / conversion Pre amp One short video for each discrete stage.

I’m an AKM fan. ✅

@PanAmStyle

18 күн бұрын

I have a Teac 301 DAC that uses AKM chips. In product cycle terms it’s “ancient” but it does sound very good. It isn’t as resolving as other DACs I’ve heard or even had, but I suspect that’s because the chips are multiple generations away from the current AKM offerings.

@IgorVicS

18 күн бұрын

​@@PanAmStyleMaybe BurrBrown PCM1795?

@PanAmStyle

17 күн бұрын

@@IgorVicS I no longer really track chip models. I’m leaning toward the Audio-GD R7-HE Mk3, which uses an FPGA plus R2R/ladder.

Makes sense

What do you think about the D&M which ended their SACD/CD mechanisms? will Luxman offer their LxDTM-i?

Hi Paul , great interesting video, another chip set is Texas which rotel now installing in for eg rotel A11 mk2 amp instead burr brown , it’s very interesting how you do yours for your equipment. Regards mark

@CaptainCrunch823

5 күн бұрын

If by “Texas” you mean Texas Instruments, they purchased Burr Brown and continue to manufacture the same chips. (Still commonly referred to Burr Brown)

Hi again on rotel websites the mk2 a11 bow uses Texas Instruments 32 to 384kbits dac instead of the burr brown in the a11 tribune, I hope that’s more explains the point I was making regards mark

So... Chip choice is the opening moves of a chess game and the analog section is the rest? (I don't play chess).

As always a good overview by Paul. If we delve a little deeper, I'd argue that everything matters and that the chip is the least impactful. They are all good. But that's like saying all red wine from Italy is good. Some are better and we all have our preferences. I like the nice old TDA1541a chip. DAC sound is primarily driven by 4 thins. 1. Power supplies and grounding. Every supply matters. Particularly the digital ones. Separate supplies for every consumer with tons of F and low noise turns digital to analogue. 2. Output stage including I/V. 3. Extremely low jitter I2S inputs. 4. The actual DAC. While they are all great, when 1 through 3 are perfected, the DAC engine really sets the sound. ESS is detailed but can lean to hard. My old TDA1541a is obsolete in terms of specs, but to some the sound cannot be beat.

@mikeg2491

18 күн бұрын

I have multiple TDA1541 DACs, some modern implementations like Abbas, MDHT Orchid and Border Patrol as well as some period era pieces from the likes of California Audio Labs and McIntosh. And yet being the same chip they all sound different but still have the soul of a smooth harmonic sound without digital glare.

For me TDA1541 chip.

@joejoejoejoejoejoe4391

18 күн бұрын

Is that NOS ?

@karlisdizbite6432

18 күн бұрын

@@joejoejoejoejoejoe4391 Nakamichi CDP-2

@necrodh

18 күн бұрын

You know!

I recently picked up a dvd audio/sacd player from 2005 that uses burr brown 1738 dacs. Compared to the ess dac in my cambridge audio integrated amp I bought in 2020, its not a contest, the burr brown is simply nicer to listen to

Has the quality of Burr-Brown chips changed since Ti bought them out? Better Worse?

Paul is very correct.

FPGAs are actually quite noisy. That's because they have unnecessary circuitry that isn't needed in a DAC.

ESS all the way. I use a twin parallel dac and it's awesome

For classical i prefer ESS ,for Rock and Reggae AKM (for the Bass) .

Isn't it ironic a company who swears by DSD uses a chip (ESS) that does not process DSD natively? :)

@ThinkingBetter

17 күн бұрын

I don't think so. The vast majority of music nowadays is in PCM and with streaming taking over, DSD as distribution format will only be in a further decline. It is only wise of ESS to put the design focus on PCM. On the input stage of their DSD supporting chips they do a fair job. These chips can do PCM decoding up to a rather irrelevant 768kHz 32 bits, which makes little sense. 96kHz 24 bits PCM already beats DSD-64. Besides, you can't actually create a native DSD master in a digital production process due to how all mixing and processing needs PCM and you end up having to transcode (lossy!) the output from PCM (e.g. DXD) to DSD. Yeah, someone could make an argument that DSD is a better ADC recording format out of older analog masters, but to gain maximum audience, it's better to ADC to PCM e.g. 192kHz 24 bits, which is already lossless streaming capable from major streaming services as FLAC lossless. So realistically, there is no future significant business in DSD and I can agree with Paul on ESS as best choice, especially considering AKM's situation.

@jorgemg1984

17 күн бұрын

@@ThinkingBetter I agree with everything you said, I'm even of the opinion that an ESS DAC with a good output stage renders moot most FPGA/discreet and R2R DACs these days - ESS sounds brilliant with PCM. But if you are aware of Paul's "opinions", he's constantly saying DSD is way better than PCM - and given that it's ironic he chooses a DAC chip (ESS) that converts DSD to a 6-bil format (so, NOT native). I'm just pointing out the irony... and BTW AKM is back in business for a while now (the Eversolo DMP-A8 uses it, amongst others), and their new chipset has a DSD Direct option on their firmware.

@ThinkingBetter

17 күн бұрын

@@jorgemg1984 Reality is PS Audio needs some unique differentiators but also some that are speaking tomorrow’s solutions rather than yesterday’s solutions. DSD made sense in the era of SACDs, in the past.

@necrodh

17 күн бұрын

Sometimes Paul is kilometers away from the truth, is not an expert in some audio departments

@jorgemg1984

17 күн бұрын

@@necrodh That seems to be the case

Gato is between very good and outstanding.

@oliverbeard7912

18 күн бұрын

I've enjoyed what I've heard from Gato.I think they're cosmetically nice to look at too.

@mikeg2491

18 күн бұрын

@@oliverbeard7912looks like a typical AliExpress hunk of silver case to me and for $4k no less lol but I won’t knock their sound quality until I’ve heard it.

why did your dac perform so poorly in Amir's tests

Interesting to hear your answer. I have 3 items that have the same DAC chip, the ESS sabre es9038q2m. A Topping x3 Pro+ headphone amp, a Arcam CDS50 CD player/streamer, and the Burson Audio Conductor x3 Performance. With the Topping and Arcam, the sound is pretty much identical, very clinical and neutral, whereas with the Burson Audio, the sound has much more life, and is much more enjoyable to listen to. Which makes sense considering the massive difference in cost and how the DAC is implemented.

I think Cambridge Audio uses Wolfson chips in some of their DACs. Are they any good and how do they compare to the one's discussed here? Or is it time for an upgrade lol

@Yiannis2112

18 күн бұрын

You're fine. Instead get a couple of diffusers and/or absorbers for your listening room.

@randomtube8226

18 күн бұрын

@@Yiannis2112 How many should I buy and what are the most critical placements I should start with?

@Yiannis2112

18 күн бұрын

@@randomtube8226 Most common places are behind your speakers and on the adjacent from your speakers, walls. For the latter for example, you'll have to use a mirror, which you'll have to drag on each wall, till you see the idol of each speaker on that mirror. That would be the center of the side panels. For small rooms generally absorption comes first. Probably absorption and/or diffusion on the ceiling too. All the above, are of course very, very general rules. The whole thing needs quite a bit of research from your part, to be honest. Its not really that straightforward, as the whole thing heavily depends on room geometry, number&size of windows, doors, furniture or lack of it. In terms, of what and exactly where and of course, why. Every room needs different treatment. Plenty of great guides-videos on yt on acoustic treatment. Vicoustic and GIK are two very good manufacturers with vast variety of products. Hope this helps to at least get started.

@Yiannis2112

18 күн бұрын

@@randomtube8226 Most common places are behind your speakers and on the adjacent from your speakers, walls. For the latter for example, you'll have to use a mirror, which you'll have to drag on each wall, till you see the idol of each speaker on that mirror. That would be the center of the side panels. For small rooms generally absorption comes first. Probably absorption and/or diffusion on the ceiling too. All the above, are of course very, very general rules. The whole thing needs a quite a bit of research from your part, to be honest. Its not really that straightforward, as the whole thing heavily depends on room geometry, number&size of widows, doors, furniture or lack of it. In terms, of what and exactly where and of course, why. Every room needs different treatment. Plenty of great guides-videos on yt on acoustic treatment. Vicoustic and GIK are two very good manufacturers with vast variety of products. Hope this helps to at least get started.

@Yiannis2112

18 күн бұрын

@@randomtube8226 Most common places are behind your speakers and on the adjacent from your speakers, walls. For the latter for example, you'll have to use a mirror, which you'll have to drag on each wall, till you see the idol of each speaker on that mirror. That would be where the first reflections are and where you'll place the side panels. For small rooms generally absorption comes first. Probably absorption and/or diffusion on the ceiling too.

RME had to switch from AKM to ESS for the DAC chip in their Adi-2 FS DAC. I heard both versions in a demo on a good system in a shop. Difference was hardly noticeable.

@ThinkingBetter

17 күн бұрын

Yes, the difference between well designed DACs is way way less audible than the differences between speakers and room acoustics.

DAC’s R2R high end it’s the best option up to now. No doubt.Let’s be honest.

@TimMihalko

15 күн бұрын

I too, believe R2R Dacs are the best way to go

From my limited personal experience, in general, ESS chip DACs have better detail retrieval, almost sharper sounding, compared to AKM chip DACs. It might be the chip, it might be the implementation might be the output stage, but out of the 15, 20 DACs that have sat in my system I found this to be the case.

Actually processing sound is not a super hard process for a modern dac or any modern dip chip. Your phone can output 4K 10 bit video at 60fps in real time and you don’t need fans or massive cooling. And actually inside the digital bit if you process with no errors or lag then the sound will be the same. It’s when you convert to analog where things can change. And this is why we moved to digital. It reduces the number of variables and possible errors

Obviously those S's in ESS stand for "sharp" and "shrill". If it was called ERW then it would be "relaxing" and "warm".

Thanks for a really great and incredibly succinct explanation of what must be a very large subject. I hadn't really thought of the processing before as a little computer inside our hi-fi equipment but it's mind boggling when you think of all that analogue (I'm English!) signal being chopped up in nanoseconds and being spewed out as 0's and 1's! It's even more mind-boggling when you think of how our ear's do this so neatly with the inner ear bones and hairs in the cochlea. Digital hi-fi is actually a bit of a retrograde stage in this, isn't it? We start of with analogue sound which hairs in the cochlea convert into pulses our brain responds to. Traditional hi-fi used vinyl records (or analogue radio signal), amplifiers and speakers whilst keeping it all analogue until it reaches our ears. Nowadays we worship introducing digital stages to all this: we convert our performances to digital and do all sorts of amazing processing on it and then convert the digital back to analogue before the last amplifier and speaker stages. And we call this progress!

Opampsmale the biggest dofference

Ultra Linear was the best I’ve heard even today. Too bad it was put out by the industry.

I have a usb burr brown dac with no extra op amps or circuit and it sounds marvelous, so its a lie that dacs doesn't matter or they depend on the circuit

does he do a better job? amir proved that your dac is weak😂

I am a BB63 fan. I do not like ESS.

ESS sounds fake and produce tinnitus, Philips was the most realistic and natural, also wolfson waz very natural, the rest just linear boring with no life and they need to compensate with filters and stuff sounding fake at the end of the day

@craigaust3306

17 күн бұрын

Wow! You sound like a real expert.

@necrodh

17 күн бұрын

@@craigaust3306 i am, i been tried a loooot of dacs in the last 10 years, is so confortable to say "all sound the same because there is only 1 and 0's and is not that simple.

PCM1792/1794 is still disliked (sometimes even hated) by most of audiophiles. Because it's still the one of rare examples of precisely working DACs, not bringing in his own trash into sound. The only competitors to these DACs are AKMs top of the line DACs. ESS is not even close to these beasts. Not in terms of sound perfection.

@housepianist

18 күн бұрын

Are you making these claims about differences in DAC’s within a dedicated circuit? It’s likely that your experiences with these different DACs comes from the analog output more than the DAC itself. Most of the time when manufacturers change/upgrade a DAC, they also change/upgrade the circuitry. Like Paul said, that makes the biggest difference. This is not to discredit your claim. I’m just saying there’s probably more to it than just the DAC itself.

@Ashen2501

18 күн бұрын

@@housepianist It's true that there's "more to it than just the DAC itself". But you'll always bump a limit eventually, and whatever you do, whichever parts you'll use, however you engineer your schematic your bottleneck and the main source of problems will be the DAC itself. It's not necessary to use extremely expensive caps, resistors with femtoapmere noise level and 0,0000001% level of precision, supa-dupa good power supply based on a private nuclear reactor to achieve a full potential of a good and properly working DAC. But even spending millions of dollars won't help you turn a shitty DAC like ESS into a good device. It WILL sound decently at some point, after your engineers pour a bunch of effort in it. But it still be a shitty overpriced DAC. Main problem is - most of audiophiles are uneducated people. Their mantra is "My multimillion system CAN'T sound bad, i've spent too much to admit that i'm that dumb".

@GBatya

18 күн бұрын

My favourite is pcm1792. It has very strong oversampling filter with 130dB stopband attenuation.

I prefer R2R ladder, to my ears they sound more musical than delta sigma. I know that delta sigma is technically better but hey, this hobby is subjective and it comes down to taste.