This page uses extracted DeltaWave aligned-spectrum curves. Each chart compares decoded MP3 output against the same original WAV, making it easier to see high-frequency roll-off across archived releases and the later cVBRb boost extensions.
It is a lab page rather than a listening test. The current charts combine the earlier official releases with aligned-spectrum curves extracted from the submitted LAME 4.0 reports.
3.98.4 rolls off earliest, with 3.99.5 showing a later but still obvious cut in the extracted response curve.
3.100 and 4.0 remain close against the original WAV in the extracted aligned-spectrum curves.
The LAME 4.0 cVBRb boost levels extend the high-frequency response progressively in the submitted 192 kbps minimum reports.
The derived metric charts back up the visual story by showing later drop points and lower average upper-band loss for the boosted variants.
This is still not a listening-test verdict. These are extracted DeltaWave aligned-spectrum curves, useful for illustrating behaviour against the original WAV rather than proving subjective transparency.
These charts are extracted from embedded DeltaWave aligned-spectrum plots. They provide a direct visual comparison between the decoded MP3 outputs and the original WAV reference.
How to read it: the dashed line is the original WAV reference extracted from the same DeltaWave plots. Solid lines are the decoded build outputs. Click any response chart to open a larger version.
These derived charts make fuller use of the extracted aligned-spectrum curves by quantifying where each build begins to fall away from the original WAV and how much average upper-band loss remains.
How to read it: lower deviation means the extracted curve is closer to the original WAV in the upper band. Click any response chart to open a larger version.
| Series | 3 dB drop | 6 dB drop | 12 dB drop | Mean loss 18–21 kHz | Mean loss 19–22 kHz |
|---|---|---|---|---|---|
| 3.98.4 MinGW | 18.98 kHz | 19.00 kHz | 19.00 kHz | 26.61 dB | 37.57 dB |
| 3.99.5 MinGW | 19.30 kHz | 20.15 kHz | 21.53 kHz | 3.84 dB | 7.54 dB |
| 3.100 MinGW | 19.30 kHz | 20.15 kHz | 21.53 kHz | 3.84 dB | 7.54 dB |
| 4.0 V0 | 19.33 kHz | 20.15 kHz | 21.53 kHz | 3.86 dB | 7.56 dB |
| 4.0 V0 · 192 kbps minimum | 19.33 kHz | 20.15 kHz | 21.53 kHz | 3.85 dB | 7.56 dB |
| 4.0 V0 · 224 kbps minimum | 18.68 kHz | 20.15 kHz | 21.53 kHz | 4.09 dB | 7.81 dB |
| 4.0 cVBRb boost 1 · 192 minimum | 19.35 kHz | 20.53 kHz | 21.53 kHz | 2.99 dB | 6.65 dB |
| 4.0 cVBRb boost 2 · 192 minimum | 19.35 kHz | 20.53 kHz | 21.53 kHz | 3.23 dB | 6.97 dB |
| 4.0 cVBRb boost 3 · 192 minimum | 19.98 kHz | 20.80 kHz | 21.55 kHz | 2.20 dB | 5.84 dB |
| 4.0 cVBRb boost 1 · 224 minimum | 19.35 kHz | 20.53 kHz | 21.53 kHz | 3.00 dB | 6.65 dB |
| 4.0 cVBRb boost 2 · 224 minimum | 19.35 kHz | 20.53 kHz | 21.53 kHz | 3.23 dB | 6.97 dB |
| 4.0 cVBRb boost 3 · 224 minimum | 19.98 kHz | 20.80 kHz | 21.58 kHz | 2.20 dB | 5.85 dB |
These metrics are derived from the extracted aligned-spectrum curves. They are based on extracted chart geometry rather than direct FFT export data.
The source is 44.1 kHz, so the Nyquist limit is 22.05 kHz. This bar chart is retained as a quick proxy summary; the aligned-spectrum charts above are the primary view on this page.
How to read it: higher kHz means the image-derived proxy found activity extending higher in frequency. It does not automatically mean better audible quality.
| Compile / configuration | Approx. high-frequency extent |
|---|---|
| 3.98.4 MinGW | 18.95 kHz |
| 3.99.5 MinGW | 19.73 kHz |
| 3.100 MinGW | 19.73 kHz |
| 4.0 V0 | 19.73 kHz |
| 4.0 V0 · 192 kbps minimum | 19.73 kHz |
| 4.0 V0 · 224 kbps minimum | 19.70 kHz |
| 4.0 cVBRb boost 1 · 192 minimum | 19.75 kHz |
| 4.0 cVBRb boost 2 · 192 minimum | 19.75 kHz |
| 4.0 cVBRb boost 3 · 192 minimum | 20.10 kHz |
| 4.0 cVBRb boost 1 · 224 minimum | 19.75 kHz |
| 4.0 cVBRb boost 2 · 224 minimum | 19.75 kHz |
| 4.0 cVBRb boost 3 · 224 minimum | 20.10 kHz |
VBR/cVBRb visual proxy derived from the embedded DeltaWave report image. Treat this as approximate, not as a laboratory frequency-response measurement.
These charts compare LAME 4.0 V0 with a 192 kbps minimum against cVBRb boost levels 1–3 using the same minimum. They are extracted from the aligned-spectrum plots in the submitted DeltaWave reports.
Important: these are extracted from DeltaWave aligned-spectrum images. They are much closer to the intended original-WAV comparison story, but they are still extracted curves rather than direct raw FFT exports.
The main response charts on this page are extracted from embedded DeltaWave aligned-spectrum SVG plots. They compare each decoded output with the original WAV reference while keeping the workflow lightweight.
The response charts use the IBM Carbon data-visualization palette for categorical series, with a neutral dashed reference line for the original WAV.
The supporting high-frequency-extent bar chart is retained as a compact summary, but it is secondary. The new response charts and derived metrics are the primary narrative: they show the roll-off differences across archived releases, the closeness of 3.100 and 4.0, and the extension introduced by the LAME 4.0 cVBRb boost levels.
The derived metrics here include 3 dB / 6 dB / 12 dB drop points and average loss across the 18–21 kHz and 19–22 kHz bands. Those measures are approximate because they come from extracted plot geometry rather than direct numeric spectral exports.
For more exact cutoff, roll-off, and high-frequency energy measurements, numeric spectral exports or a repeatable FFT-based workflow would be preferable.
No. The main charts are extracted from DeltaWave aligned-spectrum images, not from raw FFT exports. They are useful for comparing build behaviour, but they remain extracted visual data rather than lab-grade response measurements.
No. A wider visible range is not automatically a quality verdict. The chart shows output behaviour against this reference sample, not listener preference or transparency.
No. Both the cVBRb and ABR boost charts on this page are extracted from DeltaWave aligned-spectrum plots. They are useful for comparing the tested builds, but they remain extracted chart data rather than direct raw FFT exports.