• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Why do many "high-end" DACs have low specs?

curiouspeter

Addicted to Fun and Learning
Joined
Dec 30, 2020
Messages
623
Likes
401
Location
San Francisco Bay Area
Just for fun, I was browsing the specs of a few "high-end" Roon Ready DACs. It turns out many of them are limited at PCM 24/192 and/or DSD64 over network. Why is that?

Are those companies deliberately "keeping the menu small" to give the illusion of quality?

One would assume those who can afford these fixtures would try to feed them with the best source material. Many downloadable files were recorded at DSD256 or DXD. So they expect people to downsample? It is so funny.
 
The limitation has puzzled me too since higher performance DAC chips are easy enough to incorporate into a new design. Manufacturing cost wise, I'm at a loss to see how some relatively expensive DACs use older generation DAC chips. I get why some very expensive DACs use the relatively ancient resistive ladder chips (NOS) from decades ago - its just purely marketing mystique BS which sells to some people.
 
Many downloadable files were recorded at DSD256 or DXD.

Are they?

The DSD catalogue is rather limited.
Only a few companies record in this format.

DXD is simply 24 bit / 352.8 kHz PCM.
It was invented because DSD is not editable.
Convert to PCM, edit and convert back to DSD is the only option to edit a DSD recording.
 
Are they?

The DSD catalogue is rather limited.
Only a few companies record in this format.

DXD is simply 24 bit / 352.8 kHz PCM.
It was invented because DSD is not editable.
Convert to PCM, edit and convert back to DSD is the only option to edit a DSD recording.
The catalog is quite sizable for classical and jazz.
 
Are they?

The DSD catalogue is rather limited.
Only a few companies record in this format.

DXD is simply 24 bit / 352.8 kHz PCM.
It was invented because DSD is not editable.
Convert to PCM, edit and convert back to DSD is the only option to edit a DSD recording.
I don't know if DSD is editable personally, but in several PS Audio videos, Paul mentions that they can edit DSD on their Sonoma system. It might be BS like in a lot of his videos.
 
It matters when files won't play without extra processing.
More of an inconvenience, isn't it? Where this are downloads, there is generally software that can resample things during playback and/or offline as well. It's not too different from keeping a secondary lossy version for your portable player.
Many downloadable files were recorded at DSD256 or DXD.
Not sure about the "many" part. DSD is such a silly, impractical format, you can't really do anything with it. I know there are ADCs supporting it, but anything but the most rudimentary editing will involve converting to PCM. You would see levels exactly as recorded, no compression or even normalization at all, nor any fadeouts. If there is evidence of any of that, chances are the material was just converted.
Manufacturing cost wise, I'm at a loss to see how some relatively expensive DACs use older generation DAC chips.
Hmm. Care to name some examples? I can only think of a few "oldies" I would still consider competitive, e.g. PCM1792/94(A) and CS4398.

Indeed, going multibit is generally about "mystique", even if you could in principle apply the same techniques that modern multilevel delta-sigma converters use.
 
I don't know if DSD is editable personally, but in several PS Audio videos, Paul mentions that they can edit DSD on their Sonoma system. It might be BS like in a lot of his videos.
1-bit DSD cannot be edited, but multi-bit PDM can be used to some extent.

Even Roon can do DSP on DSD using multi-bit PDM.
 
Not sure about the "many" part. DSD is such a silly, impractical format, you can't really do anything with it. I know there are ADCs supporting it, but anything but the most rudimentary editing will involve converting to PCM. You would see levels exactly as recorded, no compression or even normalization at all, nor any fadeouts. If there is evidence of any of that, chances are the material was just converted.

Some types of editing can be done in PDM without converting to PCM. Even volume leveling and EQ.

DSD is a reasonable format since most ADCs and DACs are D/S.
 
2 possibilities:

1. bad engineering

2. clever, euphonic engineering

3. Clever marketing

They cannot match the spec/price ratios of upstarts like Topping/Gustard/Matrix so they reduce the specs to highlight everything else, like that "deep soundstage" and "warmth".
 
Why not? Frequently you only have a choose of various DSD formats and maybe DXD on NativeDSD. It's one of the best sites.

DSD was a solution in search of a problem that didn't exist in consumer audio in order to sell a licensing scheme. Since people began figuring that out, it is and has been on the way out.

So now you know why DSD support in DACs is no longer popular.
 
Last edited:
DSD was a solution in search of a problem that didn't exist in consumer audio in order to sell a licensing scheme. Since people began figuring that out, it is and has been on the way out.

So now you won't know why DSD support in DACs is no longer popular.
I don't know what you're talking about because I still get notifications for new DSD releases all the time.

It's not unreasonable to choose a consumer format that matches the mastering format, especially in classical genres.

Almost all new mid-fi DACs have DSD support.
 
I don't know what you're talking about because I still get notifications for new DSD releases all the time.

It's not unreasonable to choose a consumer format that matches the mastering format, especially in classical genres.

Almost all new mid-fi DACs have DSD support.

Yep. There's still some audiophiles out there buying DSD while most people have moved on.
 
If the sites offer 24/96 just get that .

Many DAC's do support DSD and very high PCM rates on their USB , but network streamers for whole house audio rarely do because it would be nuts to stream that kind of bitrate over the network, many also wants wifi for their endpoints ?
Most software intended for streaming have facilities built in for down sampling . LMS and Roon for example.

You need that anyway because Room correction and EQ and subwoofer xover and other facilities you have that are considerable boons for good sound while insisting on DSD down to the endpoint is not .

The raspberry-Pi i use with picore player supports 24/192 for the digital coax board i use . IN the case of my headphone DAC i can use 24/192 my main hifi can use 24/96 because that's the limit for the meridian DSP speakers and their digital pre-pro.

I also EQ the headphones so thet system is also feed by my PC but because the EQ DSD is not an option as output .

I do have some very few DSD releases because that's what the options i had .

Buy the music you like be a bit more format pragmatic :) don't look yourself in to certain solutions.
 
I don't know if DSD is editable personally, but in several PS Audio videos, Paul mentions that they can edit DSD on their Sonoma system. It might be BS like in a lot of his videos.
DSD is indeed impossible to do even the most basic of edits on. Sonoma does it by converting to PCM and back behind the scenes.

1-bit DSD cannot be edited, but multi-bit PDM can be used to some extent.
That doesn't make sense. Are you sure you didn't mean PCM?

DSD is a reasonable format since most ADCs and DACs are D/S.
DSD is unreasonable because it is a 1-bit format. Modern converters are multi-level sigma-delta, so there's no sense in using DSD anywhere.

Yep. There's still some audiophiles out there buying DSD while most people have moved on.
Most people never went there in the first place.
 
If the sites offer 24/96 just get that .

Many DAC's do support DSD and very high PCM rates on their USB , but network streamers for whole house audio rarely do because it would be nuts to stream that kind of bitrate over the network, many also wants wifi for their endpoints ?
Most software intended for streaming have facilities built in for down sampling . LMS and Roon for example.

You need that anyway because Room correction and EQ and subwoofer xover and other facilities you have that are considerable boons for good sound while insisting on DSD down to the endpoint is not .

The raspberry-Pi i use with picore player supports 24/192 for the digital coax board i use . IN the case of my headphone DAC i can use 24/192 my main hifi can use 24/96 because that's the limit for the meridian DSP speakers and their digital pre-pro.

I also EQ the headphones so thet system is also feed by my PC but because the EQ DSD is not an option as output .

I do have some very few DSD releases because that's what the options i had .

Buy the music you like be a bit more format pragmatic :) don't look yourself in to certain solutions.

Wi-Fi is crazy fast nowadays. I only have a Google Wifi network but 100+ Mbps is easily sustainable. Also, most "high-end" streaming DACs are too snobbish for wi-fi anyway and they insist on Ethernet. That is easily 500 Mbps or higher.

PEQ is quite efficient on the DSD stream. My core (Synology NAS) cannot do native processing with convolution but about 9x with an 8-band PEQ, which does a pretty good job correcting my room. I also throw in volume-leveling because I hate adjusting things all the time. :)

An i7 desktop (or even a new Mac Mini) should be able to do convolution on native DSD256.



DSD is indeed impossible to do even the most basic of edits on. Sonoma does it by converting to PCM and back behind the scenes.


That doesn't make sense. Are you sure you didn't mean PCM?

I meant PDM, Pulse Density Modulation. Multi-bit DSD is not PCM. Are you sure Sonoma is not converting everything to wide-DSD?

To be fair, for most genres, I would not care to look for even hi-res. However, for certain music, like solo Piano/Harp/Harpsichord, I just feel like getting a version with an interesting format. I probably cannot tell the difference, but hey, it's a ritual. :p
 
Back
Top Bottom