• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Sheffield Labs "My Disc" -70dB Test Track

Joined
Jan 23, 2020
Messages
60
Likes
38
Location
VA
#1
Amirm, are you (or anyone) familiar with the Sheffield Labs / A2TB test disc? (It’s available on Amazon Music HD to stream in FLAC) Sans any test equipment that you have (really enjoy all the measurement based reviews btw), I’ve used it over the years to confirm improvements (or not) in my system. There is a track of digital black (41) which is useful to get a baseline of system noise floor with the volume turned up to max. It also has a series of tracks (33-40) progressively reducing the recorded volume of a track down to -70dB. This would seem an excellent way to hear the difference between great and poor DACs.
It’s a HT/Stereo system compromise, but I use a Denon X4400H as a preamp (in 11.1 mode which disconnects the amps) feeding a Parasound A21 and Tekton Pendragons. So, my SINAD is at best in the 90’s. With the volume on max, the -70db track is 50/50 noise and moderately distorted music. So, I got a Topping DX3Pro (106 SINAD) and wired it direct to A21 expecting a decent improvement on the -70dB track. To my disappointment, it was much fainter and even more distorted. Any thoughts on why? Maybe the track is recorded distorted? Maybe you could analyze it and see. Also wonder what it sounds like with your Mark Levinson DAC? According to the booklet that came with the test CD, a good system should reproduce the track clearly.
 

Frank Dernie

Major Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
3,122
Likes
5,886
Location
Oxfordshire
#2
Amirm, are you (or anyone) familiar with the Sheffield Labs / A2TB test disc? (It’s available on Amazon Music HD to stream in FLAC) Sans any test equipment that you have (really enjoy all the measurement based reviews btw), I’ve used it over the years to confirm improvements (or not) in my system. There is a track of digital black (41) which is useful to get a baseline of system noise floor with the volume turned up to max. It also has a series of tracks (33-40) progressively reducing the recorded volume of a track down to -70dB. This would seem an excellent way to hear the difference between great and poor DACs.
It’s a HT/Stereo system compromise, but I use a Denon X4400H as a preamp (in 11.1 mode which disconnects the amps) feeding a Parasound A21 and Tekton Pendragons. So, my SINAD is at best in the 90’s. With the volume on max, the -70db track is 50/50 noise and moderately distorted music. So, I got a Topping DX3Pro (106 SINAD) and wired it direct to A21 expecting a decent improvement on the -70dB track. To my disappointment, it was much fainter and even more distorted. Any thoughts on why? Maybe the track is recorded distorted? Maybe you could analyze it and see. Also wonder what it sounds like with your Mark Levinson DAC? According to the booklet that came with the test CD, a good system should reproduce the track clearly.
A 16-bit sound at -70dB has quite a high distortion level.
Personally I set the volume control at a level suited to my music listening before trying any low level tones since that gives a much better idea, IMHO, of how audible/intrusive they may be when listening.
 
OP
C
Joined
Jan 23, 2020
Messages
60
Likes
38
Location
VA
Thread Starter #3
Good point. I didn't realize at 16 bit it would be inherently distorted. Agreed, it'd be inaudible at normal listening levels. Which makes me realize although my receiver as a pre-amp could be technically improved upon, an investment here probably wouldn't realize any real world audible improvement. Speakers are probably by far the biggest contributors to overall system distortion ......not to mention the room! Best spend my money there.
 

BDWoody

Major Contributor
Forum Donor
Joined
Jan 9, 2019
Messages
2,168
Likes
4,123
Location
Mid-Atlantic, USA. (Maryland)
#4
Good point. I didn't realize at 16 bit it would be inherently distorted. Agreed, it'd be inaudible at normal listening levels. Which makes me realize although my receiver as a pre-amp could be technically improved upon, an investment here probably wouldn't realize any real world audible improvement. Speakers are probably by far the biggest contributors to overall system distortion ......not to mention the room! Best spend my money there.
Excellent conclusion!
So many could realize better sound by doing exactly that. The Room/Speaker system is so often neglected... Makes my musical heart hurt when I see a high quality 'system' with speakers poorly placed in a room that's treated as an afterthought.
 
OP
C
Joined
Jan 23, 2020
Messages
60
Likes
38
Location
VA
Thread Starter #5
Absolutely! I'm lucky enough to have a dedicated room and an afternoon with REW shuffling things around can make a mind blowing difference to in-room response. I just use Audyssey XT32 below 300Hz for room modes which makes a big difference but unfortunately down samples everything to 48/24 in the process.
 

Herbert

Active Member
Joined
Nov 26, 2018
Messages
194
Likes
168
#6
I have the YEDS-18 Test Disc from Sony, which also has a -60dB Test Signal and digital silence-track.
On an older resistor-network based CD-Player, this will be like a sawtooth.
On modern designs, it ia (almost) a sinewave again.
But the digital silence - track is very revealing. Very interstingly, the most quiet player is a second gen. Philips (CD-304MKII)
No hiss at all when the volume of my amp is all way up,
but the servo can be heard "ticking / kicking". A 1-bit based Sony (XA-50ES) has a lot of hiss...
 

SIY

Technical Expert
Technical Expert
Joined
Apr 6, 2018
Messages
4,244
Likes
8,174
Location
Phoenix, AZ
#7
A 16-bit sound at -70dB has quite a high distortion level.
Noise level with respect to signal amplitude, not distortion, assuming dither (which is pretty universal except for people deliberately trying to demonstrate quantization distortion).
 
OP
C
Joined
Jan 23, 2020
Messages
60
Likes
38
Location
VA
Thread Starter #8
Interesting. Does anybody know of, or ideally could create, a 16 & 24 bit -70dB test track? You could then appreciate the improvement of 16-24 bit depth on the same DAC and also compare a low SINAD DAC to a higher one? I think all the review measurements are great but it's always nice to boil some of it down to an audible experiment to demonstrate the good from bad. I'm sure there are better ideas to do this also for different performance metrics.
 

Herbert

Active Member
Joined
Nov 26, 2018
Messages
194
Likes
168
#9
I should be able to do this. But I need to check if my DAW can produce a testtone from scratch.
But maybe there are already files to be found on the net.
 
OP
C
Joined
Jan 23, 2020
Messages
60
Likes
38
Location
VA
Thread Starter #10
That's great!! Another idea is to also do the tracks at different sampling rates to see if that can be perceived more easily at low levels........
 

SIY

Technical Expert
Technical Expert
Joined
Apr 6, 2018
Messages
4,244
Likes
8,174
Location
Phoenix, AZ
#11
Interesting. Does anybody know of, or ideally could create, a 16 & 24 bit -70dB test track? You could then appreciate the improvement of 16-24 bit depth on the same DAC and also compare a low SINAD DAC to a higher one? I think all the review measurements are great but it's always nice to boil some of it down to an audible experiment to demonstrate the good from bad. I'm sure there are better ideas to do this also for different performance metrics.
It's a simple thing, but already well-known. The difference is noise floor.. So if you set 0 dB to "loudest my system will play" (say 110dB SPL), then the 16 bit will have a noise floor at 96 dB below that, or 14dB SPL, well below the noise of any real-world room. Further improvements won't be heard unless you artificially twiddle the volume knob.
 
Joined
Jan 8, 2020
Messages
86
Likes
46
#12
I have a torrent download link to the Sheffield Labs disc image. If anyone is interested and it does not violate the forum's terms, I can provide a link.
 

Fluffy

Addicted to Fun and Learning
Joined
Sep 14, 2019
Messages
705
Likes
1,031
#13
Interesting. Does anybody know of, or ideally could create, a 16 & 24 bit -70dB test track? You could then appreciate the improvement of 16-24 bit depth on the same DAC and also compare a low SINAD DAC to a higher one? I think all the review measurements are great but it's always nice to boil some of it down to an audible experiment to demonstrate the good from bad. I'm sure there are better ideas to do this also for different performance metrics.
Here you go: https://we.tl/t-zvhtH3gRRC

I took the number 33 track ("music recorded with normal transfer level of 0 db"), and did the following:

First of all, there is the original track, unchanged, at 16 bit (file named "original").

Then I converted it to 24 bit and 32 bit, and applied a -70 db attenuation to all 3 bit depths (files names "16" "24" 32").

Then I renormalized those three files back to 0 db (files named "16/24/32 – normalized").

As you can hear, normalizing the 16 bit file sounds awful, since the -70 db decrease in volume pushed it well into the noise floor of the 16 bit file. The normalized 24 bit sounds ok, but there is still an increased noise level. The normalized 32 bit is the only one that preserved the entire dynamic range of the original and is completely lossless.
 
OP
C
Joined
Jan 23, 2020
Messages
60
Likes
38
Location
VA
Thread Starter #14
Awesome! Really appreciate you doing this! For some reason the 16/24/32 track just say buffering using JR16 but the normalized ones work.
The 16 bit normalized gives about the same noise to music ratio I get with the AVR turned up to max on the original -70dB track on the disc but I get much more distortion (sounds like clipping). When I used the topping DX3 pro, which should be much better, I also got the hard clipping. When you play the original track on your equipment turned up, do you get that? Again, many thanks for taking the time!
 
OP
C
Joined
Jan 23, 2020
Messages
60
Likes
38
Location
VA
Thread Starter #15
.......got the 16/24/32 tracks to play using media center of all things. Wow, I think the tracks you created are much better quality than the originals!!! - no distortion to speak of. Just as you describe, and the 32 bit track has very little background noise and is very clear. My Denon X4440H seems to be doing a pretty good job as a pre-amp after all....... better than the DX3pro if this is a fair test.
 

Fluffy

Addicted to Fun and Learning
Joined
Sep 14, 2019
Messages
705
Likes
1,031
#16
I actually had difficulty achieving 70 db of analog gain with anything I own. My speaker set up is not up to it, so it took connecting high sensitivity headphones to a powerful headphone amp. Even then, the attenuated 16 bit file at maximum gain was quieter than the normalized 16 bit at minimum gain. It only goes to show how large a 70 db difference is. It's a dynamic range no one ever really needs.

Anyway, with my setup the max gain attenuated file sounded like the minimum gain normalized file, no added distortion/clipping and such. It just sounds like a lot of added noise. This is how it supposed to be. If you test this and have added distortions/obvious clipping or noise that is not originating from the digital noise level, you have a problem. it's either the device itself (like a dac that can't actually reproduce 16 bit worth of dynamic range), or maybe there is a problem with your gain structure. For example, a dac feeding a too-hot of a signal to a pre-amp, getting distorted there or at the amplifier input.

This test can determine if the device is capable of better than 16 bit performance. If the dac has a SNR of 96 db or lower, that means it can only cover 16 bit dynamic range (or worse, if it's lower). In this test, that means amplifying the 16 bit and 24 bit samples would sound the same – the SNR is determined by the limits of the DAC, not the file. If the dac is capable of higher than 16 bit, like SNR of 110 db, than amplifying the 24 sample would sound better than the 16 bit one. I doubt any dac in the world could surpass 24 bit for it would require over 144 db of SNR. Such a dac would show the difference even between the 24 and 32 bit files.

By the way, I forgot to mention that the original track was not actually normalized to 0 DBFS. That means that when I normalized it, it became louder than the original. You shouldn't confuse loudness with better quality – that's a common mistake, and that's why you should always level-match when doing comparisons. The tracks I normalized have the same exact sound quality as the original (minus the added digital noise, of course), just a bit louder.
 
OP
C
Joined
Jan 23, 2020
Messages
60
Likes
38
Location
VA
Thread Starter #17
Got it. For whatever reason the 16 bit -70dbfs track you created has the same noise/music level as the original when turned up to max but yours has much less (clipping) distortion. Yes, it put's it into perspective how irrelevant much higher SINAD performance is under normal listening conditions. It makes me wonder then what the sound quality difference is between DACs that people claim to hear clearly. Maybe it's the analog output stage (quality of?) and how it interacts with the amplifier more than the DA conversion itself?
 

Fluffy

Addicted to Fun and Learning
Joined
Sep 14, 2019
Messages
705
Likes
1,031
#18
Got it. For whatever reason the 16 bit -70dbfs track you created has the same noise/music level as the original when turned up to max but yours has much less (clipping) distortion.
Ok, that's weird. I did something I haven't done the first time – I took the original track in -70 db from the album (track 40) and normalized it. It does indeed sound very different, as you can hear here: https://we.tl/t-AgVFDxlxr7

The original is from the album (track 40 normalized). The digitally is what I made from track 33.

It's now clear to me they didn't go through the same process. It's not a simple digital attenuation by 70 db, but something more complex. Maybe they used dithering or something, or maybe it actually went through a physical pot that attenuated the level. I wonder if they say that anywhere.

Anyway, that means that the distortion you heard comes from their file, not from your system.
 
OP
C
Joined
Jan 23, 2020
Messages
60
Likes
38
Location
VA
Thread Starter #19
All it says in the booklet is "Even track 40 at -70dB, can be clearly heard on an excellent system". Glad you helped me get to the bottom of this before I dropped $10k on a DAC:)
 
Top Bottom