• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. 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!

Talk to me about bandwidth

Joined
Aug 1, 2020
Messages
7
Likes
1
Location
Bellingham, WA
Hello! My TEAC AI101-DA is dying so I need to replace it in my desktop computer setup. It handles everything for me - USB DAC, headphone amp, and speaker amp. I posted a separate thread for specific component recommendations, but during my research I had a question that I can't seem to find a clear answer to. Basically how much does bitrate/frequency matter, will any bottlenecks be audibly perceptible? Hopefully I'm not muddling/misusing too many terms, please don't hesitate to correct me or ask for clarification. For all examples assume I'm using a USB DAC capable of 32bit/768khz. I need to drive a pair of HiVi 3.1As (passive 4ohm 10w-150w speaker) and an SVS SB-2000 Pro powered sub. None of my headphones are anything special but I have the THX/Drop Pandas on order.

Say I send the audio for my passive speakers from a DAC into a Crown XLS1002. This will run an incoming analog signal through a 48khz AD/DA process to facilitate DSP. But 48khz is DVD quality/above CD quality, and my hearing stops at like 15khz, how much would I really be missing? Is there still a benefit from running a DAC even at 98khz into such an amp? What about 768khz?

In another potential build I may employ a MiniDSP 2x4HD, which also runs an analog signal through AD/DA, albeit at 96khz. Will there be a huge difference versus a bottleneck of 48khz? Or any perceptible difference, all other things being equal?

In both builds, the DSP will let me take the bass out of my speakers. My sub has a crossover but my speakers do not. If there is in fact a perceptible benefit to not bottlenecking at 48/96 at any point, is what I would lose with a bottleneck worth any potential clarity I would gain by crossing over the speakers to take out the bass? I've heard this will make for a cleaner and flatter sound in general.

TL;DR are either the "bottlenecked" AD/DA conversion or not taking bass out of my speakers really bad? If so, how bad? Which is the lesser of two evils?

Thanks a bunch!
 
Last edited:

NTK

Major Contributor
Forum Donor
Joined
Aug 11, 2019
Messages
2,656
Likes
5,819
Location
US East
The transparency/quality of AD/DA process in your Crown is more a matter of the implementation than the bit depth/sampling rate capabilities of the ADC/DAC chips. The bit depth/sampling rate of the Crown is sufficient. It is the actual implementation of the electronics that is usually the bottleneck. If your XLS 1002 is of similar performance as the XLS 1502 Amir measured, the bottleneck is probably in the amp section.

It is highly beneficial to high pass the mains when using sub(s). Main speakers, even "full range" ones, often struggle to reproduce deep bass. If that duty is offloaded to the sub(s), the bass drivers of the main speakers will have a much easier time having to only reproduce mid-bass and above. The result should be cleaner (less distorted) and higher mid-bass SPL capability from the mains [see notes for further explanations below].

If you have full DSP in your audio chain (as with the MiniDSP 2x4HD), in addition to relieving your mains from the deep bass duty, you can also apply room EQ to tame your room modes for much better bass response in your room, plus EQ your speaker if you have access to its anechoic measurements (e.g. if your speaker is one that Amir has measured). For room EQ, you'll need to invest in a measurement mic, such as the MiniDSP UMIK-1. The benefits of having DSP far outweighs any negatives. The process is not trivial and the learning curve is steep, especially for a beginner, but the reward will worth the trouble.

[Notes]
When reproducing sound at the same SPL (sound pressure level), physics dictates that the travel (movement) of the diaphragm has to increase when frequency is lowered. For example, if a driver produces 80 db SPL at 80 Hz with a diaphragm travel of 5 mm, it can only produce 68 dB SPL at 40 Hz with the same travel. Or, alternatively, it will need to travel 20 mm to produce 80 dB SPL at 40 Hz.

If the driver needs to reproduce both deep bass and mid-bass, the large diaphragm travel necessitated by the deep bass will cause higher distortions (because nonlinearities in the driver goes up quickly with travel), which will also be detrimental to the quality of the mid-bass.

Here is a hand-waving explanation on why SPL goes down by 12 dB when frequency is reduced by half (for the same travel):
SPL is related to pressure, and pressure is force per unit area. Force is directly proportional to the acceleration of the diaphragm, and acceleration is proportional to the square of frequency. Therefore, if frequency is reduced by half, pressure is reduced to a quarter. SPL is 20 log10(p / p_reference), therefore, when frequency is halved, there is reduction of 12 dB SPL at the same diaphragm travel.
 
OP
vertical_mammal
Joined
Aug 1, 2020
Messages
7
Likes
1
Location
Bellingham, WA
Awesome, thanks! Really helpful. I'm hearing the main point as DSP is worth it.

For specific implementation, I'm definitely off the idea of the Crown stuff. I think I will end up with an SMSL amp, either the DA-8s or the SA300. I was considering using an SMSL M500 DAC which would let me send RCA to my powered sub (which has its own DSP and crossover) and XLR to the amp, but I would not be able to take the bass out of the mains in this setup unless there is a highpass on the DA-8s or SA300 that I'm missing - they do both have limited DSP. And this would also leave me without room correction. Would you say that my PC sending USB to an iFi Zen DAC, then RCA to a MiniDSP 2x4HD which will serve as room correction and crossover for both sub and mains and send RCA directly to my sub as well as to my speaker amp of choice, definitely be better? In this example I am trading a higher-resolution and more modern DAC and a balanced signal path for room correction and high-passed mains. If I'm understanding you properly you'd probably say that yes, the second setup will provide a quantifiably better listening experience?

Thanks again!!
 

NTK

Major Contributor
Forum Donor
Joined
Aug 11, 2019
Messages
2,656
Likes
5,819
Location
US East
IMO if you are using a MiniDSP 2x4HD, you shouldn't need the iFi DAC (unless you need it for headphones). The MiniDSP will just A/D the signal, process it, and D/A it out. You are ultimately limited by the A/D D/A capability of the MiniDSP. Why not just feed digital signals directly to the MiniDSP, and save the D/A process (by the iFi DAC) and A/D process (by the MiniDSP).
 

dc655321

Major Contributor
Joined
Mar 4, 2018
Messages
1,597
Likes
2,235
Here is a hand-waving explanation on why SPL goes down by 12 dB when frequency is reduced by half (for the same travel):
SPL is related to pressure, and pressure is force per unit area. Force is directly proportional to the acceleration of the diaphragm, and acceleration is proportional to the square of frequency. Therefore, if frequency is reduced by half, pressure is reduced to a quarter. SPL is 20 log10(p / p_reference), therefore, when frequency is halved, there is reduction of 12 dB SPL at the same diaphragm travel.

That's not hand-wavey at all - that's just dimensional analysis. Engineers and physicists use that all the time.
 
OP
vertical_mammal
Joined
Aug 1, 2020
Messages
7
Likes
1
Location
Bellingham, WA
Yeah, headphones need to come into the mix somewhere. I don't have balanced headphones but may eventually, and am considering the THX 789 headphone amp since I have some Drop credit. Right now I have Sennheiser HD 380 Pros, with the THX Pandas on order, and a mind to get some open-backed cans further down the road. I think it would be nice to be able to feed the 789 with XLR even with these headphones, but maybe that is not worth the price of a DAC that can do that. If I forego that I could use the MiniDSP as the DAC with two channels out to my mains amp and two channels out to the 789, passed through to the powered sub. Besides losing XLR to the 789 it also splits up my DSP, forcing me to do all room corrections for the sub in the sub, instead of being able to do everything in the MiniDSP. But it saves me a few hundred dollars on a separate DAC, like for instance the SMSL M300 mk2. What do you think?
 

NTK

Major Contributor
Forum Donor
Joined
Aug 11, 2019
Messages
2,656
Likes
5,819
Location
US East
That's not hand-wavey at all - that's just dimensional analysis. Engineers and physicists use that all the time.
Haha! You got a fancy name for that. Obviously I am no expert in dimensional analysis. Any knowledge of it has been long returned to my fluids dynamics professor (only place I remember seeing it).

Yeah, headphones need to come into the mix somewhere. I don't have balanced headphones but may eventually, and am considering the THX 789 headphone amp since I have some Drop credit. Right now I have Sennheiser HD 380 Pros, with the THX Pandas on order, and a mind to get some open-backed cans further down the road. I think it would be nice to be able to feed the 789 with XLR even with these headphones, but maybe that is not worth the price of a DAC that can do that. If I forego that I could use the MiniDSP as the DAC with two channels out to my mains amp and two channels out to the 789, passed through to the powered sub. Besides losing XLR to the 789 it also splits up my DSP, forcing me to do all room corrections for the sub in the sub, instead of being able to do everything in the MiniDSP. But it saves me a few hundred dollars on a separate DAC, like for instance the SMSL M300 mk2. What do you think?
That's is a difficult question without a simple answer, IMO. The difficulty is that headphones EQ (if you decide to EQ your headphones) and speakers/room correction EQ/bass management are totally different and separate things.

Warning: I have never worked with one myself, and am thinking out load based on what I read from the 2x4HD manual. So no guarantee that it will work.
If you can accommodate some inconvenience, I think you can setup different preset configurations with the MiniDSP 2x4HD. For example, you can configure preset 1 for headphone use, and route channels 1 & 2 to your headphone amp and mute channels 3 & 4. You can then setup preset 2 for speakers (with the speaker EQ settings), and run channels 1 & 2 (using y-splitters) to your speaker amp. Channels 3 & 4 go to the sub(s). You'll obviously need to mute/power off the speaker amp if you are listening to headphones. I propose channels 3 & 4 for the sub(s) with the hope that no manual switching will be needed for the sub(s). You can toggle between the presets with the remote control (a USD$5 option).
 
OP
vertical_mammal
Joined
Aug 1, 2020
Messages
7
Likes
1
Location
Bellingham, WA
Haha! You got a fancy name for that. Obviously I am no expert in dimensional analysis. Any knowledge of it has been long returned to my fluids dynamics professor (only place I remember seeing it).


That's is a difficult question without a simple answer, IMO. The difficulty is that headphones EQ (if you decide to EQ your headphones) and speakers/room correction EQ/bass management are totally different and separate things.

Warning: I have never worked with one myself, and am thinking out load based on what I read from the 2x4HD manual. So no guarantee that it will work.
If you can accommodate some inconvenience, I think you can setup different preset configurations with the MiniDSP 2x4HD. For example, you can configure preset 1 for headphone use, and route channels 1 & 2 to your headphone amp and mute channels 3 & 4. You can then setup preset 2 for speakers (with the speaker EQ settings), and run channels 1 & 2 (using y-splitters) to your speaker amp. Channels 3 & 4 go to the sub(s). You'll obviously need to mute/power off the speaker amp if you are listening to headphones. I propose channels 3 & 4 for the sub(s) with the hope that no manual switching will be needed for the sub(s). You can toggle between the presets with the remote control (a USD$5 option).

Thanks again! For a variety of reasons I have decided to still go with a DAC, including foregoing the inconveniences you mentioned. Really the biggest part of it for me is a master volume control. The DAC I've chosen will serve as a preamp, which is useful to me because if I go straight from the PC to the MiniDSP then I'll either have to set sub and mains volume separately (ugg) or use Windows volume as my master gain (double ugg). The other big benefit of the DAC I've chosen is that it is much higher resolution than the MiniDSP and will serve balanced XLR directly to my chosen headphone amp. For headphones specifically I'd rather have the numerous benefits of that setup over any correction the MiniDSP will offer me. Another big factor that I haven't mentioned now is that my chosen headphone amp won't even be delivered until November, and my chosen DAC has a headphone amp that is good enough until then.

So currently the planned endgame is PC > SMSL M500 > XLR to THX 789 for headphones. Then PC > SMSL M500 > RCA to MiniDSP 2x4HD for correction and crossover > 2 channels to SMSL SA300 speaker amp serving my HiVi 3.1a passive speakers and 1 channel to SVS SB2000 Pro powered sub. Pretty stoked!
 

NTK

Major Contributor
Forum Donor
Joined
Aug 11, 2019
Messages
2,656
Likes
5,819
Location
US East
Awesome. I'm sure your setup will work great. For me, the lack of a physical volume knob with the MiniDSP is a major usability hit too. Sometimes these seemingly minor non-technical flaws actually matter a lot.
 
OP
vertical_mammal
Joined
Aug 1, 2020
Messages
7
Likes
1
Location
Bellingham, WA
Awesome. I'm sure your setup will work great. For me, the lack of a physical volume knob with the MiniDSP is a major usability hit too. Sometimes these seemingly minor non-technical flaws actually matter a lot.

For sure. I'm super-stoked, and have started to order gear. I have to spread it out over several checks, but that's actually kind of a benefit here. My plan lets me leave the MiniDSP until last, since my amp has a sub out of its own. So I can get to know the full system. Then I can put the MiniDSP in the path and listen for awhile to see if I can hear the loss of resolution. From there I'll use a UMik to run room correction and also introduce the highpass to the mains. So I can go just one step at a time, and try to hear the changes at each and every step. And then once it's all said and done, it's as easy as taking the MiniDSP back out of the chain to do final A/B. I think it's called science? Going to be really fun, did I mentioned that I'm super-stoked? ;p
 
  • Like
Reactions: NTK

KeithPhantom

Addicted to Fun and Learning
Forum Donor
Joined
May 8, 2020
Messages
641
Likes
658
TL;DR are either the "bottlenecked" AD/DA conversion or not taking bass out of my speakers really bad? If so, how bad? Which is the lesser of two evils?
Interpolation techniques in DAC/ADC are used to simplify the filter characteristics and improve measurements such as SINAD, SFDR, etc. They do not affect anything in the realms of FR. Something that can have some issues in some badly designed software is decimation by non-perfectly-divisible factors (96-to-44.1 kHz and other factors, a non-issue nowadays), but for all of these issues, better use dither because it will decorrelate possible quantization errors from the noise floor, averaging these results to conform something with less energy than the correlated noise floor of an undithered signal.
 
Last edited:
Top Bottom