A few weeks ago NewRisingSun made an heroic effort to measure at least one copy of every N163 game. I measured what I had at the time too.
I want to keep going with this, so the last few days I wrote a set of simple and easy volume test ROMs for all the current expansions. If you don't have a Famicom or any expansion carts, just measure your APU. That still helps.
You can find the ROMs in the build folder here: https://github.com/bbbradsmith/nes-audio-tests
I can add more tests if I get my hands on other expansion audio carts, or if you've got some you can write tests for you can submit them to me. The source is there, the individual code for each test is pretty small.
The measuring process is like this:
1. Load the ROM in your dev-cart, hear the buzz when it starts.
2. Quickly pull it out*, and insert the cart to be measured. (You have a little more than 4 seconds.)
3. If you hear a second buzz, you succeeded. Otherwise start over.
4. The test will play and then loop, each time starting this a quick buzz.
5. Record the test and compare the RMS of the two tones. (Audacity's Analyze > Contrast tool will do this.)
6. Post here with your results.
* In my personal experience hotswapping cartridges has been a safe procedure, but some people are uncertain about this. I cannot make any guarantee, so if you are not comfortable with the risk, please do not do these tests. (The APU test does not require a hotswap, though. Neither does FDS if you can use the .FDS image.)
Each ROM plays an APU square wave at 440Hz, followed by an "equivalent" sound on the expansion. The sounds are played on a single channel at their maximum volume. The APU test doesn't require a hotswap (or a Famicom), it compares against the triangle. The VRC7 can't produce a square wave, so I used an instrument that has a 2:1 modulator at 50% modulation strength that makes a dull square-ish sound. (Maybe not the ideal comparison, but should be simple and reliable to execute.) The 5B has a strange amplifier, so its test uses volume 12 instead of 15, where it appears to be linear. The N163 always uses 1-channel mode.
Here's my test results. I have an AV modded Famicom (all the expansions are tested on this), and a Canadian NES. All the expansion audio tests were done with the Famicom. The APU tests were done with each of the machines:
Keep in mind that every time I do measurements like this the results are slightly different. Temperature variation etc. affects this, so we don't need to be super precise here. This is a shotgun survey approach. Accurate to within .1dB is probably already too fine. The last time I measured my N163 carts they both came out 0.3dB louder, for example.
This is partly motivated by wanting to get mixing data for N163 into the NSF format, but it would be great to have better survey data on this stuff just so we can get a better average for emulation.
Edit: the 5B test was revised after investigating its amplifier. VRC7 test was revised after feedback was suggested.
I want to keep going with this, so the last few days I wrote a set of simple and easy volume test ROMs for all the current expansions. If you don't have a Famicom or any expansion carts, just measure your APU. That still helps.
You can find the ROMs in the build folder here: https://github.com/bbbradsmith/nes-audio-tests
I can add more tests if I get my hands on other expansion audio carts, or if you've got some you can write tests for you can submit them to me. The source is there, the individual code for each test is pretty small.
The measuring process is like this:
1. Load the ROM in your dev-cart, hear the buzz when it starts.
2. Quickly pull it out*, and insert the cart to be measured. (You have a little more than 4 seconds.)
3. If you hear a second buzz, you succeeded. Otherwise start over.
4. The test will play and then loop, each time starting this a quick buzz.
5. Record the test and compare the RMS of the two tones. (Audacity's Analyze > Contrast tool will do this.)
6. Post here with your results.
* In my personal experience hotswapping cartridges has been a safe procedure, but some people are uncertain about this. I cannot make any guarantee, so if you are not comfortable with the risk, please do not do these tests. (The APU test does not require a hotswap, though. Neither does FDS if you can use the .FDS image.)
Each ROM plays an APU square wave at 440Hz, followed by an "equivalent" sound on the expansion. The sounds are played on a single channel at their maximum volume. The APU test doesn't require a hotswap (or a Famicom), it compares against the triangle. The VRC7 can't produce a square wave, so I used an instrument that has a 2:1 modulator at 50% modulation strength that makes a dull square-ish sound. (Maybe not the ideal comparison, but should be simple and reliable to execute.) The 5B has a strange amplifier, so its test uses volume 12 instead of 15, where it appears to be linear. The N163 always uses 1-channel mode.
Here's my test results. I have an AV modded Famicom (all the expansions are tested on this), and a Canadian NES. All the expansion audio tests were done with the Famicom. The APU tests were done with each of the machines:
- Gimmick! - db_5b: -1.3dB
- Erika to Satoru no Yumebouken: db_n163: 19.2dB
- Rolling Thunder - db_n163: 17.2dB
- Uchuu Keibi Tai SDF - db_mmc5: 1.0dB
- Just Breed - db_mmc5: 1.0dB
- Esper Dream 2 - db_vrc6b: 0.2dB
- Lagrange Point - db_vrc7: 11.3dB
- FDS - db_fds: 6.6
- Famicom - db_apu: -0.3dB
- NES - db_apu: 0.1dB
Keep in mind that every time I do measurements like this the results are slightly different. Temperature variation etc. affects this, so we don't need to be super precise here. This is a shotgun survey approach. Accurate to within .1dB is probably already too fine. The last time I measured my N163 carts they both came out 0.3dB louder, for example.
This is partly motivated by wanting to get mixing data for N163 into the NSF format, but it would be great to have better survey data on this stuff just so we can get a better average for emulation.
Edit: the 5B test was revised after investigating its amplifier. VRC7 test was revised after feedback was suggested.