• 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!

DIYSG HTM-12v1 Speaker Review

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
Does your receiver have digital PEQ already? Many do. Also, does your receiver have room correction? That’s another way to get it to make corrections, even up in the speaker range. I would experiment with that first.

In this case, I would let it correct the whole curve, not just below Schroeder frequency. This is because the speaker badly needs correcting. The built in room eq might have enough horsepower to calculate all the corrections you need. I would re-measure at the end with REW to confirm you’re getting a nice smooth curve. That might be all you need. iIf your receiver’s DSP capability is too limited, it might not be able to make all of the corrections needed.

IMO, an “affordable“ amp+DSP combo that can also perform surround processing duties is the Monoprice HTP-1 with hypex amps like @Buckeye Amps. This will give you 16 PEQ slots of DSP per channel, plus state of the art Dirac room correction. And it can decode everything and has nice upmixers like Auro3D. It is not “the most affordable”—you can definitely remove features and pay less. But it’s very affordable for what you get.

For your multichannel situation, there is really no benefit with going fully active. You would need 2 amps per speaker, doubling amp costs, and more advanced hardware/software to make the digital crossovers. You could do this with a Trinnov unit, using 2 outputs per speaker and having it domthe crossovers, speaker eq, and room eq—but that is $$$!

What's the point of having the Klippel data if you're not going to use it at all and let the AVR just do the good old one fits all room correction?

I think the best way is to get around an issue like this is a MiniDSP 2*4 if your AVR have both pre-outs and pre-ins. it's cheap, and versatile.
 

yourmando

Active Member
Joined
Aug 14, 2020
Messages
150
Likes
178
What's the point of having the Klippel data if you're not going to use it at all and let the AVR just do the good old one fits all room correction?

I think the best way is to get around an issue like this is a MiniDSP 2*4 if your AVR have both pre-outs and pre-ins. it's cheap, and versatile.
Great question. I asked the same question myself and @mitchco had a great answer:

https://audiophilestyle.com/forums/...designer/page/2/?tab=comments#comment-1126710

In short, if you are using a state of the art room eq on a platform with enough processing power, then using the anechoic Klippel data is not needed. This is why I suggested he *try* his room EQ over the whole speaker curve, then *confirm* the curve is nice and smooth with REW. Then he can decide whether to add a device or upgrade is pre-pro.

Adding the miniDSP to add PEQs would work fine, if needed. As would my suggestion of using an HTP-1, which has the DSP capability plus SOTA room correction and his surround processing requirements, all in 1 unit with no extra conversions or additional clipping concerns. But it’s worth trying what he has first is all I’m suggesting.
 
OP
hardisj

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
Not trying to rain on the parade but I do want to remind everyone watching that the test here is of the "v1" version. And that the "v2" version review is coming soon and shows more promise (albeit, at a loss of sensitivity). So, I wouldn't make any purchase decisions just yet. It might be worth it to simply retrofit a v1 design with v2 components.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
Great question. I asked the same question myself and @mitchco had a great answer:

https://audiophilestyle.com/forums/...designer/page/2/?tab=comments#comment-1126710

In short, if you are using a state of the art room eq on a platform with enough processing power, then using the anechoic Klippel data is not needed. This is why I suggested he *try* his room EQ over the whole speaker curve, then *confirm* the curve is nice and smooth with REW. Then he can decide whether to add a device or upgrade is pre-pro.

Adding the miniDSP to add PEQs would work fine, if needed. As would my suggestion of using an HTP-1, which has the DSP capability plus SOTA room correction and his surround processing requirements, all in 1 unit with no extra conversions or additional clipping concerns. But it’s worth trying what he has first is all I’m suggesting.

it really depends on your objective, if you're optimizing for one location for one listener then yes you don't need the anechoic data, as long as your speaker and amplifier have enough headroom to allow the necessary modifications then you can achieve whatever response you want in that one location.

however the Olive/PIR score and model originated from the spatial average of 9 measurements and the location of each measurement was 3 meters apart from its next neighbor. it's a massive difference, and it's vital that both the on-axis and off-axis responses are consistent with each other to score high on the model.

that's why I like to optimize for PIR first, so the speaker's performance is more or less consistent throughout the whole room, and target curve deviations are minimized. this way something like Dirac would not need to make unnecessary assumption. Mitch says that it's important that the room correction is SOTA, I don't think any affordable room correction on the market is SOTA.
 

yourmando

Active Member
Joined
Aug 14, 2020
Messages
150
Likes
178
it really depends on your objective, if you're optimizing for one location for one listener then yes you don't need the anechoic data, as long as your speaker and amplifier have enough headroom to allow the necessary modifications then you can achieve whatever response you want in that one location.

however the Olive/PIR score and model originated from the spatial average of 9 measurements and the location of each measurement was 3 meters apart from its next neighbor. it's a massive difference, and it's vital that both the on-axis and off-axis responses are consistent with each other to score high on the model.

that's why I like to optimize for PIR first, so the speaker's performance is more or less consistent throughout the whole room, and target curve deviations are minimized. this way something like Dirac would not need to make unnecessary assumption. Mitch says that it's important that the room correction is SOTA, I don't think any affordable room correction on the market is SOTA.
I agree with all of that. And SOTA room correction does take care not to overcorrect by taking multiple measurements and focusing correction on deviations with low variance across the entire listening space, as you suggest.

Affordable is subjective but IMO something like the HTP-1 is quite affordable for multichannel + features and would allow for both the anechoic speaker PEQ correction, + Dirac Room RQ.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
I agree with all of that. And SOTA room correction does take care not to overcorrect by taking multiple measurements and focusing correction on deviations with low variance across the entire listening space, as you suggest.

Affordable is subjective but IMO something like the HTP-1 is quite affordable for multichannel + features and would allow for both the anechoic speaker PEQ correction, + Dirac Room RQ.

any room correction service that doesn't hire Mitch for you to make the filters for you is not SOTA. :p
 

yourmando

Active Member
Joined
Aug 14, 2020
Messages
150
Likes
178
any room correction service that doesn't hire Mitch for you to make the filters for you is not SOTA. :p

:p
1622732785459.gif
 

ooheadsoo

Active Member
Forum Donor
Joined
Aug 14, 2020
Messages
198
Likes
163
a digital crossover is exactly the same as an analog crossover (in function), it actually should be after AVR processing just like an analog crossover would.

Unless you're actually DESIGNING the crossover that is.



the industry is VERY QUICKLY moving away from channel based audio to object based audio (Atmos). you cannot have any output or sound at all without processing.
Going fully active, you would be designing the crossover, or implementing someone else's design at the same stage of signal processing.

Point taken about digitally encoded multichannel. I'm still stuck in 2.1 land.
 

mitchco

Addicted to Fun and Learning
Audio Company
Joined
May 24, 2016
Messages
640
Likes
2,397
it really depends on your objective, if you're optimizing for one location for one listener then yes you don't need the anechoic data, as long as your speaker and amplifier have enough headroom to allow the necessary modifications then you can achieve whatever response you want in that one location.

however the Olive/PIR score and model originated from the spatial average of 9 measurements and the location of each measurement was 3 meters apart from its next neighbor. it's a massive difference, and it's vital that both the on-axis and off-axis responses are consistent with each other to score high on the model.

that's why I like to optimize for PIR first, so the speaker's performance is more or less consistent throughout the whole room, and target curve deviations are minimized. this way something like Dirac would not need to make unnecessary assumption. Mitch says that it's important that the room correction is SOTA, I don't think any affordable room correction on the market is SOTA.

Well, it simply isn't true about optimising for just for one location as I have shown here. I have tested/compared just about every DSP/DRC product on the market and Audiolense, Acourate and Focus Fidelity are head and shoulders above everything else. EDIT: I see Richard beat me to it. Yah, all under $US 500. Here is a blurb on single vs multiple measurements - know your DRC software :) Finally, here is why no h/w AVR or PrePro will ever match purpose built DSP software.
 

q3cpma

Major Contributor
Joined
May 22, 2019
Messages
3,060
Likes
4,416
Location
France
Well, it simply isn't true about optimising for just for one location as I have shown here. I have tested/compared just about every DSP/DRC product on the market and Audiolense, Acourate and Focus Fidelity are head and shoulders above everything else. EDIT: I see Richard beat me to it. Yah, all under $US 500. Here is a blurb on single vs multiple measurements - know your DRC software :) Finally, here is why no h/w AVR or PrePro will ever match purpose built DSP software.
Do you have any opinion on the quality of brand specific software like GLM or Neumann/Fraunhofer's new MA-1?
 

johnp98

Active Member
Joined
Jan 29, 2021
Messages
133
Likes
201
Does your receiver have digital PEQ already? Many do. Also, does your receiver have room correction? That’s another way to get it to make corrections, even up in the speaker range. I would experiment with that first.

In this case, I would let it correct the whole curve, not just below Schroeder frequency. This is because the speaker badly needs correcting. The built in room eq might have enough horsepower to calculate all the corrections you need. I would re-measure at the end with REW to confirm you’re getting a nice smooth curve. That might be all you need. If your receiver’s DSP capability is too limited, it might not be able to make all of the corrections needed.!

I compared all the EQs that I have available to me (REW+EAPO, Audyssey, Dirac) in this post and it certainly showed that the DSP capabilities of my older version of Audyssey left a lot to be desired, I don't have a way to implement Dirac unfortunately, and REW+EAPO gave great results but only would work with PC based audio sources.
So I am looking for a way I could use high resolution EQ filters based on the upcoming klippel measurements of the HTM-12 v2 for EQ above the Schroder frequency and then use my own measurements for below the transition frequency.

I certainly will look into the Monoprice HTP-1. Thanks for the suggestion.
 
Last edited:

johnp98

Active Member
Joined
Jan 29, 2021
Messages
133
Likes
201
Not trying to rain on the parade but I do want to remind everyone watching that the test here is of the "v1" version. And that the "v2" version review is coming soon and shows more promise (albeit, at a loss of sensitivity). So, I wouldn't make any purchase decisions just yet. It might be worth it to simply retrofit a v1 design with v2 components.
Yeah I am very excited for the v2 review as that is the version that I have. I am just thinking that with the Klippel measurements of the v2 I could use to add a layer of active EQ to get it as flat as possible on axis, as I assume the v2 will have the same impressive directivity given that is uses the same waveguide.
 

mtg90

Member
Joined
May 24, 2021
Messages
53
Likes
127
Location
Illinois
Yeah I am very excited for the v2 review as that is the version that I have. I am just thinking that with the Klippel measurements of the v2 I could use to add a layer of active EQ to get it as flat as possible on axis, as I assume the v2 will have the same impressive directivity given that is uses the same waveguide.

One issue with that is I gave Erin a crossover with some minor tweaks, (adjusted the value of a cap and resistor slightly). I had intended to implement that version of the v2 crossover if it measured well as it looked a little better with the measurements I took before sending him the drivers & crossover but from the preview he shared with me it seems just leaving the original values would have measured slightly better. So any EQ generated from the data would be off a little (dB here or there) unless you made similar adjustments to the crossover.

When that review is posted I will show what the small change I had made on the crossover I gave him did to the response and why it would probably would have been better just leaving the crossover in its stock form.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
Well, it simply isn't true about optimising for just for one location as I have shown here. I have tested/compared just about every DSP/DRC product on the market and Audiolense, Acourate and Focus Fidelity are head and shoulders above everything else. EDIT: I see Richard beat me to it. Yah, all under $US 500. Here is a blurb on single vs multiple measurements - know your DRC software :) Finally, here is why no h/w AVR or PrePro will ever match purpose built DSP software.

if you were hired to optimize the response in a ball room where everyone in the room should have acceptable sound, do you think that anechoic data are not necessary in that situation?


One issue with that is I gave Erin a crossover with some minor tweaks, (adjusted the value of a cap and resistor slightly). I had intended to implement that version of the v2 crossover if it measured well as it looked a little better with the measurements I took before sending him the drivers & crossover but from the preview he shared with me it seems just leaving the original values would have measured slightly better. So any EQ generated from the data would be off a little (dB here or there) unless you made similar adjustments to the crossover.

When that review is posted I will show what the small change I had made on the crossover I gave him did to the response and why it would probably would have been better just leaving the crossover in its stock form.

@hardisj I hope you’re getting compensated for this level of fuckery.
 
OP
hardisj

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
@hardisj I hope you’re getting compensated for this level of fuckery.

If I'm being honest, it didn't sit well with me at first. I wanted a representative example of what is currently available. Unfortunately, I wasn't told what I had received until after I had tested it and sent Matt some preliminary results. That bothered me because I didn't test what I intended and thought I was testing.

But, at the end of the day, part of the goal (ultimately) is to also help improve the product for the community. So, I told myself it's toward the greater good and went on about my day.


That said, nope. Ain't no one compensating me anything for this.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
If I'm being honest, it didn't sit well with me at first. I wanted a representative example of what is currently available. Unfortunately, I wasn't told what I had received until after I had tested it and sent Matt some preliminary results. That bothered me because I didn't test what I intended and thought I was testing.

But, at the end of the day, part of the goal (ultimately) is to also help improve the product for the community. So, I told myself it's toward the greater good and went on about my day.


That said, nope. Ain't no one compensating me anything for this.

I mean it would have benefited everyone if he asked you to measure the drivers in the cabinet individually without a crossover at all, this way he would have more than enough data to build either an analog or a digital crossover (or anyone reading the reviews for that matter), I mean lets face it this speaker needs saving.

Although i wonder if you would agree to something like that since he will be earning all the $$$.

but this is just 'yikes' imo. it's predatory with very little long term benefit.
 

johnp98

Active Member
Joined
Jan 29, 2021
Messages
133
Likes
201
One issue with that is I gave Erin a crossover with some minor tweaks, (adjusted the value of a cap and resistor slightly). I had intended to implement that version of the v2 crossover if it measured well as it looked a little better with the measurements I took before sending him the drivers & crossover but from the preview he shared with me it seems just leaving the original values would have measured slightly better. So any EQ generated from the data would be off a little (dB here or there) unless you made similar adjustments to the crossover.
When that review is posted I will show what the small change I had made on the crossover I gave him did to the response and why it would probably would have been better just leaving the crossover in its stock form.

Well that is most unfortunate.
So we will have measurements to a "version 3" that never has nor will see the light of day o_O
I really hope that the changes in the crossover were subtle and truly only a few dB here or there as its very unfortunate that we wont get klippel measurements for the current and ongoing version.
 

johnp98

Active Member
Joined
Jan 29, 2021
Messages
133
Likes
201
If I'm being honest, it didn't sit well with me at first. I wanted a representative example of what is currently available. Unfortunately, I wasn't told what I had received until after I had tested it and sent Matt some preliminary results. That bothered me because I didn't test what I intended and thought I was testing.

That said, nope. Ain't no one compensating me anything for this.

Man I feel for you. That really sucks!
I was going to wait for the review to be up before donating to your cause. But yeah donation incoming right now, as this maturity is commendable. I hope it has not burnt future bridges with the DIY community in general.
 
Top Bottom