Author Topic: Do nice flat graph=Good Sound  (Read 1582 times)

Offline desray

  • Global Moderator
  • Trade Count: (+4)
  • Full Member
  • *
  • Posts: 19423
  • Bring the "Cinema experience" back home...
Re: Do nice flat graph=Good Sound
« Reply #45 on: October 08, 2019, 20:09 »
I think sometimes we just need to "really" sit down and get to enjoy our fruits of labor. Striving to achieve perfection is neigh impossible...our home is not an anechoic chamber but with off shapes and different furnishings etc...but we can do is to do our best to optimise the HT experience using whatever tools, skill sets and knowledge we have.

Offline kennyluck2000

  • Trade Count: (0)
  • Full Member
  • Posts: 236
Re: Do nice flat graph=Good Sound
« Reply #46 on: October 09, 2019, 01:49 »
If I may share my findings regarding TS question on flat graph and how it sounds.

Many who know me know I depend heavily on measurements to fine tune the tonal balance of my system. This is even more important for a full diy active setup where is no baseline voicing or tonal balance like a commercial speaker.

The challenge is drawing a correlation between the measurements, and what you actually hear as a good tonal balance. Here's what I found.

1.A flat curve from about 200hz to 20khz (measured nearfield 0.5m)is what I start as a baseline when I build the active crossover . Typically, this sounds a bit bright, and very slightly downward 1-2db slope from 1-2khz to 20khz is more balanced. This is related to the (harman) X curve but is actually due to air loss and room dissipation of higher frequency in larger rooms. As a result, the final balance that sounds 'right' is dependent on both room size as well as relative liveliness of the room furnishing. Not to mention personal preferences.

2. For the upper bass and low mids, next I dial in the BSC, baffle step compensation. For commercial speakers this is already part of the crossover, and is the reason why some speakers need more distance from walls to avoid sounding boomy and thick.
For diy, in typical rooms, a 3-4db low shelf at ~300hz, Q=0.6-0.7 is a good starting point. Due to the room, effects, this is difficult to measure, typically I adjust by ear with some choice test tracks that I'm familiar with.

3. Finally the low bass. Maybe a slight 3-5db shelf from 100hz to 50hz, then flat to the system rolloff. This is measured at MLP and this is when I EQ cut the room modal peaks as part of the response shaping.

With the above, I have tried active combinations of many different woofer, mid and tweeter, and come up with a system that actually sounds more similar than different. Tonal balance is similar based on the above steps , with remaining subjective differences in sound due to the relative directivity of the system (eg dome tweeter vs horn vs ribbon ), and diffraction nasties etc.

I see this as science supporting art. At the end of the day, it's about what sounds good, and getting there systematically instead of trial and error.


PS : everytime I do the above process with a new active speaker setup, I laugh at myself, because I'm literally an 'audiophile' who uses high quality music to listen to my system..

Here's some pics of my latest build.

Sent from my X9009 using Tapatalk

Thanks wechnivag for sharing your knowledges and experiances  8) cool